# Copilot Section
Source: https://docs.cognigy.com/agent-copilot/configure/copilot-section

Learn how to configure the Copilot section in Endpoint settings to integrate Agent Copilot with a contact center platform.

<a href="/release-notes/2026.17"><Badge>Updated in 2026.17</Badge></a>

The **Copilot** section in the [Endpoint settings](/ai/agents/deploy/endpoints/overview) is used to configure the Agent Copilot workspace.

The following Endpoints include the **Copilot** section:

* [NiCE CXone Agent Assist Hub (AAH)](/agent-copilot/deploy/nice-aah)
* [LINE](/ai/agents/deploy/endpoint-reference/line-messenger)
* [Facebook (Meta) Messenger](/ai/agents/deploy/endpoint-reference/facebook-messenger)
* [8x8](/ai/agents/deploy/endpoint-reference/8x8)
* [Voice Gateway](/ai/agents/deploy/endpoint-reference/voice-gateway)
* [Slack](/ai/agents/deploy/endpoint-reference/slack)
* [Sunshine Conversations](/ai/agents/deploy/endpoint-reference/sunshine-conversations)
* [Socket.IO](/ai/agents/deploy/endpoint-reference/socketio)
* [Azure Bot Services](/ai/agents/deploy/endpoint-reference/azure-bot-services)
* [Genesys](/ai/agents/deploy/endpoint-reference/genesys)
* [Webhook](/ai/agents/deploy/endpoint-reference/webhook)
* [Webchat v3](/webchat/v3/configuration)

## Prerequisites

* If you use **Token Secret** authentication, you need a 256-bit string secret for signing.

## Limitations

* Cognigy caches JWKS responses for up to 5 key store URLs at once. If an additional URL is used, the new URL replaces the oldest one in the cache.
* Cognigy sends up to 10 requests/minute to any single JWKS URL.

## Restrictions

* EdDSA (Ed25519/Ed448) keys aren't supported for **Public Key** or **Key Store** authentication.

## Parameters

| Parameter                    | Type   | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
| ---------------------------- | ------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Copilot Type                 | List   | Activates and sets how to configure Agent Copilot. Select one of the following options:<ul><li>**None** — deactivates Agent Copilot for the Endpoint. This option is selected by default.</li><li>**Workspace** — activates Agent Copilot based on the [workspace](/agent-copilot/configure/workspace) configuration.</li><li>**Whisper** — activates [Agent Copilot Whisper](/live-agent/assistants/agent-copilot-whisper).</li></ul>                                                                                                                                                                                                                                                         |
| Copilot Flow                 | List   | Sets the Flow in which the Agent Copilot workspace or the Agent Copilot Whisper is configured.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
| Copilot Config               | List   | Sets the configuration for the Agent Copilot workspace. Valid for workspaces configured via [API](/api-reference/agent-copilot-configurations/get-agent-copilot-configurations).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
| Authentication               | List   | Sets the method to verify the JWT passed in the Agent Copilot URL. Enforce the method by activating **Accept Only JWT Requests**. Select one of the following methods:<ul><li>**Default (built-in)** — built-in authentication using an internal Agent Copilot secret.</li><li>**Token Secret** — symmetric authentication using a shared secret stored through a [connection](/ai/platform-features/connections).</li><li>**Public Key** — asymmetric authentication using one or more PEM-encoded public keys.</li><li>**Key Store** — asymmetric authentication using a remote JWKS endpoint.</li></ul>For more information, see [JWT Authentication Methods](#jwt-authentication-methods). |
| Agent Copilot Authentication | List   | This parameter appears only when **Token Secret** is selected in **Authentication**. Defines the JWT connection to use for symmetric authentication. Select a JWT connection or create one by clicking <img alt="add JWT connection" /> and configuring the connection:<ul><li>**Connection name** — enter a unique name.</li><li>**JWT Secret** — enter the string to be used as the JWT secret.</li></ul>                                                                                                                                                                                                                                                                                    |
| Public Keys                  | Text   | This parameter appears only when **Public Key** is selected in **Authentication**. Defines one or more PEM-encoded public keys. A token is accepted if it verifies with any of the configured public keys. Use multiple keys to support key rotation. Remove a key to immediately invalidate tokens signed with it.                                                                                                                                                                                                                                                                                                                                                                            |
| Key Store URL                | Text   | This parameter appears only when **Key Store** is selected in **Authentication**. Defines the URL of your JWKS key store. The verifying key is selected by the `kid` header of the incoming token. A JWT without a `kid` header is rejected. Cognigy caches the JWKS document for up to 10 minutes per URL. After rotating a key, tokens signed with the new key can fail and tokens signed with the removed key can still succeed for up to 10 minutes.                                                                                                                                                                                                                                       |
| Accept Only JWT Requests     | Toggle | Activates JWT validation according to the method selected in **Authentication**. Requests that include only a user ID and session ID are blocked. This parameter is deactivated by default. Activating this parameter improves security by preventing unauthenticated access. When deactivated, Agent Copilot still extracts `userId` and `sessionId` if a JWT is present in the URL, but the signature isn't verified. For more information, see [JWT Authentication Methods](#jwt-authentication-methods).                                                                                                                                                                                   |

<Note>
  As of Cognigy 2026.15.0, the **Authentication** list replaces the former **Enable Copilot Authentication** toggle. Previously configured Endpoints are migrated automatically. If the former toggle was:

  * Turned on — **Token Secret** is selected.
  * Turned off — **Default (built-in)** is selected.

  No action is required.
</Note>

## JWT Authentication Methods

To secure access to Agent Copilot, the contact center platform signs a JWT containing `userId` and `sessionId`, then passes that signed JWT in the Agent Copilot URL as the `auth` parameter:

```txt theme={null}
https://${AgentCopilotBaseUrl}/?URLToken=${URLToken}&auth={signed_JWT}
```

Each contact center platform has its own requirements for how the JWT signature is configured. Refer to your contact center platform documentation.

Depending on the option you select in the **Authentication** list, you configure the authentication method differently:

<Tabs>
  <Tab title="Default (built-in)">
    Agent Copilot handles the JWT authentication between the contact center platform and Cognigy. Activate **Accept Only JWT Requests** in the [Endpoint settings](#parameters).
  </Tab>

  <Tab title="Token Secret">
    Symmetric authentication. The contact center platform and Cognigy share a single secret, used to both sign and verify the JWT.

    <AccordionGroup>
      <Accordion title="1. On the Contact Center Side">
        1. Generate a 256-bit secret to use as the shared JWT secret. For example, you can use [OpenSSL](https://openssl-library.org/):

           ```sh theme={null}
           # Generate 32 random bytes (256 bits) and print them as a hex string
           openssl rand -hex 32
           ```

           Example output: `a3f7b9c2d8e14f6a0b5c9d3e7f2a8b4c6d0e9f3a7b1c5d8e2f4a6b9c3d7e1f85`

        2. Copy the secret for later use and store it securely.

        3. Sign JWTs with this secret. Agent Copilot accepts the `HS256`, `HS384`, and `HS512` HMAC algorithms. For example, use the [`jsonwebtoken` Node.js library](https://www.npmjs.com/package/jsonwebtoken):

           <CodeGroup>
             ```js HS256 theme={null}
             const jwt = require('jsonwebtoken'); // JWT signing library

             const token = jwt.sign(
               { userId: 'user-123', sessionId: 'session-456' }, // payload
               'a3f7b9c2d8e14f6a0b5c9d3e7f2a8b4c6d0e9f3a7b1c5d8e2f4a6b9c3d7e1f85', // shared secret
               { algorithm: 'HS256' } // HMAC using SHA-256
             );
             ```

             ```js HS384 theme={null}
             const jwt = require('jsonwebtoken'); // JWT signing library

             const token = jwt.sign(
               { userId: 'user-123', sessionId: 'session-456' }, // payload
               'a3f7b9c2d8e14f6a0b5c9d3e7f2a8b4c6d0e9f3a7b1c5d8e2f4a6b9c3d7e1f85', // shared secret
               { algorithm: 'HS384' } // HMAC using SHA-384
             );
             ```

             ```js HS512 theme={null}
             const jwt = require('jsonwebtoken'); // JWT signing library

             const token = jwt.sign(
               { userId: 'user-123', sessionId: 'session-456' }, // payload
               'a3f7b9c2d8e14f6a0b5c9d3e7f2a8b4c6d0e9f3a7b1c5d8e2f4a6b9c3d7e1f85', // shared secret
               { algorithm: 'HS512' } // HMAC using SHA-512
             );
             ```
           </CodeGroup>
      </Accordion>

      <Accordion title="2. On the Cognigy Side">
        1. In the **Copilot** section in the Endpoint settings, select **Token Secret** in **Authentication**.
        2. Next to **Agent Copilot Authentication**, click <img alt="add JWT connection" /> and configure the connection:
           * **Connection name** — enter a unique name, for example, `copilot-token-secret`.
           * **JWT Secret** — enter the exact same 256-bit secret you previously generated for the contact center platform, for example, `a3f7b9c2d8e14f6a0b5c9d3e7f2a8b4c6d0e9f3a7b1c5d8e2f4a6b9c3d7e1f85`.
        3. Activate **Accept Only JWT Requests**.
        4. Click **Create**, then **Save**.
      </Accordion>
    </AccordionGroup>
  </Tab>

  <Tab title="Public Key">
    Asymmetric authentication. The contact center signs the JWT with a private key, and Cognigy verifies it with the corresponding public key.

    <AccordionGroup>
      <Accordion title="1. On the Contact Center Side">
        1. Generate a public-private key pair. For example, you can use [OpenSSL](https://openssl-library.org/):

           **RSA:**

           ```sh theme={null}
           # Generate a 2048-bit RSA private key
           openssl genpkey -algorithm RSA -pkeyopt rsa_keygen_bits:2048 -out private.pem
           # Derive the matching public key
           openssl pkey -in private.pem -pubout -out public.pem
           ```

           **ECDSA P-256:**

           ```sh theme={null}
           # Generate a private key on the P-256 curve
           openssl ecparam -name prime256v1 -genkey -noout -out private.pem
           # Derive the matching public key
           openssl ec -in private.pem -pubout -out public.pem
           ```
        2. *(Optional)* If you want to use key rotation, generate multiple public-private key pairs.
        3. Keep `private.pem` secret and store it in your own system.
        4. Sign JWTs with `private.pem`. The algorithm you use depends on the key type. For example, use the [`jsonwebtoken` Node.js library](https://www.npmjs.com/package/jsonwebtoken):

           <CodeGroup>
             ```js RS256 theme={null}
             const jwt = require('jsonwebtoken'); // JWT signing library
             const fs = require('fs');

             const privateKey = fs.readFileSync('private.pem'); // RSA private key

             const token = jwt.sign(
               { userId: 'user-123', sessionId: 'session-456' }, // payload
               privateKey,
               { algorithm: 'RS256' } // RSA key, PKCS#1 v1.5 padding, SHA-256
             );
             ```

             ```js RS384 theme={null}
             const jwt = require('jsonwebtoken'); // JWT signing library
             const fs = require('fs');

             const privateKey = fs.readFileSync('private.pem'); // RSA private key

             const token = jwt.sign(
               { userId: 'user-123', sessionId: 'session-456' }, // payload
               privateKey,
               { algorithm: 'RS384' } // RSA key, PKCS#1 v1.5 padding, SHA-384
             );
             ```

             ```js RS512 theme={null}
             const jwt = require('jsonwebtoken'); // JWT signing library
             const fs = require('fs');

             const privateKey = fs.readFileSync('private.pem'); // RSA private key

             const token = jwt.sign(
               { userId: 'user-123', sessionId: 'session-456' }, // payload
               privateKey,
               { algorithm: 'RS512' } // RSA key, PKCS#1 v1.5 padding, SHA-512
             );
             ```

             ```js PS256 theme={null}
             const jwt = require('jsonwebtoken'); // JWT signing library
             const fs = require('fs');

             const privateKey = fs.readFileSync('private.pem'); // RSA private key

             const token = jwt.sign(
               { userId: 'user-123', sessionId: 'session-456' }, // payload
               privateKey,
               { algorithm: 'PS256' } // RSA key, RSA-PSS padding, SHA-256
             );
             ```

             ```js PS384 theme={null}
             const jwt = require('jsonwebtoken'); // JWT signing library
             const fs = require('fs');

             const privateKey = fs.readFileSync('private.pem'); // RSA private key

             const token = jwt.sign(
               { userId: 'user-123', sessionId: 'session-456' }, // payload
               privateKey,
               { algorithm: 'PS384' } // RSA key, RSA-PSS padding, SHA-384
             );
             ```

             ```js PS512 theme={null}
             const jwt = require('jsonwebtoken'); // JWT signing library
             const fs = require('fs');

             const privateKey = fs.readFileSync('private.pem'); // RSA private key

             const token = jwt.sign(
               { userId: 'user-123', sessionId: 'session-456' }, // payload
               privateKey,
               { algorithm: 'PS512' } // RSA key, RSA-PSS padding, SHA-512
             );
             ```

             ```js ES256 theme={null}
             const jwt = require('jsonwebtoken'); // JWT signing library
             const fs = require('fs');

             const privateKey = fs.readFileSync('private.pem'); // ECDSA private key, P-256 curve

             const token = jwt.sign(
               { userId: 'user-123', sessionId: 'session-456' }, // payload
               privateKey,
               { algorithm: 'ES256' } // ECDSA key on the P-256 curve, SHA-256
             );
             ```
           </CodeGroup>

           `ES384` and `ES512` work the same way, but need a key generated on the P-384 or P-521 curve instead (`openssl ecparam -name secp384r1 ...` or `-name secp521r1 ...`) — the algorithm is tied to the curve the key was generated on.
      </Accordion>

      <Accordion title="2. On the Cognigy Side">
        1. In the **Copilot** section in the Endpoint settings, select **Public Key** in **Authentication**.
        2. Enter the contents of the `public.pem` file into the **Public Keys** field.

        <Accordion title="Public Key">
          ```txt theme={null}
          -----BEGIN PUBLIC KEY-----
          MIICITANBgkqhkiG9w0BAQEFAAOCAg4AMIICCQKCAgBTPbIGqvbkfo55bY7IRJ8U
          RUysKvl9X3qgSoJyDJhuydav7ssnxWnYFtBLziDQNeh1NsWF97bP7RkuCpyhnGYd
          GuKfPa6NNX8bVrzDwMtI6nwQ8KShhyiYfkQAHz7hTh/yfF5W5rg1oRH0nuUtp/w4
          ukwaOHvDuz2td9pdZO6cNq1e1K96dOd7P+VRu3W27xM0cx5u5NHSFargz2WM/6+M
          -----END PUBLIC KEY-----
          ```
        </Accordion>

        After you enter a public key in the field, another input field appears. Use the new input field to add another public key.

        3. Activate **Accept Only JWT Requests**.
        4. Click **Save**.
      </Accordion>
    </AccordionGroup>
  </Tab>

  <Tab title="Key Store">
    Asymmetric authentication. The contact center signs the JWT with a private key, and Cognigy verifies it with the corresponding public key by fetching it from a JWKS (JSON Web Key Set) endpoint you host.

    <AccordionGroup>
      <Accordion title="1. On the Contact Center Side">
        1. Generate a public-private key pair, the same way as for **Public Key**. For example, you can use [OpenSSL](https://openssl-library.org/):

           ```sh theme={null}
           # Generate a 2048-bit RSA private key
           openssl genpkey -algorithm RSA -pkeyopt rsa_keygen_bits:2048 -out private.pem
           # Derive the matching public key
           openssl pkey -in private.pem -pubout -out public.pem
           ```

        2. Host a JWKS endpoint that returns your public keys as JSON, each with a unique `kid`, for example:

           ```json theme={null}
           {
             "keys": [
               {
                 "kty": "RSA",
                 "use": "sig",
                 "kid": "copilot-key-2026-01",
                 "alg": "RS256",
                 "n": "0vx7agoebGcQSuuPiLJXZptN9nndrQmbXEps2aiAFbWhM78LhWx4cbbfAAtVT86zwu1RK7aPFFxuhDR1L6tSoc_BJECPebWKRXjBZCiFV4n3oknjhMstn64tZ_2W-5JsGY4Hc5n9yBXArwl93lqt7_RN5w6Cf0h4QyQ5v-65YGjQR0_FDW2QvzqY368QQMicAtaSqzs8KJZgnYb9c7d0zgdAZHzu6qMQvRL5hajrn1n91CbOpbISD08qNLyrdkt-bFTWhAI4vMQFh6WeZu0fM4lFd2NcRwr3XPksINHaQ-G_xBniIqbw0Ls1jF44-csFCur-kEgU8awapJzKnqDKgw",
                 "e": "AQAB"
               }
             ]
           }
           ```

        3. Sign JWTs with the matching private key, and set the `kid` header to the same value published in the JWKS. A JWT without a `kid` header is rejected. The algorithm you use depends on the key type, the same as for **Public Key**:

           <CodeGroup>
             ```js RS256 theme={null}
             const token = jwt.sign(
               { userId: 'user-123', sessionId: 'session-456' }, // payload
               privateKey,
               {
                 algorithm: 'RS256', // RSA key, PKCS#1 v1.5 padding, SHA-256
                 keyid: 'copilot-key-2026-01' // must match a "kid" in the JWKS
               }
             );
             ```

             ```js RS384 theme={null}
             const token = jwt.sign(
               { userId: 'user-123', sessionId: 'session-456' }, // payload
               privateKey,
               {
                 algorithm: 'RS384', // RSA key, PKCS#1 v1.5 padding, SHA-384
                 keyid: 'copilot-key-2026-01' // must match a "kid" in the JWKS
               }
             );
             ```

             ```js RS512 theme={null}
             const token = jwt.sign(
               { userId: 'user-123', sessionId: 'session-456' }, // payload
               privateKey,
               {
                 algorithm: 'RS512', // RSA key, PKCS#1 v1.5 padding, SHA-512
                 keyid: 'copilot-key-2026-01' // must match a "kid" in the JWKS
               }
             );
             ```

             ```js PS256 theme={null}
             const token = jwt.sign(
               { userId: 'user-123', sessionId: 'session-456' }, // payload
               privateKey,
               {
                 algorithm: 'PS256', // RSA key, RSA-PSS padding, SHA-256
                 keyid: 'copilot-key-2026-01' // must match a "kid" in the JWKS
               }
             );
             ```

             ```js PS384 theme={null}
             const token = jwt.sign(
               { userId: 'user-123', sessionId: 'session-456' }, // payload
               privateKey,
               {
                 algorithm: 'PS384', // RSA key, RSA-PSS padding, SHA-384
                 keyid: 'copilot-key-2026-01' // must match a "kid" in the JWKS
               }
             );
             ```

             ```js PS512 theme={null}
             const token = jwt.sign(
               { userId: 'user-123', sessionId: 'session-456' }, // payload
               privateKey,
               {
                 algorithm: 'PS512', // RSA key, RSA-PSS padding, SHA-512
                 keyid: 'copilot-key-2026-01' // must match a "kid" in the JWKS
               }
             );
             ```

             ```js ES256 theme={null}
             const token = jwt.sign(
               { userId: 'user-123', sessionId: 'session-456' }, // payload
               privateKey, // ECDSA private key, P-256 curve
               {
                 algorithm: 'ES256', // ECDSA key on the P-256 curve, SHA-256
                 keyid: 'copilot-key-2026-01' // must match a "kid" in the JWKS
               }
             );
             ```
           </CodeGroup>

           `ES384` and `ES512` work the same way, with a key generated on the P-384 or P-521 curve instead.
      </Accordion>

      <Accordion title="2. On the Cognigy Side">
        1. In the **Copilot** section in the Endpoint settings, select **Key Store** in **Authentication**.

        2. Enter your JWKS URL in the **Key Store URL** field. The domain is specific to your system; the path is only a convention, not a requirement — any URL that returns a valid JWKS document works. For example:

           ```txt theme={null}
           https://<your-domain>/.well-known/jwks.json
           ```

        3. Activate **Accept Only JWT Requests**.

        4. Click **Save**.
      </Accordion>
    </AccordionGroup>
  </Tab>
</Tabs>

## More Information

* [Endpoint Settings](/ai/agents/deploy/endpoints/overview)


# Grid Example: Knowledge Retrieval
Source: https://docs.cognigy.com/agent-copilot/configure/examples/knowledge-retrieval

Configure Agent Copilot for knowledge retrieval to surface relevant information to human agents during live conversations.

This guide shows how to configure Agent Copilot for knowledge retrieval using [Knowledge AI](/ai/agents/develop/knowledge-ai/overview). In this guide, you will use a [Copilot: Knowledge Tile](/ai/agents/develop/node-reference/agent-copilot/knowledge-tile) Node to integrate RAG capabilities into your contact center platform. This allows Agent Copilot to bring up information from your knowledge base to human agents to answer user questions. The AI Agent hands over a conversation to the human agent. The human agent receives the inquiry via chat and can search the answer in an embedded knowledge widget in the Agent Copilot workspace.

In this example, you can use this [CTXT sample file](/_assets/ai/empower/knowledge-ai/knowledge-ai-pdf-sample.ctxt) about Knowledge AI to create a [Knowledge Source](/ai/agents/develop/knowledge-ai/knowledge-source/overview#working-with-knowledge-sources).

## Prerequisites

* Configure a contact center platform that you can integrate with a [handover provider](/ai/escalate/handover-reference/overview). For example, for [Live Agent](/ai/escalate/handover-reference/live-agent), you need to have an [Inbox](/live-agent/settings/inboxes) with the [Project ID and the Webhook URL set up](/live-agent/settings/inboxes#edit-inboxes) and, if necessary, the **Enable Agent Copilot Workspace** option activated.
* Create a [handover provider](/ai/escalate/handover-reference/overview) that integrates with your contact center platform.
* Add a [Large Language Model (LLM) to your Project](/ai/agents/develop/gen-ai-and-llms/providers/all-providers) that supports [Knowledge Search and Answer Extracting](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature).
* Create a [Knowledge Store](/ai/agents/develop/knowledge-ai/knowledge-store) and [Knowledge Sources](/ai/agents/develop/knowledge-ai/knowledge-source/overview) in the Knowledge Store.

## Configure Knowledge Retrieval for Agent Copilot

### Configure the AI Agent Flow

<Steps>
  <Step title="Create a Flow and Add a Say Node for Greeting">
    1. In **Build > Flows**, create a Flow and give it a name, for example, `Knowledge Retrieval Agent Flow`.
    2. Add a Say Node and configure the following settings:
       * **Text** — enter a greeting message, for example, `Hi, do you have any questions regarding Knowledge AI?`. Save the Node.
  </Step>

  <Step title="Add a Handover to Human Agent Node">
    1. Below the Say Node, add a [Handover to Human Agent Node](/ai/agents/develop/node-reference/service/handover-to-agent) and configure the following settings:
       * **Handover Settings** section, select your [handover provider](/ai/escalate/handover-reference/overview) from the **Handover Provider** list. Depending on the handover provider, you need to configure additional settings.
       * **Handover Accepted Message** — enter a message to be displayed when the handover is accepted, for example, `I'll hand over to a human agent who can look into this matter for you.`. Save the Node.
  </Step>
</Steps>

### Configure the Agent Copilot Flow

The Copilot: Knowledge Tile Node requires a specific configuration to work correctly. In this example, you need an [If Node](/ai/agents/develop/node-reference/logic/if), a [Once Node](/ai/agents/develop/node-reference/logic/once) and two Copilot: Knowledge Tile Nodes. For more information, see [Copilot: Knowledge Tile Node](/ai/agents/develop/knowledge-ai/knowledge-search-and-extraction/agent-copilot-knowledge).

<Steps>
  <Step title="Create a Flow and Set the If Node">
    1. In **Build > Flows**, create a Flow and give it a clear name, for example, `Knowledge Retrieval Copilot Flow`.
    2. After the **Start** Node, add an [If Node](/ai/agents/develop/node-reference/logic/if). Under the **Condition** field, click **Use Advanced Editor** and enter the following:

       ```txt theme={null}
       input.data._cognigy._agentAssist.payload.tileId === "knowledge"
       ```

       If you set this condition to `knowledge`, the ID of the Copilot: Knowledge Tile Node needs to be set to `knowledge`.
  </Step>

  <Step title="Add the First Copilot: Knowledge Tile Node">
    After the **Then** Child Node of the If Node, add a [Copilot: Knowledge Tile Node](/ai/agents/develop/node-reference/agent-copilot/knowledge-tile) and configure the following settings:

    * **Tile ID** — enter `knowledge`.
    * **Knowledge Source** — select a Knowledge Source, for example, the Knowledge Source you created with the CTXT sample file about Knowledge AI.
  </Step>

  <Step title="Add a Set Grid Node">
    1. After the **Else** Child Node, add a [Copilot: Set Grid Node](/ai/agents/develop/node-reference/agent-copilot/set-grid) and set the **Copilot Grid Configuration** as follows:

       ```json theme={null}
       {
         "grid": {
           "columns": 2,
           "rows": 2,
           "gap": 5
         },
         "tiles": {
           "next-action": {
             "x": 1,
             "y": 1,
             "rows": 1,
             "columns": 1
           },
           "knowledge": {
             "x": 1,
             "y": 2,
             "rows": 1,
             "columns": 1
           }
         }
       }
       ```
  </Step>

  <Step title="Add the Copilot Nodes to the Flow">
    1. Below the Copilot: Set Grid Node, add an [LLM Prompt Node](/ai/agents/develop/node-reference/service/llm-prompt) and configure it to generate the next best response for the agent based on the user's latest input and to store the result in the Input object:

       * **System Prompt** — enter the following:
         ```txt theme={null}
         Based on the user's last message, suggest the most helpful next response for a human agent.
         ```
       * In the **Storage & Streaming** section, configure the following:
         * **How to handle the result** — set to **Store in Input**.
         * **Input Key to store Result** — set to `input.promptResult`.

    2. Add a [Copilot: Next Action Tile Node](/ai/agents/develop/node-reference/agent-copilot/next-action-tile) after the LLM Prompt Node and configure the Node as follows:
       * **Tile ID**: `next-action`
       * **Text**:
       ```txt theme={null}
       {{input.promptResult.detailedResult.choices[0].message.content}}
       ```

    3. Add a [Copilot: Knowledge Tile Node](/ai/agents/develop/node-reference/agent-copilot/knowledge-tile) and configure it as follows:
       * **Tile ID**: `knowledge`
       * **Knowledge Source**: select a Knowledge Source, for example, the Knowledge Source you created with the CTXT sample file about Knowledge AI. Save the Node.
  </Step>

  <Step title="Add a Once Node and the Second Copilot: Knowledge Tile Node">
    1. Below the If Node, add a [Once Node](/ai/agents/develop/node-reference/logic/once).
    2. Copy the first Copilot: Knowledge Tile Node by right-clicking it and selecting **Copy**.
    3. Below the **On the First Time** Node, right-click <img alt="add node" /> and select **Paste** to add a [Copilot: Knowledge Tile Node](/ai/agents/develop/node-reference/agent-copilot/knowledge-tile).
  </Step>
</Steps>

With this grid, the Agent Copilot workspace displays a widget with LLM-generated answers and a widget where you can search the answer in the Knowledge Source. You can test the Knowledge Tile with [Agent Copilot for chat applications](/agent-copilot/getting-started/chat).

## More Information

* [Knowledge AI](/ai/agents/develop/knowledge-ai/overview)
* [Knowledge Source](/ai/agents/develop/knowledge-ai/knowledge-source/overview)
* [Copilot: Knowledge Tile Node](/ai/agents/develop/node-reference/agent-copilot/knowledge-tile)
* [Copilot: Next Action Tile Node](/ai/agents/develop/node-reference/agent-copilot/next-action-tile)
* [Copilot: Set Grid Node](/ai/agents/develop/node-reference/agent-copilot/set-grid)
* [Handover to Human Agent Node](/ai/agents/develop/node-reference/service/handover-to-agent)


# Configuration Examples
Source: https://docs.cognigy.com/agent-copilot/configure/examples/overview

Explore Agent Copilot configuration examples including real-time translation and knowledge retrieval setups for contact center workspace integration.

This section provides examples of Agent Copilot configurations that you can use to configure the Agent Copilot workspace.

<CardGroup>
  <Card title="Real-Time Translation" href="/agent-copilot/configure/examples/real-time-translation">
    This configuration example shows how to configure the Agent Copilot workspace for real-time translation with chat-to-voice applications.
  </Card>

  <Card title="Knowledge Retrieval" href="/agent-copilot/configure/examples/knowledge-retrieval">
    This configuration example shows how to configure the Agent Copilot workspace for knowledge retrieval with Knowledge AI.
  </Card>
</CardGroup>


# Grid Example: Real-Time Translation
Source: https://docs.cognigy.com/agent-copilot/configure/examples/real-time-translation

Configure Agent Copilot for real-time translation in voice-to-chat conversations with multilingual support.

This guide shows how to configure Agent Copilot for real-time translation in a voice-to-chat application. In this example, the AI Agent receives calls, recognizes the language of the caller, and hands them over to human agents. The human agent can chat and the user receives voice responses in their own language, both in their respective languages.

This guide includes configuring the following:

* AI Agent Flow — configures the AI Agent to receive calls, recognize the caller's language, and hand them over to human agents.
* Agent Copilot Flow — configures the Agent Copilot workspace and the real-time translation widget.

For a comprehensive guide to set up a voice-to-chat application with Agent Copilot, see [Getting Started with Agent Copilot for Voice-to-Chat](/agent-copilot/getting-started/voice-to-chat).

## Prerequisites

* Access to [Voice Gateway](/voice-gateway/index).
* Configure a speech provider in [Cognigy](/ai/agents/test/voice-preview#set-up-a-speech-provider).
* Configure settings in the Voice Gateway Self-Service Portal:
  * The [phone number](/voice-gateway/webapp/phone-numbers) you use in your contact center.
  * A Voice Gateway [Application](/voice-gateway/webapp/applications).
  * A [speech provider](/voice-gateway/webapp/speech-services).
* Create a [Voice Gateway Endpoint](/ai/agents/deploy/endpoint-reference/voice-gateway).
* Create a [handover provider](/ai/escalate/handover-reference/overview) that integrates with your contact center platform.
* Configure an LLM that supports [LLM Prompt Nodes](/ai/agents/develop/node-reference/service/llm-prompt).

## Configure Real-Time Translation for Agent Copilot

### Configure the AI Agent Flow

<Steps>
  <Step title="Create a Flow and Add a Say Node for Greeting">
    1. In **Build > Flows**, create a Flow and give it a name, for example, `Real-Time Translation Agent Flow`.
    2. Add a Say Node and configure the following settings:
       * **Text** — enter a greeting message, for example, `Hello! Welcome to our customer service line!`. Save the Node.
  </Step>

  <Step title="Add a Question Node for Language Recognition">
    1. Below the Say Node, add a Question Node and configure the following settings:
       1. **Question Type** — select **Text**.
       2. Click <img alt="add channel" /> and select **Voice Gateway** from the list.
       3. **Text** — enter `What is your preferred language?`.
       4. Activate **Set Activity Parameters**.
       5. In the **Recognizer (STT)** section, select your speech provider from the **STT Vendor** list.
       6. **STT Language** — select the Flow language, for example, English (United States).
       7. Activate **Recognize Language** and select the languages you want to recognize, for example, `German (Germany)`, `Portuguese (Brazil)`, from the **Alternative Language 1** and **Alternative Language 2** lists. Save the Node.
  </Step>

  <Step title="Add a Lookup Node for Language-Specific Voice Output ">
    1. Below the Question Node, add a Lookup Node and configure the following settings:
       * **Type** — select **CognigyScript**.
       * **Operator** — enter the following:
       ```javascript theme={null}
       input.data.payload?.speech?.language_code
       ```
       This value verifies the language code for language recognition. Save the Node.
    2. Set the Case Nodes as follows:
       * Case Node 1:
         * **Value** — enter `de-DE`. For language recognition, use the full [BCP-47 format](https://datatracker.ietf.org/doc/html/rfc5646#appendix-A).
       * Case Node 2:
         * **Value** — enter `pt-PT`. For language recognition, use the full [BCP-47 format](https://datatracker.ietf.org/doc/html/rfc5646#appendix-A).
    3. Below each Case Node, add a Set Translation Node and configure as follows:
       1. Set Translation Node for German:
          * **User Input Language** — enter `de`. For translation, use the [ISO 639-1 language code](https://en.wikipedia.org/wiki/List_of_ISO_639_language_codes).
          * **Flow Language** — enter `en`.
          * *(Optional)* In the **Settings** section, enter a name to distinguish the Node in the **Label** field, for example, `EN/DE`. Save the Node.
       2. Set Translation Node for Portuguese:
          * **User Input Language** — enter `pt`. For translation, use the [ISO 639-1 language code](https://en.wikipedia.org/wiki/List_of_ISO_639_language_codes).
          * **Flow Language** — enter `en`.
          * *(Optional)* In the **Settings** section, enter a name to distinguish the Node in the **Label** field, for example, `EN/PTBR`. Save the Node.
    4. Below each Set Translation Node, add a [Set Session Config Node](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details) and configure the **Synthesizer (TTS)** settings as follows:
       1. Set Session Config Node for German:
          * **TTS Vendor** — select your speech provider.
          * **TTS Language** — select **German**.
          * **TTS Voice** — select a voice for your AI Agent to use when speaking in German.
       2. Set Session Config Node for Portuguese:
          * **TTS Vendor** — select your speech provider.
          * **TTS Language** — select **Portuguese** from the **TTS Language** list.
          * **TTS Voice** — select a voice for your AI Agent to use when speaking in Portuguese.
  </Step>

  <Step title="(Optional) Add an LLM Prompt Node for a Human-Like Handover Message">
    1. Below the Lookup Node, add an LLM Prompt Node and configure the **System Prompt** field as:

    ```text theme={null}
    Based on the user's message, reply kindly and advise you are handing over to a human agent.
    ```

    Save the Node.
  </Step>

  <Step title="Configure Handover">
    Below the Lookup or the LLM Prompt Node, add a Handover to Human Agent Node and configure the following:

    * **Handover Settings** section, select your [handover provider](/ai/escalate/handover-reference/overview) from the **Handover Provider** list. Depending on the handover provider, you need to configure additional settings.
  </Step>
</Steps>

### Configure the Agent Copilot Flow

<Steps>
  <Step title="Configure the Agent Copilot Grid">
    1. In **Build > Flows**, create a Flow and give it a clear name, for example, `Real-Time Translation Copilot Flow`.
    2. Add a [Copilot: Set Grid Node](/ai/agents/develop/node-reference/agent-copilot/set-grid) and set the **Copilot Grid Configuration** as follows:

    <Accordion title="Agent Copilot Grid JSON">
      ```json theme={null}
      {
        "grid": {
          "columns": 2,
          "rows": 2,
          "gap": 5
        },
        "tiles": {
          "next-action": {
            "x": 1,
            "y": 1,
            "rows": 1,
            "columns": 1
          },
          "transcript": {
            "x": 1,
            "y": 2,
            "rows": 1,
            "columns": 1
          }
        }
      }
      ```
    </Accordion>

    3. Below the Copilot: Set Grid Node, add an [LLM Prompt Node](/ai/agents/develop/node-reference/service/llm-prompt) and configure it to generate the next best response for the agent based on the user's latest input and to store the result in the Input object:

       * **System Prompt** — enter the following:
         ```txt theme={null}
         Based on the user's last message, suggest the most helpful next response for a human agent.
         ```
       * In the **Storage & Streaming** section, configure the following:
         * **How to handle the result** — set to **Store in Input**.
         * **Input Key to store Result** — set to `input.promptResult`.

    4. Add a [Copilot: Next Action Tile Node](/ai/agents/develop/node-reference/agent-copilot/next-action-tile) after the LLM Prompt Node and configure the Node as follows:
       * **Tile ID**: `next-action`
       * **Text**:
       ```txt theme={null}
       {{input.promptResult.detailedResult.choices[0].message.content}}
       ```

    5. Add a [Copilot: Transcript Tile Node](/ai/agents/develop/node-reference/agent-copilot/transcript-tile) and set **Tile ID** to `transcript`.
  </Step>

  <Step title="Create a Voice Copilot Endpoint">
    1. In **Deploy > Endpoints**, click **+ New Endpoint** and select **Voice Copilot**.
    2. In the **New Endpoint** section, configure the following:
       * **Name** — enter a unique name, for example, `Real-Time Translation Copilot Endpoint`.
       * **Flow** — select the Agent Copilot Flow you created from the list.
  </Step>
</Steps>

With this grid, the Agent Copilot workspace displays a widget with the conversation transcript and a widget with an LLM-generated next action that is translated to the user's language when sent to the user. You can test the Real-Time Translation with [Agent Copilot for voice-to-chat applications](/agent-copilot/getting-started/voice-to-chat).

## More Information

* [Agent Copilot for Voice-to-Chat](/agent-copilot/getting-started/voice-to-chat)
* [Set Translation Node](/ai/agents/develop/node-reference/other-nodes/set-translation)


# Agent Copilot Nodes
Source: https://docs.cognigy.com/agent-copilot/configure/nodes

Control the Agent Copilot workspace from your Cognigy Flow using configuration, plug-and-play widget, and customization Nodes.

<a href="/release-notes/4.60"><Badge>Updated in 4.60</Badge></a>

Agent Copilot Nodes allow you to configure and control the Agent Copilot workspace from the Flow.

Agent Copilot Nodes include:

* **Configuration Node** — the [Copilot: Set Grid Node](/ai/agents/develop/node-reference/agent-copilot/set-grid) to configure and arrange the tiles in the Agent Copilot grid.
* **Plug-and-play widgets** — plug-and-play widgets are Agent Copilot Nodes that require minimal configuration to integrate into the Agent Copilot workspace.
* **Customization Nodes** — Nodes that can provide advanced capabilities to the widgets in the Agent Copilot workspace.

## Configuration Node

<CardGroup>
  <Card title="Copilot: Set Grid" href="/ai/agents/develop/node-reference/agent-copilot/set-grid">
    This Node allows you to configure and arrange the tiles in the Agent Copilot grid.
  </Card>
</CardGroup>

## Plug-and-Play Widgets

<CardGroup>
  <Card title="Copilot: Identity Tile" href="/ai/agents/develop/node-reference/agent-copilot/identity-tile">
    This Node displays and processes user-specific information directly in the widget of your Agent Copilot workspace.
  </Card>

  <Card title="Copilot: Next Action Tile" href="/ai/agents/develop/node-reference/agent-copilot/next-action-tile">
    This Node determines the appropriate next step or action in a conversation or process based on the user's input or the current context.
  </Card>

  <Card title="Copilot: Sentiment Tile" href="/ai/agents/develop/node-reference/agent-copilot/sentiment-tile">
    This Node performs sentiment analysis on the most recent input and provides insights into the emotional tone of the user.
  </Card>

  <Card title="Copilot: Transcript Tile" href="/ai/agents/develop/node-reference/agent-copilot/transcript-tile">
    This Node captures the most recent user answer and displays it on a designated widget in the Agent Copilot workspace.
  </Card>
</CardGroup>

## Customization Nodes

<CardGroup>
  <Card title="Copilot: Adaptive Card Tile" href="/ai/agents/develop/node-reference/agent-copilot/set-adaptive-card-tile">
    This Node allows you to create and display adaptive cards in your Agent Copilot workspace.
  </Card>

  <Card title="Copilot: HTML Tile" href="/ai/agents/develop/node-reference/agent-copilot/set-html-tile">
    This Node allows you to display customized HTML content in the Agent Copilot workspace.
  </Card>

  <Card title="Copilot: IFrame Tile" href="/ai/agents/develop/node-reference/agent-copilot/set-iframe-tile">
    This Node allows you to embed external websites directly into the Agent Copilot workspace.
  </Card>

  <Card title="Copilot: Knowledge Tile" href="/ai/agents/develop/node-reference/agent-copilot/knowledge-tile">
    This Node allows you to provide human agents with potential answers to user questions.
  </Card>
</CardGroup>

## More Information

* [Agent Copilot](/agent-copilot/overview)
* [Agent Copilot Configuration](/agent-copilot/configure/workspace)
* [Embedding Agent Copilot](/agent-copilot/deploy/embedding)


# Agent Copilot Configuration
Source: https://docs.cognigy.com/agent-copilot/configure/overview

Learn how to configure Agent Copilot with an AI Agent, Voice Gateway, and contact center integration.

To integrate Agent Copilot with an AI Agent, you need to configure different components:

* [Agent Copilot](#agent-copilot)
* [Voice Gateway](#voice-gateway), for voice implementations.
* [Contact center integration](#contact-center-integration)

## Agent Copilot

The Agent Copilot configuration includes:

<AccordionGroup>
  <Accordion title="Agent Copilot Flow">
    The Agent Copilot Flow is a Flow separate from the AI Agent Flow to which you add Agent Copilot Nodes. The Copilot: Set Grid Node is required in this Flow to set the workspace arrangement on the contact center side and the widgets in the workspace. Other Agent Copilot Nodes provide functionality to the widgets in the workspace.
  </Accordion>

  <Accordion title="Agent Copilot Grid">
    The Agent Copilot Grid is a JSON object you can configure in the [Copilot: Set Grid Node](/ai/agents/develop/node-reference/agent-copilot/set-grid) that defines the layout of the Agent Copilot workspace. For more information, see [Configure the Workspace](/agent-copilot/configure/workspace).
  </Accordion>

  <Accordion title="Agent Copilot URL">
    To view the Agent Copilot workspace in the contact center interface, configure the Agent Copilot URL. The workspace is rendered at this URL as soon as the handover is triggered.

    To configure the URL, use the following format:

    `https://${AgentCopilotBaseUrl}/?userId=${userId}&sessionId=${sessionId}&URLToken=${URLToken}`

    Where:

    * `AICopilotBaseUrl` — the base URL used to access the Agent Copilot workspace.
    * `sessionId` — the Cognigy session ID. A session represents the interaction between the human agent and the Agent Copilot workspace.
    * `userId` — the Cognigy user ID. It helps Agent Copilot associate the interaction with a specific user.
    * `URLToken` — the Endpoint token from Cognigy. The Endpoint must be linked to the Handover Flow. To find this token, go to the Endpoint that you created previously. In the **Endpoint URL** field, copy the token after `https://endpoint-trial.cognigy.ai/`. For example, in the URL `https://endpoint-trial.cognigy.ai/f38791ae20d4961acf0e97d9f377c4fe3df92894e1eff1c7a774a8ed089a4590`, the token is `f38791ae20d4961acf0e97d9f377c4fe3df92894e1eff1c7a774a8ed089a4590`. Only authorized users can access the Agent Copilot system through this URL.

    Example URL:

    ```txt wrap theme={null}
    https://ai-copilot-trial.cognigy.ai/?sessionId=session123&userId=user123&URLToken=a4d5c86c98f27730311591f28d194510e05ffed30ca148e3344970defd418e7d
    ```

    You need this URL for further steps in the [Deployment section](/agent-copilot/deploy/overview) where you choose how you want to deploy your Agent Copilot workspace.
  </Accordion>
</AccordionGroup>

## Voice Gateway

For voice implementations, you need to configure:

<AccordionGroup>
  <Accordion title="Voice Gateway Endpoint">
    Set up a [Voice Gateway Endpoint](/ai/agents/deploy/endpoint-reference/voice-gateway) and [application](/voice-gateway/webapp/applications) to connect the AI Agent to Voice Gateway, and Voice Gateway to the contact center.
  </Accordion>

  <Accordion title="Voice Copilot Endpoint">
    The [Voice Copilot Endpoint](/ai/agents/deploy/endpoint-reference/voice-copilot) receives the transcript of the voice call and sends it to the Agent Copilot Flow.
  </Accordion>
</AccordionGroup>

## Third-Party Platform Integration

You can integrate Agent Copilot with a third-party platform in the following ways:

<AccordionGroup>
  <Accordion title="Handover to Human Agents">
    For chat applications, you need to configure the handover to human agents by adding a [Handover to Human Agent Node](/ai/agents/develop/node-reference/service/handover-to-agent) to the AI Agent Flow and by using the [Copilot section](/agent-copilot/configure/copilot-section) in the Endpoint settings.
  </Accordion>

  <Accordion title="Call Transfer to Human Agents">
    To transfer calls from the AI Agent to human agents, you need to configure the call transfer by adding a [Transfer Node](/ai/agents/develop/node-reference/voice/voice-gateway/transfer) to the AI Agent Flow.
  </Accordion>
</AccordionGroup>

Both implementations require a [handover provider](/ai/escalate/handover-reference/overview) that supports the contact center platform you use.

## More Information

* [Agent Copilot Workspace](/agent-copilot/overview)
* [Getting started with Agent Copilot for Chat](/agent-copilot/getting-started/chat)
* [Getting started with Agent Copilot for Voice-to-Chat](/agent-copilot/getting-started/voice-to-chat)
* [Getting started with Agent Copilot for Voice](/agent-copilot/getting-started/voice)
* [Agent Copilot Nodes](/ai/agents/develop/node-reference/agent-copilot/overview)


# Configure the Workspace
Source: https://docs.cognigy.com/agent-copilot/configure/workspace

Configure the Agent Copilot workspace grid layout, widget size, and tile arrangement to display content in your contact center platform.

<a href="/release-notes/4.65"><Badge>Updated in 4.65</Badge></a>

The Agent Copilot workspace is the interface that displays the Agent Copilot on the contact center side. The workspace includes the widgets that represent the [Agent Copilot Nodes](/agent-copilot/configure/nodes) that you configure in the [Agent Copilot Flow](/agent-copilot/configure/overview). You can configure widget size and arrangement in the workspace, and display different types of content in it, including HTML, adaptive cards, or iframes.

To configure the Agent Copilot workspace, you need to set the Agent Copilot grid. Other Agent Copilot Nodes provide functionality to the widgets in the workspace.

## Agent Copilot Grid

The Agent Copilot grid is a JSON object that defines the size and arrangement of the tiles, through tile IDs, in the workspace.

The following table shows the JSON properties you use to configure the Agent Copilot grid:

<Accordion title="Agent Copilot Grid JSON Properties">
  | Property |             |           | Type   | Description                                                                                                                            |
  | -------- | ----------- | --------- | ------ | -------------------------------------------------------------------------------------------------------------------------------------- |
  | `grid`   |             |           | Object | Set the grid size and gap between tiles.                                                                                               |
  |          | `columns`   |           | Number | Set the number of columns in the grid. The number of columns corresponds to the grid's x-axis.                                         |
  |          | `rows`      |           | Number | Set the number of rows in the grid. The number of rows corresponds to the grid's y-axis.                                               |
  |          | `gap`       |           | Number | Set the gap size in pixels between the tiles. If you set the value to `0`, Agent Copilot uses the default value of `10`.               |
  | `tiles`  |             |           | Object | Set the tiles configuration.                                                                                                           |
  |          | `<tile ID>` |           | Object | Enter the [tile ID](#tile-id) and set the tile configuration. The tile ID must match the tile ID in the respective Agent Copilot Node. |
  |          |             | `x`       | Number | Set the coordinate of the tile in the grid's x-axis.                                                                                   |
  |          |             | `y`       | Number | Set the coordinate of the tile in the grid's y-axis.                                                                                   |
  |          |             | `rows`    | Number | Set the number of rows the tile takes up.                                                                                              |
  |          |             | `columns` | Number | Set the number of columns the tile takes up.                                                                                           |
</Accordion>

To configure the Agent Copilot grid, you have the following options:

<Tabs>
  <Tab title="GUI">
    Use the [Copilot: Set Grid Node](/ai/agents/develop/node-reference/agent-copilot/set-grid) to configure the grid directly in the Agent Copilot Flow:

    1. In **Build > Flows**, create an [Agent Copilot Flow](/agent-copilot/configure/overview#agent-copilot)
    2. Add a [Copilot: Set Grid Node](/ai/agents/develop/node-reference/agent-copilot/set-grid) after the **Start** Node. You can use the following example as a reference:

    <Accordion title="Agent Copilot Grid Example">
      ```json theme={null}
      "grid": {
          "columns": 3,
          "rows": 8,
          "gap": 10
      },
      "tiles": {
          "identity-assist": {
            "x": 1,
            "y": 1,
            "rows": 4,
            "columns": 1
          },
          "transcript-assist": {
            "x": 2,
            "y": 1,
            "rows": 2,
            "columns": 2
          },
          "next-action": {
            "x": 2,
            "y": 3,
            "rows": 2,
            "columns": 2
          },
          "knowledge-assist": {
            "x": 2,
            "y": 5,
            "rows": 4,
            "columns": 2
          },
          "event-assist": {
            "x": 1,
            "y": 5,
            "rows": 4,
            "columns": 1
          }
      }
      ```
    </Accordion>
  </Tab>

  <Tab title="API">
    Use the [Cognigy API](https://api-trial.cognigy.ai/openapi) to create a configuration for the grid through an API request. You'll need to select the configuration in the Endpoint settings. This grid configuration is used only if there's no Copilot: Set Grid Node in the Agent Copilot Flow.

    You can configure the grid via API and implement it by selecting it in the **Copilot** section in the [Endpoint settings](/ai/agents/deploy/endpoints/overview).

    To configure the grid via API, follow these steps:

    <Steps>
      <Step title="Create an API Key">
        1. On the [Main page](/ai/overview/user-interface/main-page), click **<img alt="user menu" /> > My Profile** in the upper-right corner.
        2. In the **API Keys** section, click **+**.
        3. Enter a unique name for the API key and click **Confirm**. The API key will appear in the list. Copy the API key for later use.
      </Step>

      <Step title="Copy the Project ID">
        On the Main page, click <img alt="menu" />**> Copy Project ID**. Save the copied Project ID for later use.
      </Step>

      <Step title="Authenticate with the API Key">
        1. In the [Cognigy API](/ai/for-developers/developers/api-and-cli), go to the **AUTHENTICATION** section and paste the API key you copied earlier in the following fields:
           * **API Key (X-API-Key)**
           * **API Key (api\_key)**
        2. Click **Set** next to both fields.
      </Step>

      <Step title="Create the Grid with a POST Request">
        1. Go to the [POST](https://api-trial.cognigy.ai/openapi#post-/v2.0/aicopilotconfigs) request and configure the JSON payload:
           * `name` — enter a unique name for the Agent Copilot configuration. This name is displayed in the [Endpoint settings](/ai/agents/deploy/endpoints/overview).
           * *(Optional)* `description` — enter a description for the Agent Copilot configuration.
           * `config` — configure the [Agent Copilot grid](#agent-copilot-grid).
        2. Paste the Project ID you copied earlier into the `projectId` field in the JSON body of the request.
        3. Click **Try** to send the POST request with the JSON payload. If your request is successful, you receive the `201` status code in the response body and the Agent Copilot configuration is added to the **Copilot Config** list in the Endpoint settings.

        You can use the following example as a reference:

        <Accordion title="Agent Copilot Grid Example">
          ```json theme={null}
          {
              "name": "Agent Copilot Grid Example",
              "description": "A sample configuration for Agent Copilot showcasing the layout of tiles",
              "config": {
                "grid": {
                    "columns": 3,
                    "rows": 8,
                    "gap": 10
                },
                "tiles": {
                    "identity-assist": {
                      "x": 1,
                      "y": 1,
                      "rows": 4,
                      "columns": 1
                    },
                    "transcript-assist": {
                      "x": 2,
                      "y": 1,
                      "rows": 2,
                      "columns": 2
                    },
                    "next-action": {
                      "x": 2,
                      "y": 3,
                      "rows": 2,
                      "columns": 2
                    },
                    "knowledge-assist": {
                      "x": 2,
                      "y": 5,
                      "rows": 4,
                      "columns": 2
                    },
                    "event-assist": {
                      "x": 1,
                      "y": 5,
                      "rows": 4,
                      "columns": 1
                    }
                }
              },
              "projectId": "your-project-id"
          }
          ```
        </Accordion>
      </Step>
    </Steps>
  </Tab>
</Tabs>

## Tile ID

Each tile must have a unique tile ID, for example, `transcript-tile`, because the tile ID corresponds to a JSON key. You need to use the tile ID in the respective Agent Copilot Node to map it correctly to the tile in the workspace.

If Agent Copilot receives updates for a tile ID that isn't in the Agent Copilot configuration, the tile isn't rendered. Additionally, a warning indicator is displayed in the lower-right corner of the workspace, indicating that an unknown tile update has been received. In these cases, check the tile IDs in your Agent Copilot grid.

## Transcript Tile

The Transcript Tile acts as a chat interface and allows human agents to read messages from end users and reply to them directly in the workspace. The tile ID of the Transcript Tile is `transcript` and must be included in your Agent Copilot configuration. This tile is intended for [chat use cases](/agent-copilot/getting-started/chat) in [standalone Agent Copilot applications](/agent-copilot/deploy/standalone). For voice-related scenarios, use the [Copilot: Transcript Tile Node](/ai/agents/develop/node-reference/agent-copilot/transcript-tile).

To activate the Transcript Tile, go to the **Copilot** section in the Endpoint settings and use the settings in **Transcript Tile**:

| Setting                                                                                            | Description                                                                                                                                                                                           |
| -------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Enable Transcript Tile with ID 'transcript' in the Agent Copilot workspace                         | Activates the Transcript Tile.                                                                                                                                                                        |
| Enable the chat input for the Transcript Tile to enable the agent to send messages to the customer | Allows human agents to send messages to the end user from the Transcript Tile. The following handover providers don't support this setting: Genesys Cloud Open Messaging, NiCE CXone, storm, and 8x8. |
| Enable redaction of Transcript Tile messages                                                       | Redacts credit card numbers in messages in the Transcript Tile and replaces them with the `CREDIT_CARD_NUMBER` placeholder.                                                                           |


# Agent Copilot Custom Integration
Source: https://docs.cognigy.com/agent-copilot/deploy/custom-integration



This guide shows how to embed Agent Copilot into a custom contact center or agent desktop using the Voice Copilot Endpoint. It helps developers test and validate Copilot quickly without a full contact-center setup. Works for both chat and voice.

## Prerequisites

* Configured the Voice Gateway integration:
  * If a Voice Copilot Endpoint is used for voice or chat use cases.
  * If a Webhook Endpoint is used for voice use cases.

## Restrictions

* If you use a Webhook Endpoint, you will be charged for two conversations: one with the webhook and one with Agent Copilot.

## Create an Endpoint

Mock a request via Postman or curl to validate the webhook accepts your session.

<Tabs>
  <Tab title="Voice Copilot">
    1. In the left-side menu of your Project, click **Deploy > Endpoints**.
    2. On the **Endpoints** page, click **+ New Endpoint**.
    3. In the **New Endpoint** section, do the following:
       1. Select the **Voice Copilot** Endpoint type.
       2. Specify a unique name.
       3. Select a Flow from the list. Save changes.
    4. Go to the **Configuration Information** section and copy the URL from the **Endpoint URL** field.
  </Tab>

  <Tab title="Webhook">
    1. In the left-side menu of your Project, click **Deploy > Endpoints**.
    2. On the **Endpoints** page, click **+ New Endpoint**.
    3. In the **New Endpoint** section, do the following:
       1. Select the **Webhook** Endpoint type.
       2. Specify a unique name.
       3. Select a Flow from the list. Save changes.
    4. Go to the **Configuration Information** section and copy the URL from the **Endpoint URL** field.
  </Tab>
</Tabs>

## Mock a Request

Send a `POST` request to the Endpoint which you configured in the previous step.
Your web server should accept `POST` requests and process the JSON payload sent by Cognigy. For testing purposes, you can use [webhook.site](https://webhook.site) as a temporary web server.

<Tabs>
  <Tab title="cURL">
    Replace `https://<your-endpoint-url>` with the Endpoint URL from your Endpoint's settings.

    ```bash theme={null}
    curl -X POST https://<your-endpoint-url> \
      -H "Content-Type: application/json" \
      -d '{
        "userId": "user123",
        "sessionId": "session123",
        "text": "Hello, I need help with my order",
        "data": {
          "exampleKey": "exampleValue"
        }
      }'
    ```
  </Tab>

  <Tab title="Postman">
    1. Open the Postman collection, select **Add a request**, then set the request type to `POST`.
    2. Enter the Endpoint URL as the request URL.
    3. Go to the **Headers** tab and add `Content-Type: application/json`.
    4. Go to the **Body** tab, select **raw**, then select **JSON** as the format.
    5. Paste the request body:

    ```json theme={null}
    {
      "userId": "user123",
      "sessionId": "session123",
      "text": "Hello, I need help with my order",
      "data": {
        "exampleKey": "exampleValue"
      }
    }
    ```

    | Parameter | Type   | Description                                                                                                                                                                                                 | Required                                             |
    | --------- | ------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------- |
    | userId    | String | The ID of the end user.                                                                                                                                                                                     | Yes                                                  |
    | sessionId | String | The ID used to track the current session and maintain its state. Generate a new unique ID for each new session. For testing, you can use any string and change it whenever you want to start a new session. | Yes                                                  |
    | text      | String | The text that the assigned Flow should process.                                                                                                                                                             | No, if `data` is specified<sup>[1](#footnote1)</sup> |
    | data      | Object | The data that the assigned Flow should process.                                                                                                                                                             | No, if `text` is specified<sup>[1](#footnote1)</sup> |
  </Tab>
</Tabs>

## Configure the Agent Copilot URL

When setting up Agent Copilot, you can choose between embedded or standalone Agent Copilot versions.

For standalone Agent Copilot, use the URL format to access the workspace:

```txt theme={null}
https://${AgentCopilotBaseUrl}/?userId=${userId}&sessionId=${sessionId}&URLToken=${URLToken}
```

For the configuration details, refer to the [Agent Copilot URL](/agent-copilot/configure/overview#agent-copilot-url) section.

### (Optional) Embed in the Agent Desktop

You can embed Agent Copilot into any contact center interface using an iframe. Some platforms may require additional setup, such as enabling scripts or granting API access.

Embedding is specific to the contact center, so verify the integration on the provider's side to ensure proper functionality.

```html theme={null}
<iframe
  src="https://${AgentCopilotBaseUrl}/?userId=${userId}&sessionId=${sessionId}&URLToken=${URLToken}"
  width="100%"
  height="720"
  style="border:0;">
</iframe>
```

## More Information

* [Overview](/agent-copilot/deploy/overview)

***

<sup>1</sup>: You must provide at least one of `text` or `data`. You can send either, or both. If both are missing or invalid, the REST Endpoint throws an error.


# Embedded Agent Copilot
Source: https://docs.cognigy.com/agent-copilot/deploy/embedding



<a href="/release-notes/2026.15"><Badge>Updated in 2026.15</Badge></a>

You can embed Agent Copilot directly into your contact center interface.

In this mode, Copilot runs inside the agent workspace, allowing agents to interact with AI assistance without leaving their primary contact center application. This option is useful if you want to provide a seamless in-app experience and reduce context switching for agents.

All features remain available, including real-time guidance, transcript sharing, and support for both chat and voice use cases.

## How to Use

1. Make sure you configured an Agent Copilot URL using the following pattern: `https://${AgentCopilotBaseUrl}/?userId=${userId}&sessionId=${sessionId}&URLToken=${URLToken}`. For the configuration details, refer to the [Agent Copilot URL](/agent-copilot/configure/overview#agent-copilot-url) section.
2. Pass this URL along with the handover to the human agent:

   * For native integrations, the embedded version is supported automatically.
   * For custom integrations, include the link in your handover configuration and embed it via **iFrame**:

   ```html theme={null}
   <iframe src="https://copilot.example.com/?userId=12345&sessionId=abcde-67890&URLToken=f38791ae20d4961acf0e97d9f377c4fe3df92894e1eff1c7a774a8ed089a4590"
           width="100%"
           height="800px"
           frameborder="0">
   </iframe>
   ```
3. If you are using a **voice use case**, check the **Transfer Node** in your Flow and ensure the **UUIValue** is forwarded via SIP headers during the transfer to the human agent.

### Secure Access with JWT

You can secure  access to Agent Copilot by passing a JWT in the URL. This approach ensures that only authenticated users can open the Agent Copilot workspace.

```txt theme={null}
https://${AgentCopilotBaseUrl}/?URLToken=${URLToken}&auth={JWT}
```

Where `auth` is the JWT used for authentication.

In the **Copilot** section of the Endpoint settings, you can control how the JWT is verified with the **Authentication** and **Accept Only JWT Requests**. The following methods support JWT authentication:

* **Default (built-in)** — Agent Copilot handles the JWT authentication between the contact center platform and Cognigy. Activate **Accept Only JWT Requests** in the [Endpoint settings](/agent-copilot/configure/copilot-section#parameters).
* **Token Secret** — the contact center and Cognigy share a 256-bit JWT secret. Configure it by selecting or creating a connection in the **Agent Copilot Authentication** field. The contact center platform must sign the JWT with the shared secret before passing it in the URL.
* **Public Key** — enter one or more PEM-encoded public keys into the **Public Keys** field. The JWT must be signed with the corresponding private key. Multiple keys are supported to allow for key rotation.
* **Key Store** — provide a JWKS URL in the **Key Store URL** field. The Agent Copilot backend retrieves the public key from the key store using the `kid` claim in the JWT header.

The contact center integration must proceed as follows:

1. Creates values for `userId` and `sessionId`.
2. Uses `URLToken` from the respective Cognigy Endpoint.
3. Reads the signing key from an internal storage.
4. Signs a JWT with a payload containing the `userId` and `sessionId`.

The Agent Copilot backend proceeds as follows:

1. Checks if the `auth` parameter is passed in the URL.
2. If `auth` is passed, the value is treated as a JWT and the Agent Copilot backend verifies the JWT signature using the configured authentication method.
3. During verification, Agent Copilot decodes the payload of the JWT to retrieve the `userId` and `sessionId` parameters, and reads `URLToken` from the URL query string.

For more information about JWTs, read the [JWT documentation](https://www.jwt.io/introduction).

## More Information

* [Agent Copilot](/agent-copilot/overview)
* [Getting started with Agent Copilot](/agent-copilot/getting-started/chat)
* [Agent Copilot Configuration](/agent-copilot/configure/overview)
* [Agent Copilot Nodes](/ai/agents/develop/node-reference/agent-copilot/overview)
* [Integration with Contact Centers](/agent-copilot/configure/overview)


# NiCE CXone AAH
Source: https://docs.cognigy.com/agent-copilot/deploy/nice-aah



<a href="/release-notes/2026.4"><Badge>Added in 2026.4</Badge></a>

<Frame>
  <img alt="NiCE CXone AAH Endpoint configuration panel in Cognigy" />
</Frame>

The NiCE CXone AAH Endpoint enables you to deploy [Agent Copilot](/agent-copilot/overview) directly through [NiCE CXone Agent Assist Hub (AAH)](https://help.nicecxone.com/content/aiassistantsandbots/agentassisthub/cognigyagentassist/cognigyagentassist.htm). This deployment mode lets you use Agent Copilot in the NiCE CXone Mpower contact center without a Cognigy Agent.

## Prerequisites

* An account on the [NiCE CXone](https://cxone.niceincontact.com/) platform.
* Access to the [CXone Mpower Agent](https://help.nicecxone.com/content/agent/agentapplicationadministration/cxoneagent/cxoneagentforadministrators.htm) in NiCE CXone AAH.
* Install the [NiCE CXone Agent WebRTC extension](https://help.nicecxone.com/content/agent/cxoneagent/addcxawebrtcext.htm).
* Generate a [Cognigy API key](/ai/administer/user-menu/my-profile#api-keys).
* Configure an [Agent Copilot Flow](/agent-copilot/configure/overview#agent-copilot).
* A 256-bit string secret to sign a JSON Web Token (JWT). Both [Cognigy](#setup-on-the-cognigy-ai-side) and [NiCE CXone](#setup-on-the-nice-cxone-side) use the same secret to authenticate via JWT and establish a secure connection. For more information, read the [JWT documentation](https://www.jwt.io/introduction).

## Limitations

* When the end user hangs up, the Agent Copilot window on the NiCE CXone AAH platform closes, and the human agent can no longer access it.

## General Endpoint Settings

| Setting         | Description                                                                                                                                                                                                                                                                                         |
| --------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Flow            | Select the [Agent Copilot Flow](/agent-copilot/configure/overview#agent-copilot) to use for the Endpoint.<br /><br />You can enter manually by clicking **Manual Input** below the Flow list. Click <img alt="open flow" /> next to the **Flow** list to open the selected Flow in the Flow editor. |
| Endpoint URL    | The Endpoint URL serves as the interface to the Endpoint. You need this URL to configure the AAH in NiCE CXone.                                                                                                                                                                                     |
| Enable Endpoint | Toggle to activate or deactivate the Endpoint. This setting is enabled by default.                                                                                                                                                                                                                  |

## Specific Endpoint Settings

| Parameter                     | Description                                                                                                                                                                                                                                                                                                                                                                          |
| ----------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| NiCE CXone AAH Authentication | Select a NiCE CXone AAH connection from the list.<br /> Alternatively, click <img alt="add NiCE CXone AAH connection" /> next to the list to configure a connection to NiCE CXone AAH by filling in the following fields:<ul><li>**Connection name** — enter a unique name.</li><li>**API Key** — enter a [Cognigy API key](/ai/administer/user-menu/my-profile#api-keys).</li></ul> |

## How to Set Up

### Setup on the Cognigy Side

<Accordion title="Configure a NiCE CXone AAH Endpoint">
  1. Go to **Deploy > Endpoints**.
  2. On the **Endpoints** page, click **+ New Endpoint**.
  3. In the **New Endpoint** section, configure the following:
     1. Select the **NiCE CXone AAH** Endpoint type.
     2. Add a unique name.
     3. Select a relevant Agent Copilot Flow from the list. Save changes.
  4. In the **Configuration Information** section, copy the Endpoint URL and save it for later use in NiCE CXone. This URL is composed of two parts, the server address and the URL token. For example, if the Endpoint URL is `https://endpoint-trial.cognigy.ai/69ad2e3f9a1bc351eac7aa234cee41c2967d7b0b9dbbdba6b5711357bdgaa1a8`, then:
     * The server address is `https://endpoint-trial.cognigy.ai/`.
     * The URL token is `69ad2e3f9a1bc351eac7aa234cee41c2967d7b0b9dbbdba6b5711357bdgaa1a8`.
  5. Activate the **Enable Endpoint** setting.
  6. In the **NiCE CXone AAH Setup** section, click <img alt="add NiCE CXone AAH connection" /> next to the **NiCE CXone AAH Authentication** field and add a NiCE CXone AAH connection by filling in the following fields:
     * **Connection name** — enter a unique name.
     * **API Key** — enter a [Cognigy API key](#prerequisites).
  7. In the **Copilot** section, configure the following:
     * **Copilot Type** — select **Workspace**.
  8. Toggle on the **Enable Copilot Authentication** option.
  9. Click <img alt="add JWT connection" /> next to the **Agent Copilot Authentication** field to configure the JWT connection:
     * **Connection name** — enter a unique name.
     * **JWT Secret** — enter the [string](#prerequisites) to be used as the JWT secret.
  10. *(Optional)* Activate the **Accept Only JWT Requests** option to block any requests to the Agent Copilot URL that don't contain a valid JWT. This parameter is deactivated by default. Activating this parameter is useful for security reasons because it ensures that access requests that include only the user and session IDs are blocked.
</Accordion>

### Setup on the NiCE CXone Side

<AccordionGroup>
  <Accordion title="1. Create an Agent Assist App in NiCE CXone">
    1. On the [NiCE CXone](https://cxone.niceincontact.com/) platform, click the **Applications** selector in the upper-left corner.
    2. Go to **Others > Automation & AI**. The **Automation & AI** window opens.
    3. In the left-side **CXone Mpower Hubs** menu, select **Agent Assist Hub**. The **Agent Assist Hub** panel appears. Click **Add Agent Assist App**.
    4. At **Step 1: General Information**, configure the following:
       * **Agent Assist App Name** — enter a unique name. Copy and save the name for later use in NiCE CXone.
       * **Select Agent Assist App Type** — select **Cognigy Assist**.
    5. At **Step 2: Configuration**, configure the following:
       * **Endpoint URL** — enter the server address from the Endpoint URL that you saved previously, for example, `https://endpoint-trial.cognigy.ai/`.
       * **Agent Assist URL** — enter the base URL of Agent Copilot, for example, `https://agent-copilot-trial.cognigy.ai/`.
       * **URL Token** — enter the URL token from the Endpoint URL that you saved previously, for example, `69ad2e3f9a1bc351eac7aa234cee41c2967d7b0b9dbbdba6b5711357bdgaa1a8`.
       * **API Key** — enter the API key generated in NiCE CXone.
       * **JWT Secret Key** — enter the [string](#prerequisites) to be used as the JWT secret.
    6. Click **Create**.
  </Accordion>

  <Accordion title="2. Create a Transcription Profile">
    1. In the **CXone Mpower Hubs** menu, select **Transcription Hub > Continuous Stream** and click **Add Transcription Profile** and configure the following:
       1. At **Step 1: General Information**, configure the following:
          * **Transcription Profile Name** — enter a unique name. Copy and save the name for later use in NiCE CXone.
          * **Select Provider** — select your transcription provider.
       2. At **Step 2: Configuration**, configure according to the selected provider.
    2. Click **Create** and close the **Automation & AI** window.
  </Accordion>

  <Accordion title="3. Create a Campaign">
    1. On the [NiCE CXone](https://cxone.niceincontact.com/) platform, click the **Applications** selector in the upper-left corner and select **ACD**.
    2. In the left-side menu, select **Campaigns**.
    3. In the upper-right corner, click **+ Create Campaign** and give the campaign a unique name.
  </Accordion>

  <Accordion title="4. Create a Script">
    1. On the [NiCE CXone](https://cxone.niceincontact.com/) platform, click the **Applications** selector in the upper-left corner and select **Studio**.
    2. In the upper-right corner, click **New Script** and select **Phone**.
    3. In the left-side menu, click actions and type `Begin` in the **Search actions** field.
    4. Drag the **Begin** action onto the canvas.
    5. Hover over the **Begin** action and click the circle on the right. The **Choose Action** widget opens.
    6. Type `Agent Assist` in the **Search actions** field and select the **Agent Assist** action from the list. The **Agent Assist** is set on the canvas.
    7. Click the **Agent Assist** action and paste the Agent Assist App Name that you saved previously in the **Assist Launch Config Name** field.
    8. Hover over the **Agent Assist** action and click the circle on the right. The **Choose Action** widget opens.
    9. Type `Continuous Transcription` in the **Search actions** field and select the **Continuous Transcription** action from the list. The **Continuous Transcription** action is set on the canvas.
    10. Click the **Continuous Transcription** action and configure the following:
        * **TranscriptionProfileName** — paste the Transcription Profile Name that you saved previously.
        * **Enabled** — enter `true`.
        * **Participants** — enter `Both`.
        * **GenerateUtteranceEvents** — enter `false`.
    11. In the upper-right corner, click **Create**.
    12. Enter a unique name in the **Name** field and click **Create**.
  </Accordion>

  <Accordion title="5. Create an ACD Skill">
    1. In the left-side menu of the ACD application, select **ACD Skills**.
    2. Click **+ Create New** and select **Single Skill**.
    3. In the **Name & Other Details** section, configure the following:
       * **Media Type** — select **Phone Call**.
       * **Skill Name** — enter a unique name.
       * **Inbound/Outbound** — select **Outbound**.
       * **Campaign** — select the campaign you created previously.
       * **Custom Script** — select the script you created previously.
    4. In the upper-right corner, click **Create**.
  </Accordion>

  <Accordion title="6. Assign the Skill to the ACD User">
    1. In the left-side menu of the ACD application, select **ACD Users**.
    2. Select the ACD user you want to assign the skill to from the list. You can type their name in the **Search** field to find them.
    3. Go to the **Skills** tab, select the skill you previously created from the **Add Skills** list and click **Add Skills**. The skill is moved to the **Assigned Skills** list.
    4. In the **Proficiency** column of the skill you assigned, select **1 - Highest** from the list. Click **Save Proficiencies**.
  </Accordion>
</AccordionGroup>

## Test the Endpoint

After you've set up the Endpoint, you can test Agent Copilot by clicking the **Screen Pops** icon and making a call through the [NiCE CXone Agent interface](https://help.nicecxone.com/content/agent/cxoneagent/cxoneagent.htm?tocpath=Agents%7CAgent%20%7C_____0#TheAgentInterface).

## More Information

* [Cognigy in NiCE CXone](https://help.nicecxone.com/content/aiassistantsandbots/agentassisthub/cognigyagentassist/cognigyagentassist.htm?)
* [Integrate Cognigy with NiCE CXone](https://help.nice-incontact.com/content/aiassistantsandbots/virtualagenthub/integratecognigy.htm?tocpath=Virtual%20Agent%20Hub%7CVirtual%20Agent%20Hub%7CVirtual%20Agent%20Integrations%7CCognigy%7C_____1)
* [NiCE CXone Studio](https://help.nicecxone.com/content/studio/gettingstarted/startusingstudiocx.htm?Highlight=studio)
* [NiCE CXone ACD](https://help.nicecxone.com/content/acd/welcometoacd.htm)
* [NiCE CXone Automation and AI](https://help.nicecxone.com/content/aiassistantsandbots/aiassistantsandbots.htm)
* [NiCE CXone Agent](https://help.nicecxone.com/content/agent/cxoneagent/cxoneagent.htm)


# Deploy and Integrate Agent Copilot
Source: https://docs.cognigy.com/agent-copilot/deploy/overview

Deploy Agent Copilot as standalone or embedded, with integrations for Genesys, Salesforce, and NiCE CXone.

The Agent Copilot workspace is a web-based application that provides customer service agents with AI-powered assistance during their interactions with customers. It can be deployed in various ways to suit your contact center's architecture and requirements.

## Deployment Options

The Agent Copilot workspace is designed for flexible integration into any contact center environment. Depending on your needs, it can be deployed in two primary ways:

* **Standalone Deployment:** The Agent Copilot workspace can be used as a standalone application, accessible via a web browser.
* **Embedded Deployment:** The workspace can be embedded directly into an existing third-party application, such as a contact center agent desktop. This provides a seamless user experience for agents.

## Integrations

### Native Integrations

Cognigy also provides out-of-the-box native integrations for the following contact centers, simplifying the setup process and ensuring a streamlined connection.

* **Genesys Cloud**: [Genesys Cloud Open Messaging](/ai/escalate/handover-reference/genesys-cloud-open-messaging)
* **Salesforce**: [Salesforce Service Cloud](/ai/escalate/handover-reference/salesforce) and [Salesforce MIAW](/ai/escalate/handover-reference/salesforce-miaw)
* **NiCE CXone**: [NiCE CXone AAH](/agent-copilot/deploy/nice-aah)

### Custom Integration

For all contact centers, the Agent Copilot workspace can be integrated via a custom integration. This approach gives you full control over how the workspace is configured and deployed within your specific environment.


# Standalone Agent Copilot
Source: https://docs.cognigy.com/agent-copilot/deploy/standalone



<a href="/release-notes/2026.15"><Badge>Updated in 2026.15</Badge></a>

You can deploy Agent Copilot as a standalone app.

In this mode, Copilot is accessed via a dedicated URL and can be opened on a separate browser tab instead of being embedded into a contact center interface. This option is useful if your contact center doesn't support embedding, or if human agents prefer working with Agent Copilot in a full-screen view.

All features remain available, including real-time guidance, transcript sharing, and support for both chat and voice use cases.

## How to Use

1. Make sure you configured an Agent Copilot URL using the following pattern: `https://${AgentCopilotBaseUrl}/?userId=${userId}&sessionId=${sessionId}&URLToken=${URLToken}`. For the configuration details, refer to the [Agent Copilot URL](/agent-copilot/configure/overview#agent-copilot-url) section.
2. Pass this URL along with the handover to the human agent.
   * For native integrations, the link is passed automatically.
   * For custom integrations, include the link in your handover configuration.
3. If you work with a voice use case, check the Transfer Node in your Flow.
4. Configure the Transfer Node:
   1. Fill in the **Transcription Webhook** field with your Agent Copilot URL.
   2. Activate **Enable Copilot**.
   3. Specify the header that carries the UUIValue in the **Copilot Headers Key** field. The default is `X-Custom-Header`. This header will be sent to the Agent Copilot transcription webhook, while the UUIValue is also forwarded via SIP headers when transferring the call to the human agent.

### Secure Access with JWT

You can secure  access to Agent Copilot by passing a JWT in the URL. This approach ensures that only authenticated users can open the Agent Copilot workspace.

```txt theme={null}
https://${AgentCopilotBaseUrl}/?URLToken=${URLToken}&auth={JWT}
```

Where `auth` is the JWT used for authentication.

In the **Copilot** section of the Endpoint settings, you can control how the JWT is verified with the **Authentication** and **Accept Only JWT Requests**. The following methods support JWT authentication:

* **Default (built-in)** — Agent Copilot handles the JWT authentication between the contact center platform and Cognigy. Activate **Accept Only JWT Requests** in the [Endpoint settings](/agent-copilot/configure/copilot-section#parameters).
* **Token Secret** — the contact center and Cognigy share a 256-bit JWT secret. Configure it by selecting or creating a connection in the **Agent Copilot Authentication** field. The contact center platform must sign the JWT with the shared secret before passing it in the URL.
* **Public Key** — enter one or more PEM-encoded public keys into the **Public Keys** field. The JWT must be signed with the corresponding private key. Multiple keys are supported to allow for key rotation.
* **Key Store** — provide a JWKS URL in the **Key Store URL** field. The Agent Copilot backend retrieves the public key from the key store using the `kid` claim in the JWT header.

The contact center integration must proceed as follows:

1. Creates values for `userId` and `sessionId`.
2. Uses `URLToken` from the respective Cognigy Endpoint.
3. Reads the signing key from an internal storage.
4. Signs a JWT with a payload containing the `userId` and `sessionId`.

The Agent Copilot backend proceeds as follows:

1. Checks if the `auth` parameter is passed in the URL.
2. If `auth` is passed, the value is treated as a JWT and the Agent Copilot backend verifies the JWT signature using the configured authentication method.
3. During verification, Agent Copilot decodes the payload of the JWT to retrieve the `userId` and `sessionId` parameters, and reads `URLToken` from the URL query string.

For more information about JWTs, read the [JWT documentation](https://www.jwt.io/introduction).


# Getting Started with Agent Copilot for Chat
Source: https://docs.cognigy.com/agent-copilot/getting-started/chat

Get started with Agent Copilot for chat: configure your first workspace with widgets that suggest next actions to human agents in your contact center.

In this guide, you will learn how to configure your first embedded Agent Copilot workspace for the chat AI Agent. This configuration provides an Agent Copilot workspace that includes two widgets that suggest next actions to the human agent.

## Prerequisites

* Configure a contact center platform that you can integrate with a [handover provider](/ai/escalate/handover-reference/overview). For example, for [Live Agent](/ai/escalate/handover-reference/live-agent), you need to have an [Inbox](/live-agent/settings/inboxes) with the [Project ID and the Webhook URL set up](/live-agent/settings/inboxes#edit-inboxes) and, if necessary, the **Enable Agent Copilot Workspace** option activated.
* Add a [compatible LLM](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) for the LLM Prompt and Copilot: Sentiment Tile Nodes and configure the LLM in the [Generative AI settings](/ai/agents/develop/gen-ai-and-llms/generative-ai).

## Setup Process

The setup process includes the following steps:

1. Create a handover provider. The handover provider connects the AI Agent and Agent Copilot to your contact center.
2. Create a Handover Flow. The Handover Flow controls the interaction with the AI Agent and the handover to a human agent.
3. Create an Agent Copilot Flow. The Agent Copilot Flow controls the Agent Copilot workspace and the individual widgets in it.
4. Test the Agent Copilot workspace through a Webchat v3 Endpoint.

## Set Up Agent Copilot for Chat

<Steps>
  <Step title="Create a Handover Provider">
    In **Deploy > Handover Providers**, create a [Handover Provider](/ai/escalate/handovers) according to your contact center platform.
  </Step>

  <Step title="Create a Handover Flow">
    1. In **Build > Flows**, create a [Flow](/ai/agents/develop/projects-and-flows/overview) and give it a clear name, for example, `Handover`.
    2. In the Flow editor, add a [Say Node](/ai/agents/develop/node-reference/basic/say) and set its **Text** parameter to `Hi, let's hand you over to the human agents.`.
    3. Add a [Handover to Human Agent Node](/ai/agents/develop/node-reference/service/handover-to-agent) and configure as follows:
       * **Handover Provider** — set to the previously [created Handover Provider](#create-a-handover-provider), and [configure it accordingly](/ai/escalate/handover-reference/overview). For example, if you use [Live Agent](/ai/escalate/handover-reference/live-agent), you need to set the **Live Agent Inbox ID**.
       * **Handover Accepted Message** — enter `Transferring to human agents...`.
  </Step>

  <Step title="Create an Agent Copilot Flow">
    1. In **Build > Flows**, create a Flow and give it a clear name, for example, `Agent Copilot`.
    2. Add a [Copilot: Set Grid Node](/ai/agents/develop/node-reference/agent-copilot/set-grid) and configure the grid as follows:

    ```json theme={null}
    {
      "grid": {
        "columns": 2,
        "rows": 2,
        "gap": 5
      },
      "tiles": {
        "next-best-action": {
          "x": 1,
          "y": 1,
          "rows": 1,
          "columns": 1
        },
        "sentiment": {
          "x": 2,
          "y": 1,
          "rows": 1,
          "columns": 1
        }
      }
    }
    ```

    3. After the Copilot: Set Grid Node, add an [LLM Prompt Node](/ai/agents/develop/node-reference/service/llm-prompt). Configure it to generate the next best response for the agent based on the user's latest input and to store the result in the Input object:

    * **System Prompt** — enter the following:

    ```txt theme={null}
    Based on the user's last message, suggest the most helpful next response for a human agent.
    ```

    * In the **Storage & Streaming** section, configure the following:
      * **How to handle the result** — set to **Store in Input**.
      * **Input Key to store Result** — set to `input.promptResult`.

    4. Add the following Agent Copilot Nodes after the LLM Prompt Node and configure them as follows:
       * **[Copilot: Next Action Tile Node](/ai/agents/develop/node-reference/agent-copilot/next-action-tile)**:
         * **Tile ID** — `next-best-action`
         * **Text**:
         ```txt theme={null}
         {{input.promptResult.detailedResult.choices[0].message.content}}
         ```
         This expression dynamically retrieves the LLM-generated suggestion for the tile.
         * *(Optional)* In the **Settings** section, set **Label** to `Next Best Action`.
       * **[Copilot: Sentiment Tile Node](/ai/agents/develop/node-reference/agent-copilot/sentiment-tile)**:
         * **Tile ID** — `sentiment`.
         * *(Optional)* In the **Settings** section, set **Label** to `Customer Sentiment`.
  </Step>

  <Step title="Test Agent Copilot">
    To test Agent Copilot, create a Webchat v3 Endpoint:

    1. In **Deploy > Endpoints**, create a [Webchat v3 Endpoint](/webchat/v3/configuration)
    2. In the Endpoint settings, configure the following:
       * Set **Flow** to the previously created Handover Flow, for example, `Handover`.
       * In the **Copilot** section, set **Copilot Flow** to the previously created Agent Copilot Flow, for example, `Agent Copilot`.
    3. Click **Save** to activate the **Open Demo Webchat** button.

    Then, to simulate a user interaction and test Agent Copilot, follow these steps:

    1. In the Webchat v3 Endpoint settings you created, click **Open Demo Webchat** in the upper-right corner. Demo Webchat opens.
    2. Click **Start conversation** and send the AI Agent a message, for example, `Hello, I need help with my order`. You should receive a message with the text you set in the Say Node followed by the text from the Handover to Human Agent Node.
       <Note>
         At the start of a conversation, the `Initial message for AI Copilot` reply may appear in the human agent's chat as if it were sent by the end user. This reply is added automatically to trigger Agent Copilot and is visible only to the human agent.
       </Note>
    3. Go to your contact center platform and search for the session with the message you sent. The Agent Copilot workspace displays two widgets that suggest next actions.
  </Step>
</Steps>

## More Information

* [Agent Copilot](/agent-copilot/overview)
* [Agent Copilot Nodes](/ai/agents/develop/node-reference/agent-copilot/overview)
* [Agent Copilot for Voice](/agent-copilot/getting-started/voice)


# Overview
Source: https://docs.cognigy.com/agent-copilot/getting-started/overview

Get started with Agent Copilot setup guides for chat, voice, and voice-to-chat AI agent modes.

These getting started guides help you set up Agent Copilot in different modes.

<CardGroup>
  <Card title="Chat" href="/agent-copilot/getting-started/chat">
    Set up Agent Copilot for AI Agents in chat applications.
  </Card>

  <Card title="Voice" href="/agent-copilot/getting-started/voice">
    Set up Agent Copilot for voice AI Agents with Voice Gateway.
  </Card>

  <Card title="Voice-to-Chat" href="/agent-copilot/getting-started/voice-to-chat">
    Set up Agent Copilot for voice AI Agents in chat applications that respond to users with voice via TTS.
  </Card>
</CardGroup>


# Getting Started with Agent Copilot for Voice
Source: https://docs.cognigy.com/agent-copilot/getting-started/voice

Set up Agent Copilot for voice applications with a workspace including next action widgets and call transcripts for human agents in contact centers.

In this guide, you will learn how to configure your first embedded Agent Copilot workspace for a voice AI Agent. This configuration provides an Agent Copilot workspace that includes two widgets:

* A [Copilot: Next Action Tile Node](/ai/agents/develop/node-reference/agent-copilot/next-action-tile) that suggests a next action to the human agent.
* A [Copilot: Transcript Tile Node](/ai/agents/develop/node-reference/agent-copilot/transcript-tile) that displays the conversation transcript to the human agent.

## Prerequisites

* Access to [Voice Gateway](/voice-gateway/index).
* Configure a speech provider in [Cognigy](/ai/agents/test/voice-preview#set-up-a-speech-provider) and the [Voice Gateway Self-Service Portal](/voice-gateway/webapp/speech-services).
* Configure the [phone number](/voice-gateway/webapp/phone-numbers) you use in your contact center in the Voice Gateway Self-Service Portal.
* Configure a contact center platform that supports voice calls, for example, [Genesys Cloud Open Messaging](/ai/escalate/handover-reference/genesys-cloud-open-messaging), and can integrate with a [handover provider](/ai/escalate/handover-reference/overview).
* Add a [compatible LLM](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) for the LLM Prompt Node.

## Setup Process

1. Create a handover provider. The handover provider connects the AI Agent to your contact center platform.
2. Create an Agent Copilot Flow. The Agent Copilot configures the Agent Copilot workspace and the individual widgets.
3. Create a Voice Copilot Endpoint. The Voice Copilot Endpoint receives the transcript of the voice call and sends it to the Agent Copilot workspace.
4. Create a Transfer Flow. The Transfer Flow configures the user interaction with the AI Agent and the handover to a human agent.
5. Create a Voice Gateway Endpoint. The Voice Gateway Endpoint connects the AI Agent to Voice Gateway.
6. Create an [application](/voice-gateway/webapp/applications) in the Voice Gateway Self-Service Portal. The Voice Gateway application connects your contact center phone number to the Voice Gateway Endpoint.
7. Test the Agent Copilot workspace by calling the phone number you configured in the Voice Gateway application.

## Set Up Agent Copilot for Voice

<Steps>
  <Step title="Create a Handover Provider">
    In **Deploy > Handover Providers**, create a [handover provider](/ai/escalate/handovers) according to your contact center platform. Depending on the platform, you may need to configure different settings for the Transfer Flow.
  </Step>

  <Step title="Create an Agent Copilot Flow">
    1. In **Build > Flows**, create a [Flow](/ai/agents/develop/projects-and-flows/overview) and give it a clear name, for example, `Agent Copilot`.

    2. Add a [Set Grid Node](/ai/agents/develop/node-reference/agent-copilot/set-grid) and set the grid as follows:

       ```json theme={null}
       {
         "grid": {
           "columns": 2,
           "rows": 2,
           "gap": 5
         },
         "tiles": {
           "next-action": {
             "x": 1,
             "y": 1,
             "rows": 1,
             "columns": 1
           },
           "transcript": {
             "x": 1,
             "y": 2,
             "rows": 1,
             "columns": 1
           }
         }
       }
       ```

    3. After the Copilot: Set Grid Node, add an [LLM Prompt Node](/ai/agents/develop/node-reference/service/llm-prompt). Configure it to generate the next best response for the agent based on the user's latest input and to store the result in the Input object:
       * **System Prompt** — enter the following:
         ```txt theme={null}
         Based on the user's last message, suggest the most helpful next response for a human agent.
         ```
       * In the **Storage & Streaming** section, configure the following:
         * **How to handle the result** — set to **Store in Input**.
         * **Input Key to store Result** — set to `input.promptResult`.

    4. Add the following Agent Copilot Nodes after the LLM Prompt Node and configure them as follows:
       * **[Copilot: Next Action Tile Node](/ai/agents/develop/node-reference/agent-copilot/next-action-tile)**:
         * **Tile ID** — `next-action`
         * **Text**:
         ```txt theme={null}
         {{input.promptResult.detailedResult.choices[0].message.content}}
         ```
         This expression dynamically retrieves the LLM-generated suggestion for the widget.
       * **[Copilot: Transcript Tile Node](/ai/agents/develop/node-reference/agent-copilot/transcript-tile)**:
         * **Tile ID** — `transcript`
  </Step>

  <Step title="Create a Voice Copilot Endpoint">
    1. In **Deploy > Endpoints**, click **+ New Endpoint** and select **Voice Copilot**.
    2. In the **New Endpoint** section, configure the following:
       * **Name** — enter a unique name, for example, `Voice Agent Copilot Endpoint`.
       * **Flow** — select the Agent Copilot Flow you created from the list.
    3. Copy the Endpoint URL for later use in the Transfer Flow and save the changes.
  </Step>

  <Step title="Create a Transfer Flow">
    1. In **Build > Flows**, create a [Flow](/ai/agents/develop/projects-and-flows/overview) and give it a clear name, for example, `Transfer`.
    2. In the Flow editor, add a [Say Node](/ai/agents/develop/node-reference/basic/say) and set its **Text** parameter to `Hi, I'll transfer you to the humans.`.
    3. Add a [Transfer Node](/ai/agents/develop/node-reference/voice/voice-gateway/transfer) and configure the following:
       * **Transfer Type** — select `Dial`.
       * **Transfer** — enter the phone number to redirect the call to a human agent. For testing purposes, use a phone number you have access to.
       * Activate the **Enable Copilot** toggle.
       * Enter the **Copilot Headers Key**. The key is provider-dependent. For example, for [Genesys Cloud Open Messaging](/ai/escalate/handover-reference/genesys-cloud-open-messaging), the key is `User-to-User`.
       * In the **Transcribe** section, enter the Voice Copilot Endpoint URL you copied from the Voice Copilot Endpoint settings in **Transcription Webhook**.
  </Step>

  <Step title="Create a Voice Gateway Endpoint">
    1. In **Deploy > Endpoints**, click **+ New Endpoint** and select **Voice Gateway**.
    2. In the **New Endpoint** section, configure the following:
       * **Name** — enter a unique name.
       * **Flow** — select the Transfer Flow from the list, in this example, `Transfer`.
    3. Copy the Endpoint URL for later use in the Voice Gateway application and save the changes.
  </Step>

  <Step title="Create a Voice Gateway Application">
    1. In the Voice Gateway Self-Service Portal, click **Application** in the left-side menu, then **Add application**.
    2. On the **Add an application** page, configure the following:
       * **Application name** — enter a unique name, for example, `Voice App`.
       * **Calling webhook** — enter the Endpoint URL you copied from the Voice Gateway Endpoint settings and make sure to select the **POST** in the **METHOD** field.
       * **Call status webhook** — enter the Endpoint URL you copied from the Voice Gateway Endpoint settings and make sure to select the **POST** in the **METHOD** field.
       * **Speech synthesis vendor** and **Speech recognizer vendor** — make sure you configure the same speech provider as configured in [Cognigy](/ai/agents/test/voice-preview#set-up-a-speech-provider).
    3. Click **Save**, then **Phone Numbers** in the left-side menu.
    4. Select the phone number you use in your contact center.
    5. In the **Application** field, select **Voice App** and save the changes.

    Your voice AI Agent can now receive calls and hand them over to a human agent with access to the Agent Copilot workspace.
  </Step>

  <Step title="Test Agent Copilot">
    To test Agent Copilot, follow these steps:

    1. Open your contact center platform and call the number you configured in the Voice Gateway Application.
       The voice AI Agent answers the call and creates a session in the contact center platform.
    2. Say something to the voice AI Agent, for example, `Thanks, I'll wait for the human agent.`. The AI Agent will transfer your call to the phone number you configured in the Transfer Flow.
    3. Pick up the call from the phone number you configured in the Transfer Flow.
       <Note>
         At the start of a conversation, the `Initial message for AI Copilot` reply may appear in the human agent's chat as if it were sent by the end user. This reply is added automatically to trigger Agent Copilot and is visible only to the human agent.
       </Note>
    4. The Agent Copilot workspace displays a widget with the conversation transcript and a widget with the next action.
    5. Say the next action text into the phone.
  </Step>
</Steps>

## More Information

* [Agent Copilot](/agent-copilot/overview)
* [Agent Copilot for Chat](/agent-copilot/getting-started/chat)
* [Agent Copilot Nodes](/ai/agents/develop/node-reference/agent-copilot/overview)
* [Agent Copilot for Voice](/agent-copilot/getting-started/voice)


# Getting Started with Agent Copilot for Voice-to-Chat
Source: https://docs.cognigy.com/agent-copilot/getting-started/voice-to-chat

Configure Agent Copilot for voice-to-chat applications with next action suggestion and conversation transcript widgets for human agents in contact centers.

In this guide, you will learn how to configure your first embedded Agent Copilot workspace for voice-to-chat applications. In these applications, the end user calls a voice AI Agent and the AI Agent hands over to a human agent. The human agent keeps supporting the end user through a chat in the contact center platform, and the text is converted to voice by a text-to-speech provider. This configuration provides an Agent Copilot workspace that includes two widgets:

* A [Copilot: Next Action Tile Node](/ai/agents/develop/node-reference/agent-copilot/next-action-tile) that suggests a next action to the human agent.
* A [Copilot: Transcript Tile Node](/ai/agents/develop/node-reference/agent-copilot/transcript-tile) that displays the conversation transcript to the human agent.

## Prerequisites

* Access to [Voice Gateway](/voice-gateway/index).
* Configure a speech provider in [Cognigy](/ai/agents/test/voice-preview#set-up-a-speech-provider) and the [Voice Gateway Self-Service Portal](/voice-gateway/webapp/speech-services).
* Configure the [phone number](/voice-gateway/webapp/phone-numbers) you use in your contact center in the Voice Gateway Self-Service Portal.
* Configure a contact center platform that you can integrate with a [handover provider](/ai/escalate/handover-reference/overview). For example, for [Live Agent](/ai/escalate/handover-reference/live-agent), you need to have an [Inbox](/live-agent/settings/inboxes) with the [Project ID and the Webhook URL set up](/live-agent/settings/inboxes#edit-inboxes) and, if necessary, the **Enable Agent Copilot Workspace** option activated.
* Add a [compatible LLM](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) for the LLM Prompt Node.

## Setup Process

1. Create a handover provider. The handover provider connects the AI Agent to your contact center platform.
2. Create a Handover Flow. The Handover Flow configures the interaction with the AI Agent and the handover to a human agent.
3. Create an Agent Copilot Flow. The Agent Copilot Flow configures the Agent Copilot workspace and the individual widgets.
4. Create a Voice Copilot Endpoint. The Voice Copilot Endpoint receives the transcript of the voice call and sends it to the Agent Copilot workspace.
5. Create a Voice Gateway Endpoint. The Voice Gateway Endpoint connects the AI Agent to Voice Gateway.
6. Create an [application](/voice-gateway/webapp/applications) in the Voice Gateway Self-Service Portal. The Voice Gateway application connects your contact center phone number to the Voice Gateway Endpoint.
7. Test the Agent Copilot workspace by calling the phone number you configured in the Voice Gateway application.

## Set Up Agent Copilot for Voice-to-Chat

<Steps>
  <Step title="Create a Handover Provider">
    In **Deploy > Handover Providers**, create a [handover provider](/ai/escalate/handovers) according to your contact center platform. Depending on the platform, you may need different settings for the Handover Flow.
  </Step>

  <Step title="Create a Handover Flow">
    1. In **Build > Flows**, create a [Flow](/ai/agents/develop/projects-and-flows/overview) and give it a clear name, for example, `Handover`.
    2. In the Flow editor, add a [Say Node](/ai/agents/develop/node-reference/basic/say) and set its **Text** parameter to `Hi, let's hand you over to the humans.`.
    3. Add a [Handover to Human Agent Node](/ai/agents/develop/node-reference/service/handover-to-agent) and configure the following:

    * **Handover Provider** — set to the previously [created Handover Provider](#create-a-handover-provider), and [configure it accordingly](/ai/escalate/handover-reference/overview). For example, if you use [Live Agent](/ai/escalate/handover-reference/live-agent), you need to set the **Live Agent Inbox ID**.
    * **Handover Accepted Message** — enter `Transferring to humans...`.
  </Step>

  <Step title="Create an Agent Copilot Flow">
    1. In your Cognigy Project, go to **Build > Flows** and create a [Flow](/ai/agents/develop/projects-and-flows/overview) and give it a clear name, for example, `Agent Copilot`.
    2. Add a [Set Grid Node](/ai/agents/develop/node-reference/agent-copilot/set-grid) and set the **Copilot Grid Configuration** as follows:

       ```json theme={null}
       {
         "grid": {
           "columns": 2,
           "rows": 2,
           "gap": 5
         },
         "tiles": {
           "next-action": {
             "x": 1,
             "y": 1,
             "rows": 1,
             "columns": 1
           },
           "transcript": {
             "x": 1,
             "y": 2,
             "rows": 1,
             "columns": 1
           }
         }
       }
       ```
    3. After the Copilot: Set Grid Node, add an [LLM Prompt Node](/ai/agents/develop/node-reference/service/llm-prompt). Configure it to generate the next best response for the agent based on the user's latest input and to store the result in the Input object:

    * **System Prompt** — enter the following:

    ```txt theme={null}
    Based on the user's last message, suggest the most helpful next response for a human agent.
    ```

    * In the **Storage & Streaming** section, configure the following:
      * **How to handle the result** — set to **Store in Input**.
      * **Input Key to store Result** — set to `input.promptResult`.

    4. Add the following Agent Copilot Nodes after the LLM Prompt Node and configure them as follows:
       * **[Copilot: Next Action Tile Node](/ai/agents/develop/node-reference/agent-copilot/next-action-tile)**:
         * **Tile ID** — `next-action`
         * **Text**:
         ```txt theme={null}
         {{input.promptResult.detailedResult.choices[0].message.content}}
         ```
         This expression dynamically retrieves the LLM-generated suggestion for the tile.

    * **[Copilot: Transcript Tile Node](/ai/agents/develop/node-reference/agent-copilot/transcript-tile)**:
      * **Tile ID** — `transcript`
  </Step>

  <Step title="Create a Voice Copilot Endpoint">
    1. In **Deploy > Endpoints**, click **+ New Endpoint** and select **Voice Copilot**.
    2. In the **New Endpoint** section, configure the following:
       * **Name** — enter a unique name, for example, `Voice Agent Copilot Endpoint`.
       * **Flow** — select the Agent Copilot Flow you created from the list.
  </Step>

  <Step title="Create a Voice Gateway Endpoint">
    1. In **Deploy > Endpoints**, click **+ New Endpoint** and select **Voice Gateway**.
    2. In the **New Endpoint** section, configure the following:
       * **Name** — enter a unique name.
       * **Flow** — select the Handover Flow from the list, in this example, `Handover`.
    3. Copy the Endpoint URL for later use in the Voice Gateway application and save the changes.
  </Step>

  <Step title="Create a Voice Gateway Application">
    1. In the [Voice Gateway Self-Service Portal](/voice-gateway/webapp/overview), click **Application** in the left-side menu, then **Add application**.
    2. On the **Add an application** page, configure the following:
       * **Application name** — enter a unique name, for example, `Voice-to-Chat App`.
       * **Calling webhook** — enter the Endpoint URL you copied from the Voice Gateway Endpoint settings and make sure to select the **POST** in the **METHOD** field.
       * **Call status webhook** — enter the Endpoint URL you copied from the Voice Gateway Endpoint settings and make sure to select the **POST** in the **METHOD** field.
       * **Speech synthesis vendor** and **Speech recognizer vendor** — make sure you configure the same speech provider as configured in [Cognigy](/ai/agents/test/voice-preview#set-up-a-speech-provider).
    3. Click **Save**, then **Phone Numbers** in the left-side menu.
    4. Select the phone number you use in your contact center.
    5. In the **Application** field, select **Voice-to-Chat App** and save the changes.

    Your voice AI Agent can now receive calls and hand them over to a human agent with access to the Agent Copilot workspace.
  </Step>

  <Step title="Test Agent Copilot">
    To test Agent Copilot, follow these steps:

    1. Open your contact center platform and call the number you configured in the Voice Gateway Application.
       The voice AI Agent answers the call and creates a chat session in the contact center platform.
    2. Say something to the voice AI Agent, for example, `Thanks, I'll wait for the human agent.`. Your voice message is sent as text to the chat in the contact center platform.
    3. Find and select the corresponding chat in the contact center platform.
       <Note>
         At the start of a conversation, the `Initial message for AI Copilot` reply may appear in the human agent's chat as if it were sent by the end user. This reply is added automatically to trigger Agent Copilot and is visible only to the human agent.
       </Note>
    4. The Agent Copilot workspace displays a widget with the conversation transcript and a widget with the next action.
    5. Click the icon to paste the next action text into the chat and send the message. The voice AI Agent responds with the text from the next action widget.
  </Step>
</Steps>

## More Information

* [Agent Copilot](/agent-copilot/overview)
* [Getting started with Agent Copilot](/agent-copilot/getting-started/chat)
* [Agent Copilot Configuration](/agent-copilot/configure/overview)
* [Agent Copilot Nodes](/ai/agents/develop/node-reference/agent-copilot/overview)
* [Agent Copilot Embedding](/agent-copilot/deploy/embedding)
* [Agent Copilot for Voice](/agent-copilot/getting-started/voice)
* [Integration with Contact Centers](/agent-copilot/configure/overview)


# Agent Copilot Glossary
Source: https://docs.cognigy.com/agent-copilot/glossary

Definitions for Agent Copilot terms including workspace, grid, tiles, widgets, and deployment methods.

Agent Copilot terms listed in alphabetical order. All terms in the glossary are provided for informational purposes.

## A

**Agent Copilot**

A Cognigy product that offers a variety of features in a [workspace](#w) and empowers human agents to provide faster and more accurate customer support. You can customize the Agent Copilot workspace with [widgets](#w) that display the most relevant information for chat and voice use cases. When embedded into your contact center's desktop, Agent Copilot enhances the overall efficiency of your contact center. For more information, read the [Agent Copilot](/agent-copilot/overview) documentation.

**Agent Copilot Flow**

A Flow separate from the AI Agent Flow to which you add [Agent Copilot Nodes](#a). The Agent Copilot Flow controls the [workspace](#w) arrangement on the contact center side and the [widgets](#w) in the workspace. The Copilot: Set Grid Node is required in this Flow to set the workspace layout. Other Agent Copilot Nodes provide functionality to the widgets in the workspace. For more information, read the [Agent Copilot Configuration](/agent-copilot/configure/overview) documentation.

**Agent Copilot Nodes**

Nodes that allow you to configure and control the [Agent Copilot](#a) [workspace](#w) from a Flow. Agent Copilot Nodes include configuration Nodes, [plug-and-play widgets](#p), and customization Nodes that provide advanced capabilities to the [widgets](#w) in the workspace. For more information, read the [Agent Copilot Nodes](/agent-copilot/configure/nodes) documentation.

## E

**Embedded Deployment**

A deployment method where [Agent Copilot](#a) is integrated directly into the contact center's agent desktop interface, providing human agents with real-time assistance within their existing workflow. For more information, read the [Agent Copilot Embedding](/agent-copilot/deploy/embedding) documentation.

## G

**Grid**

A JSON object that defines the layout structure of the [Agent Copilot](#a) [workspace](#w). The grid specifies the number of columns and rows, the gap between [tiles](#t), and the position and size of each tile through [tile IDs](#t). You configure the grid using the Copilot: Set Grid Node or via the Cognigy API. For more information, read the [Configure the Workspace](/agent-copilot/configure/workspace) documentation.

## P

**Plug-and-Play Widgets**

Pre-built [Agent Copilot Nodes](#a) that require minimal configuration to integrate into the [Agent Copilot](#a) [workspace](#w). Examples include the Identity Tile, Next Action Tile, Sentiment Tile, and Transcript Tile. These [widgets](#w) provide ready-to-use functionality for common use cases. For more information, read the [Agent Copilot Nodes](/agent-copilot/configure/nodes) documentation.

## S

**Standalone Deployment**

A deployment method where [Agent Copilot](#a) operates as an independent application rather than being embedded into a contact center platform. For more information, read the [Standalone Agent Copilot](/agent-copilot/deploy/standalone) documentation.

## T

**Tile**

A configuration object in the [Agent Copilot](#a) [grid](#g) that represents a visual component in the Agent Copilot [workspace](#w). Each tile corresponds to an [Agent Copilot Node](#a) in the [Agent Copilot Flow](#a) and must have a unique [tile ID](#t). Tiles can display content such as HTML, adaptive cards, iframes, or specialized [widgets](#w) like sentiment analysis or knowledge retrieval. For more information, read the [Configure the Workspace](/agent-copilot/configure/workspace) documentation.

**Tile ID**

A unique identifier assigned to each [tile](#t) in the [Agent Copilot](#a) [workspace](#w). The tile ID is used in the [grid](#g) configuration to define the tile's position and size, and it must match the tile ID specified in the corresponding [Agent Copilot Node](#a). For example, `transcript`, `knowledge`, or `next-action`. For more information, read the [Tile ID](/agent-copilot/configure/workspace#tile-id) documentation.

## W

**Widget**

A visual component in the [Agent Copilot](#a) [workspace](#w) that displays information or provides interactive functionality to human agents. Widgets are created using [Agent Copilot Nodes](#a) and are arranged within [tiles](#t) in the [grid](#g). Examples include sentiment analysis widgets, knowledge retrieval widgets, and next action suggestion widgets. Also referred to as a tile. For more information, read the [Agent Copilot](/agent-copilot/overview) documentation.

**Workspace**

The customizable interface that displays [Agent Copilot](#a) on the contact center side. The workspace contains [widgets](#w) arranged in a [grid](#g) layout that show the most relevant information for chat and voice use cases. You can configure the workspace layout, widget size, and arrangement to optimize the human agent's workflow. For more information, read the [Configure the Workspace](/agent-copilot/configure/workspace) documentation.


# Cognigy Agent Copilot
Source: https://docs.cognigy.com/agent-copilot/index



<div>
  <div>
    *Cognigy Agent Copilot* offers a variety of advanced features that empower agents to provide faster and more accurate customer support.

    <p>
      The workspace is fully customizable with widgets that display the most relevant information for any chat and voice use case.
    </p>

    <p>
      When embedded into your contact center's agent desktop, it enhances the overall efficiency of your contact center.
    </p>
  </div>

  <div>
    <video aria-label="Cognigy Agent Copilot hero animation showing AI Agent capabilities: knowledge, language, wrap-up, sentiment, and identity assistance.">
      <source type="video/mp4" />
    </video>

    <video aria-label="Cognigy Agent Copilot hero animation showing AI Agent capabilities: knowledge, language, wrap-up, sentiment, and identity assistance.">
      <source type="video/mp4" />
    </video>

    <button type="button" aria-label="Pause video">
      ⏸
    </button>
  </div>
</div>

<CardGroup>
  <Card title="Explore Agent Copilot" icon="info" href="/agent-copilot/overview">
    Discover Agent Copilot's capabilities.
  </Card>

  <Card title="Build Your First Agent Copilot" icon="hammer" href="/agent-copilot/getting-started/overview">
    Build your Agent Copilot using Agent Copilot Nodes and ready-to-use widgets.
  </Card>

  <Card title="Configure Agent Copilot" icon="cog" href="/agent-copilot/configure/overview">
    Define the use case your Agent Copilot will handle.
  </Card>

  <Card title="Deploy and Connect" icon="cable" href="/agent-copilot/deploy/overview">
    Set up the connection to your contact center.
  </Card>

  <Card title="Troubleshooting Guide" icon="bug-off" href="/agent-copilot/troubleshooting">
    Diagnose and resolve common Agent Copilot issues.
  </Card>
</CardGroup>


# Browser Requirements
Source: https://docs.cognigy.com/agent-copilot/installation/browser-requirements

Supported browsers and minimum version requirements for Agent Copilot.

The Agent Copilot web frontend is compatible with most modern web browsers. The table below provides a list of supported web browsers.

| Browser         | Version       |
| --------------- | ------------- |
| Mozilla Firefox | 110 or higher |
| Google Chrome   | 110 or higher |
| Chromium        | 87 or higher  |
| Apple Safari    | 14 or higher  |
| Microsoft Edge  | 110 or higher |


# Agent Copilot
Source: https://docs.cognigy.com/agent-copilot/overview

Learn about Agent Copilot features, including customizable workspaces, rich media capabilities, and integration with contact center platforms for human agent support.

*Cognigy Agent Copilot* offers a variety of features in a workspace and empowers human agents to provide faster and more accurate customer support. You can customize the Agent Copilot workspace with widgets that display the most relevant information for chat and voice use cases. When embedded into your contact center's desktop, Agent Copilot enhances the overall efficiency of your contact center.

## Key features

* Displays interactive micro applications in a customizable workspace with rich-media capabilities.
* Standalone or embedded deployment into contact center platforms.
* Shows any type of information, such as PDFs, automatic search results, or even interactive 3D animations.
* Access to 100% of Cognigy features, including Extensions.

## More Information

* [Getting started with Agent Copilot](/agent-copilot/getting-started/overview)
* [Agent Copilot Configuration](/agent-copilot/configure/overview)
* [Agent Copilot Nodes](/ai/agents/develop/node-reference/agent-copilot/overview)
* [Agent Copilot Embedding](/agent-copilot/deploy/embedding)
* [Agent Copilot for Voice](/agent-copilot/getting-started/voice)
* [Agent Copilot for Chat](/agent-copilot/getting-started/chat)
* [Integration with Contact Centers](/agent-copilot/configure/overview)
* [Agent Copilot in Live Agent](/live-agent/assistants/agent-copilot)


# Troubleshooting
Source: https://docs.cognigy.com/agent-copilot/troubleshooting

Fix common Agent Copilot issues including missing tiles, connection errors, and sentiment tile failures.

This page lists common issues you may encounter while using Agent Copilot and provides troubleshooting steps to resolve them.

<AccordionGroup>
  <Accordion title="Why are some tiles not appearing in my Copilot workspace?">
    You may see a message like `The following tiles are not registered in the workspace layout and will not be rendered: <tileId-1>, <tileId-2>.`
    This occurs because some of the Nodes in your Flow aren't correctly configured in the [Copilot: Set Grid Node](/ai/agents/develop/node-reference/agent-copilot/set-grid).

    To fix this issue, ensure all Copilot Nodes are added to the Set Grid Node with the exact same Tile ID; check for typos. If the tiles are not needed, delete the unused tiles from the grid and related Nodes from your Flow.
  </Accordion>

  <Accordion title="My Copilot workspace shows the error `<host-name> refused to connect`">
    This usually indicates the workspace is pointing to the wrong host. Use the correct base URL (`AICopilotBaseUrl`) for your installation.
  </Accordion>

  <Accordion title="Why is the Sentiment tile failing to render?">
    Two key issues can prevent the Sentiment tile from displaying:

    * No model selected for Sentiment Analysis. In **Manage > Settings > Generative AI Settings**, make sure a model, such as `gpt-4o` is selected. Without this model, Agent Copilot can't render the tile and may return a 500 error.
    * Flow contains incompatible Nodes. Standard output Nodes, such as Say and Question Nodes, can't be placed below the [Copilot: Set Grid Node](/ai/agents/develop/node-reference/agent-copilot/set-grid) in the Agent Copilot Flow. Remove or disable incompatible Nodes to render the tile correctly.
  </Accordion>
</AccordionGroup>


# Use Cases
Source: https://docs.cognigy.com/agent-copilot/use-cases

Explore Agent Copilot use cases for sentiment analysis, knowledge retrieval, call summaries, and real-time translation.

Agent Copilot helps human agents support customers more effectively by streamlining workflows and simplifying daily tasks. The following use cases show how these features can be used in practice.

## Access Relevant Customer Data with Contextual Handover

When a live chat with a human agent begins, Agent Copilot retrieves relevant CRM data, previous queries, and interaction summaries.
All context is displayed directly in the Agent Copilot workspace. Human agents get a complete view of the customer's situation without switching tabs or navigating multiple systems, which leads to faster and more efficient issue resolution.

<iframe title="Access Relevant Customer Data with Contextual Handover" />

## Detect Customer Emotions with Sentiment Analysis

Analyze the emotional tone of customer interactions with the [Sentiment widget](/ai/agents/develop/node-reference/agent-copilot/sentiment-tile) in real time. This widget helps human agents identify frustration or positive sentiment. Human agents can respond appropriately to de-escalate situations and capture positive customer experiences.

<div>
  <video title="Sentiment analysis widget showing real-time emotion detection in Agent Copilot" aria-label="Sentiment analysis widget showing real-time emotion detection in Agent Copilot">
    <source />
  </video>
</div>

## Automate Post-Call Summaries and Updates

After a call, Agent Copilot automatically transcribes the conversation, generates summaries, collects feedback, and updates records. This automation reduces manual tasks, enabling faster follow-ups and ensuring accurate, actionable customer records.

<iframe title="Automate Post-Call Summaries and Updates" />

## Integrate Knowledge Retrieval

Use a [Copilot: Knowledge Tile](/ai/agents/develop/node-reference/agent-copilot/knowledge-tile) Node to integrate RAG capabilities into your contact center platform. This allows Agent Copilot to bring up information from your knowledge base to human agents to answer user questions.

<iframe title="Integrate Knowledge Retrieval" />

## Provide Real-Time Translation for Human Agents

In combination with Cognigy's [Real-Time Translation](/ai/platform-features/translation-and-localization/real-time-translation) feature, you can provide human agents with translation of the user input in real time. Additionally, the human agent can reply in their own language, and the reply is automatically translated back to the user's language.

<iframe title="Provide Real-Time Translation for Human Agents" />

## More Information

* [Agent Copilot Nodes](/ai/agents/develop/node-reference/agent-copilot/overview)
* [Copilot: Knowledge Tile](/ai/agents/develop/node-reference/agent-copilot/knowledge-tile)
* [Real-Time Translation](/ai/platform-features/translation-and-localization/real-time-translation)


# Access Control
Source: https://docs.cognigy.com/ai/administer/access/admin-center/access-control



<a href="/release-notes/2025.22"><Badge>Updated in 2025.22</Badge></a>

The *Access Control* interface provides a user management tool for organizational-level administration.
The interface acts as an Access Control List (ACL), enabling the assignment of specific roles to users.

In the Access Control interface, you can assign global roles only. These roles apply to all Projects within your organization and set basic access rights for all agents the user can access. If you want to assign Project-specific roles, refer to [Members](/ai/administer/access/members).

## Before you Begin

### Explore Global Roles

Global roles are assigned to a user to establish access control across the entire organization. For more granular control within a Project, assign [Project-specific roles](/ai/administer/access/members).

<ExpandableTable />

### Check Who Can Access the Access Control Page

Users with `admin` rights have full control over the Admin Center interface. However, users with one of the following roles have access only to the **Access Control** page:

* `projectAdmin` – a [Project role](/ai/administer/access/members#project-roles), read-only access.
* `memberManager` – a [Project role](/ai/administer/access/members#project-roles), read-only access.
* `userManager` – a global role, full access.
* `userDetailsViewer` – a global role, read-only access.

## Working with the Access Control interface

<AccordionGroup>
  <Accordion title="Create New Users">
    To create a new user in Cognigy for your organization, you must have the `admin` or `userManager` global role.

    To create a new user, follow these steps:

    1. In the Admin Center interface, select **Access Control** from the left-side menu, then click **+ Create new User**.
    2. In the **New User** window, fill in the following fields:
       * **Display Name** — a name that is visible to other users within the organization.
       * **Email Address** — a unique email address associated with the user's account.
       * **Password** — a user password. Passwords can be modified at any time and are securely stored. Users can change their passwords later by navigating **User Menu > My Profile** on the **Main** page or by selecting **My Profile** from the User Menu in the bottom-left corner of the **Project** page.
    3. Click **Save & Open**. A page opens where you can assign global roles and Projects to the user.

    When you create a new user, this user receives a notification at the email address specified in their profile.
  </Accordion>

  <Accordion title="Assign Global Roles to Users">
    To assign global roles to a user, follow these steps:

    1. In the upper-right corner of the **Main** page, select **User Menu > Admin Center**.
    2. In the Admin Center interface, select **Access Control** from the left-side menu. Select a user you want to assign global roles to:

    <Tabs>
      <Tab title="New user">
        1) On the **Users** page, create a new user by clicking **+ Create new User**.
        2) In the **New User** window, fill in the following fields:
           * **Display Name** — a name that is visible to other users within the organization.
           * **Email Address** — a unique email address associated with the user's account.
           * **Password** — a user password. Passwords can be modified at any time and are securely stored. Users can change their passwords later by navigating **User Menu > My Profile** on the **Main** page or by selecting **My Profile** from the User Menu in the bottom-left corner of the **Project** page.
        3) Click **Save & Open**. A page opens where you can assign global roles to the user.
        4) In the **Global Roles** section, click **+** to open a list of roles. Select a role from the list.
      </Tab>

      <Tab title="Existing user">
        1. On the **Users** page, select the user you want to assign roles to.
        2. In the **Global Roles** section, click **+** to open a list of roles. Select a role from the list.
      </Tab>
    </Tabs>

    After a role is assigned to a user, the role label is displayed in the **Global Roles** section.
  </Accordion>

  <Accordion title="Assign Projects to Users">
    To assign a Project to a user, follow these steps:

    1. In the upper-right corner of the **Main** page, select **User Menu > Admin Center**.
    2. In the Admin Center interface, select **Access Control** from the left-side menu. Select a user you want to assign global roles to:

    <Tabs>
      <Tab title="New user">
        1) On the **Users** page, create a new user by clicking **+ Create new User**.
        2) In the **New User** window, fill in the following fields:
           * **Display Name** — a name that is visible to other users within the organization.
           * **Email Address** — a unique email address associated with the user's account.
           * **Password** — a user password. Passwords can be modified at any time and are securely stored. Users can change their passwords later by navigating **User Menu > My Profile** on the **Main** page or by selecting **My Profile** from the User Menu in the bottom-left corner of the **Project** page.
        3) Click **Save & Open**. A page opens where you can assign Projects to the user.
        4) In the **Assigned Projects** section, click **+** to open a list of available Projects.
        5) Type the name of the Project, then select it from the list, or scroll to find and select a Project.
      </Tab>

      <Tab title="Existing user">
        1. On the **Users** page, select the user you want to assign Projects to.
        2. In the **Assigned Agents** section, click **+** to open a list of available Projects.
        3. Type the name of the Project, then select it from the list, or scroll to find and select a Project.
      </Tab>
    </Tabs>

    After a Project is assigned to a user, the Project label is displayed in the **Assigned Projects** section.

    When you assign a Project to an existing user, this user receives a notification at the email address specified in their profile. The email includes information about the Project name and its URL.
  </Accordion>
</AccordionGroup>

## More Information

* [Members](/ai/administer/access/members)
* [Admin Center](/ai/administer/access/admin-center/overview)


# Audit Events
Source: https://docs.cognigy.com/ai/administer/access/admin-center/audit-events



<a href="/release-notes/2026.10"><Badge>Updated in 2026.10</Badge></a>

The *Audit Events* interface allows users with the `admin` and `fullUserSupport` roles to track events at an organization-wide level. For installations integrated with NiCE CXone, this interface also tracks user, tenant, and role sync events triggered by the integration. You can also track events related to this integration in the [Activity Audit Report](https://help.nicecxone.com/content/dashboardwem/activityaudit.htm) in NiCE CXone.

You can access the **Audit Events** page using one of the following options:

* Click **User Menu > Admin Center** on the **Main** page, then select **Audit Events** from the left-side menu.
* Select **Admin Center** from the User Menu in the bottom-left corner of the **Project** page, then select **Audit Events** from the left-side menu.

## Working with the Audit Events Interface

<AccordionGroup>
  <Accordion title="View List of Audit Events">
    The list of audit events displays all events that occurred in the past 30 days. You can filter and sort the events.

    | Column Name   | Description                                                                                                                |
    | ------------- | -------------------------------------------------------------------------------------------------------------------------- |
    | Type          | The category of the event, such as `Action`, `Create`, `Replace`, `Patch`, `Delete`, `Authorization`, or `Authentication`. |
    | Action        | The specific action that triggered the event, such as `cancelTask`, `addRoleToUser`, or `addProjectMember`.                |
    | Resource Type | The type of resource involved in the event, such as Project, Node, Flow, and others.                                       |
    | Timestamp     | The exact date and time when the event occurred.                                                                           |
    | Project       | The identification number of the Project where the event occurred.                                                         |
    | User Email    | The email address of the user who triggered the event, identifying the responsible party.                                  |
  </Accordion>

  <Accordion title="View User Details">
    To view the details of the user who initiated the audit event, follow these steps:

    1. Access the **Audit Events** page using one of the following options:
       * Click **User Menu > Admin Center** on the **Main** page, then select **Audit Events** from the left-side menu.
       * Select **Admin Center** from the User Menu in the bottom-left corner of the **Project** page, then select **Audit Events** from the left-side menu.
    2. On the **Audit Events** page, click the user's email in the audit event row. The system will open a window showing the user's Projects and roles within the organization.
  </Accordion>

  <Accordion title="View Event Chain">
    To view an event chain, follow these steps:

    1. Access the **Audit Events** page using one of the following options:
       * Click **User Menu > Admin Center** on the **Main** page, then select **Audit Events** from the left-side menu.
       * Select **Admin Center** from the User Menu in the bottom-left corner of the **Project** page, then select **Audit Events** from the left-side menu.
    2. On the **Audit Events** page, hover your cursor over an audit event row to view the event's chain. The system will display the dependencies of all resources related to the event. For example, for the Node resource type, the dependencies will appear in the following format:

       ```txt theme={null}
       locale 67291073e1a6d2732253b620
         → flow 6732fb2eb23a1561e0c4a91b
            → node 6732fc9903cb134f00e349ad
       ```
  </Accordion>

  <Accordion title="View Audit Event Details">
    To view audit event details, follow these steps:

    1. Access the **Audit Events** page using one of the following options:
       * Click **User Menu > Admin Center** on the **Main** page, then select **Audit Events** from the left-side menu.
       * Select **Admin Center** from the User Menu in the bottom-left corner of the **Project** page, then select **Audit Events** from the left-side menu.
    2. On the **Audit Events** page, select an event from the list. The **Raw Event Details** window appears on the right side of the page.

    The event is displayed in JSON format. You can copy the event data by clicking <img alt="menu" /> **> Copy Event Details**.
  </Accordion>
</AccordionGroup>

## More Information

* [Admin Center Dashboard](/ai/administer/access/admin-center/dashboard)
* [Admin Center](/ai/administer/access/admin-center/overview)
* [Access Control](/ai/administer/access/admin-center/access-control)


# Admin Center Dashboard
Source: https://docs.cognigy.com/ai/administer/access/admin-center/dashboard



<a href="/release-notes/2026.15"><Badge>Updated in 2026.15</Badge></a>

The *Admin Center Dashboard* allows you to track the number of billable conversations, Knowledge Chunks, and Queries within your organization. Additionally, you can activate other Cognigy products, such as Live Agent and Voice Gateway, as well as delete organizations if necessary.

On the Admin Center Dashboard, you can perform the following actions:

* [View Charts](#charts)
* [Set up Cognigy Products](#set-up-cognigy-products)
* [Delete Organizations](#delete-organizations)

You can access the **Dashboard** page using one of the following options:

* Click **User Menu > Admin Center** on the **Main** page, then select **Dashboard** from the left-side menu.
* Select **Admin Center** from the User Menu in the bottom-left corner of the **Project** page, then select **Dashboard** from the left-side menu.

## Charts

Charts display key metrics, such as billable conversations and Knowledge AI usage, to help manage costs within the organization. Data can be filtered by month and year.

<AccordionGroup>
  <Accordion title="Billable Conversations">
    The chart shows the number of [billable conversations](/ai/administer/billing) for the organization.

    <Frame>
      <img alt="Billable conversations chart in the Admin Center Dashboard" />
    </Frame>

    Conversation counts are calculated overnight. For this reason, data for the current day isn't included in this chart. The billable conversations metrics include the following:

    1. The number of billable conversations created within the selected period.
    2. The total number of billable conversations created overall.

    <Note>
      Historical data for conversation counts is available only from 2025 onward.
    </Note>
  </Accordion>

  <Accordion title="Knowledge AI">
    Within each organization, if the Knowledge AI feature is used, you can monitor the number of Knowledge Queries and Knowledge Chunks.

    Knowledge Chunks and Knowledge Queries are billable units. Tracking their usage helps manage costs and optimize resource allocation.

    <Frame>
      <img alt="Knowledge AI metrics chart showing queries and chunks usage" />
    </Frame>

    The Knowledge AI metrics include the following:

    1. The number of Knowledge Queries performed within the selected period.
    2. The total number of Knowledge Queries performed.
    3. The number of Knowledge Chunks currently available in the organization.
    4. The maximum number of Knowledge Chunks allocated to the organization.

    For more information on how to activate Knowledge AI for your organization, read the [Knowledge AI](/ai/agents/develop/knowledge-ai/activate) documentation.
  </Accordion>
</AccordionGroup>

## Set up Cognigy Products

The **Cognigy Products** section allows you to activate additional Cognigy products for your organization, such as [Live Agent](/live-agent/index) and [Voice Gateway](/voice-gateway/index).

## Delete Organizations

<AccordionGroup>
  <Accordion title="1. Before Deleting Organizations">
    <Danger>
      Deleting an organization is irreversible and permanently erases all associated data.
    </Danger>

    Before the organization deletion, you need to set up:

    * Valid system administrator emails in the environment variable `SYS_ADMINISTRATOR_EMAIL` of the `values.yaml` file. For example, `SYS_ADMINISTRATOR_EMAIL: "admin1@domain.com,admin2@domain.com"`. Your administrators receive a notification at these email addresses regarding the organization deletion request. By following the link provided in the email, the administrators can proceed with the organization deletion.
    * The Cognigy [Management UI](/ai/administer/access/management-ui) is installed, and its path is set in the `MANAGEMENTUI_BASE_URL_WITH_PROTOCOL` variable of the `values.yaml` file. For example, `MANAGEMENTUI_BASE_URL_WITH_PROTOCOL: http://management-ui`.
  </Accordion>

  <Accordion title="2. Deletion Process for Organizations">
    This process contains the following steps:

    1. On the **Dashboard** page, a user with the admin role requests to delete the organization by clicking **Delete Organization**. This request is sent to the system administrator's email.
    2. The system administrator receives the email and follows the link to the [Management UI](/ai/administer/access/management-ui).
    3. The system administrator deletes the organization in the Management UI and then receives a confirmation email.
  </Accordion>
</AccordionGroup>


# Global Connections
Source: https://docs.cognigy.com/ai/administer/access/admin-center/global-resources/global-connections



<a href="/release-notes/4.97"><Badge>Updated in 4.97</Badge></a>

*Global Connections* are [Connections](/ai/platform-features/connections) established when you create [global LLMs](/ai/agents/develop/gen-ai-and-llms/llms).

## Working with Global LLMs

<Tabs>
  <Tab title="GUI">
    You can view, edit, and delete global Connections in the [Admin Center](/ai/administer/access/admin-center/global-resources/overview) under **Resources > Connections** as well as copy their Reference ID.
  </Tab>

  <Tab title="API">
    You can view, edit, and delete global Connections using the [Cognigy API](https://api-trial.cognigy.ai/openapi#tag--Connections-v2.0). To do so, set the `resourceLevel` parameter to `organisation` in the JSON body.
  </Tab>
</Tabs>


# Global Large Language Models
Source: https://docs.cognigy.com/ai/administer/access/admin-center/global-resources/global-llms



<a href="/release-notes/4.97"><Badge>Updated in 4.97</Badge></a>

*Global Large Language Models* (Global LLMs) are [LLMs](/ai/agents/develop/gen-ai-and-llms/llms) at [organization level](/ai/administer/access/admin-center/global-resources/overview) that you can assign to and share across different Projects in your organization.

## Restrictions

* Global LLMs don't support [LLM fallback](/ai/agents/develop/gen-ai-and-llms/fallback).
* Global LLMs can't be set as a [default model](/ai/agents/develop/gen-ai-and-llms/other-operations).
* Global LLMs can't be cloned.

## Working with Global LLMs

<Tabs>
  <Tab title="GUI">
    You can view, create, edit, and delete global LLMs in the [Admin Center](/ai/administer/access/admin-center/global-resources/overview) under **Resources > LLMs** as well as copy their Reference ID.

    To assign a global LLM to a Project, go to the global LLM page, click <img alt="assign" /> or <img alt="edit-assign" /> in the **Assigned Projects** section, and select the Project.

    You can also view global LLMs in **Build > LLMs** and edit their settings.
  </Tab>

  <Tab title="API">
    You can view, create, delete, and edit global LLMs using the [Cognigy API](https://api-trial.cognigy.ai/openapi#tag--LargeLanguageModels-v2.0). To do so, set the `resourceLevel` parameter to `organisation` in the JSON body.
  </Tab>
</Tabs>


# Global Resources
Source: https://docs.cognigy.com/ai/administer/access/admin-center/global-resources/overview



<a href="/release-notes/4.97"><Badge>Updated in 4.97</Badge></a>

*Global Resources* are Cognigy resources that you can assign to and share across different Projects in your organization. Global resources work similarly to their counterparts at Project level and have a similar [user interface as other resources](/ai/overview/user-interface/resource-pages).

Global resources include:

* [Global Large Language Models](/ai/administer/access/admin-center/global-resources/global-llms)
* [Global Connections](/ai/administer/access/admin-center/global-resources/global-connections)

## Restrictions

* Only users with `admin` [global role](/ai/administer/access/admin-center/access-control) can access global resources.
* You can't export global resources as a [Package](/ai/platform-features/packages) or [Snapshot](/ai/agents/deploy/snapshots).


# Admin Center
Source: https://docs.cognigy.com/ai/administer/access/admin-center/overview



<a href="/release-notes/4.89"><Badge>Updated in 4.89</Badge></a>

The *Admin Center* is the central hub in Cognigy that enables users with `admin`<sup>[1](#footnote1)</sup> rights to manage the organization. With the Admin Center, you can monitor key resource metrics, control access, track system events, and manage global resources.

* [Dashboard](/ai/administer/access/admin-center/dashboard)
* [Access Control](/ai/administer/access/admin-center/access-control)
* [Audit Events](/ai/administer/access/admin-center/audit-events)
* [Resources](/ai/administer/access/admin-center/global-resources/overview)

Access the Admin Center using one of the following options:

* Click **User Menu > Admin Center** on the **Main** page.
* Select **Admin Center** from the User Menu in the bottom-left corner of the **Project** page.

***

<sup>1</sup>: While users with admin rights have full control over the Admin Center, users with other roles can view or manage specific pages within the Admin Center.


# Management UI
Source: https://docs.cognigy.com/ai/administer/access/management-ui



<a href="/release-notes/2026.15"><Badge>Updated in 2026.15</Badge></a>

The *Management UI* is a web-based interface for managing system-wide and organization-specific settings in Cognigy.

Use the Management UI to configure users, organizations, quotas, password policies, data privacy settings, expiration values for sensitive data, and system messages.

The Management UI is especially useful for:

* System administrators, particularly in on-premises installations.
* Advanced users who need to manage multiple organizations or adjust global settings.

## Key Features

* **User and Organization Management**. Create, modify, remove users, and add them to specific organizations, which serve as isolated environments within Cognigy.
* **Resource Control through Quotas**. Set specific quotas for organizations to manage resource usage, including user seats, Projects, and messages. Quotas let you prevent overuse and ensure equitable resource allocation.
* **Security and Privacy Settings**. Change password policies and manage data privacy settings, ensuring compliance with security standards and protecting sensitive information through features like data masking and retention policies.
* **Multitenancy and Impersonation Features**. Cognigy supports multitenancy, enabling the management of multiple organizations under one installation, along with the ability to impersonate users for troubleshooting and assistance.

## Before you Begin

<AccordionGroup>
  <Accordion title="1. Check Availability">
    The Management UI is available for the following Cognigy installations:

    * **Cognigy SaaS** – the Management UI is managed by Cognigy technical support in most cases. Contact Cognigy support if you want to change any values.
    * **On-Premises** – the Management UI is available and can be used by the customer. For more information on how to install the Management UI, refer to the [Cognigy GitHub](https://github.com/Cognigy/cognigy-ai-helm-chart/blob/main/README.md#install-management-ui) repo.
  </Accordion>

  <Accordion title="2. Log in to the Management UI">
    To log in to the Management UI instance, you need the following information:

    | **Field**                | **Description**                                                                                                              |
    | ------------------------ | ---------------------------------------------------------------------------------------------------------------------------- |
    | Cognigy Backend Base URL | The base URL must follow this format: `https://api-<your installation domain>`. For example: `https://api-trial.cognigy.ai`. |
    | Service User             | The username provided by your Management UI administrator.                                                                   |
    | Service Password         | The password for the service user account, provided by your Management UI administrator.                                     |
  </Accordion>
</AccordionGroup>

## Manage Organizations

Organizations define the environment for users, Projects, and data in Cognigy. Use the Management UI to create and configure organizations, adjust quotas and policies, and manage data privacy and expiration settings. Each organization has its own limits, users, and settings, allowing for secure and flexible multi-tenant environments.

Each installation includes at least one organization, which is created automatically. This initial organization is named `defaultOrg-xxx`.

<AccordionGroup>
  <Accordion title="Create an Organization">
    1. In the Management UI, select **Organisations** from the left-side menu, then click <img alt="create organization" /> in the bottom-right corner.
    2. In the **Create new Organisation** window, fill in the following fields:
       * **Name** – enter the name of the organization.
       * **CXone Business Unit ID** – this field is optional. Enter the Business Unit ID if applicable for NiCE Cognigy billing.
       * **CXone Tenant ID** – this field is optional. If you want users to log in to Cognigy via their NiCE CXone accounts, enter the Tenant ID for your NiCE CXone instance. This way, users can log in via single sign-on (SSO), so they don't need a separate Cognigy username or password. Each Tenant ID is tied to a single Cognigy organization. When users log in via SSO, they can access only the organization associated with that Tenant ID.<br />
         Specifying a Tenant ID allows NiCE CXone users to log in to Cognigy under the **Automation** category in the app selector of the NiCE CXone interface.
         This option is available only if the Cognigy application is activated in NiCE CXone.
         For more information, refer to the [NiCE CXone](https://help.nicecxone.com/content/cognigyai/enablecognigyaiandssoaccess.htm#Virtual) documentation.
    3. Click **Create**. The organization will be created with the default settings and appear in the list of organizations.
  </Accordion>

  <Accordion title="View Organizations">
    After creating an organization, you can see the organization in the list of organizations.

    | **Field**                 | **Description**                                                                                                                                                                                                                              |
    | ------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | **General**               |                                                                                                                                                                                                                                              |
    | Name                      | The name and version of the organization.                                                                                                                                                                                                    |
    | Enabled                   | Whether the organization is enabled.                                                                                                                                                                                                         |
    | Business Unit ID          | The ID used for NiCE Cognigy billing.                                                                                                                                                                                                        |
    | Tenant ID                 | The NiCE CXone Tenant ID that enables SSO login to a single Cognigy organization.                                                                                                                                                            |
    | Max Users                 | The maximum number of users allowed.                                                                                                                                                                                                         |
    | Max Projects              | The maximum number of Projects allowed.                                                                                                                                                                                                      |
    | Max Channels per Project  | The maximum number of channels allowed per Project.                                                                                                                                                                                          |
    | Max Messages per Day      | The maximum number of messages allowed per day.                                                                                                                                                                                              |
    | Billing Timezone          | The time zone used for billing.                                                                                                                                                                                                              |
    | **Data Privacy Settings** |                                                                                                                                                                                                                                              |
    | Apply to all projects     | Whether the data privacy settings are applied across all Projects.                                                                                                                                                                           |
    | Contact Profiles          | Whether Contact Profiles are created for persisted user data.                                                                                                                                                                                |
    | Collect Analytics         | Whether analytics data is collected for all Endpoints.                                                                                                                                                                                       |
    | Mask Sensitive Analytics  | Whether sensitive analytics data is masked.                                                                                                                                                                                                  |
    | Mask Sensitive Logging    | Whether sensitive logging data is masked.                                                                                                                                                                                                    |
    | Collect Conversations     | Whether the conversation history is collected.                                                                                                                                                                                               |
    | Ignore List               | A list of Projects that use Endpoint-specific privacy settings instead of organization-level settings.                                                                                                                                       |
    | **NiCE CXone Info**       | This section is optional and appears only if **Business Unit ID** or **Tenant ID** is configured for the organization.                                                                                                                       |
    | CXone Business Unit ID    | This field is optional. The ID used for NiCE Cognigy billing purposes. For more information, refer to the [Create an Organization](#create-an-organization) section.                                                                         |
    | CXone Tenant ID           | This field is optional. The ID used for SSO login via NiCE CXone. For more information, refer to the [Create an Organization](#create-an-organization) section.                                                                              |
    | **Conversation Counter**  |                                                                                                                                                                                                                                              |
    | Conversation Counter      | The total number of conversations for a specific period. Conversation counts are calculated overnight. For this reason, data for the current day isn't included. Historical data for conversation counts is available only from 2025 onward. |
  </Accordion>

  <Accordion title="Change Organization Information">
    To change the organization information, such as Billing Time Zone, Business Unit ID, or Tenant ID, select **Organisations** from the left-side menu. From the organization list, select the organization you want to change. On the **Organisation Details** page, click <img alt="menu" /> and select **Change Organization Information**.

    | **Setting**       | **Description**                                                                                                                                                                |
    | ----------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Billing Time Zone | The time zone used for billing. This setting affects conversation counts. Before changing it, make sure you understand how [conversations are billed](/ai/administer/billing). |
    | Business Unit ID  | The ID used for NiCE Cognigy billing.                                                                                                                                          |
    | Tenant ID         | The NiCE CXone Tenant ID that enables SSO login to a single Cognigy organization.                                                                                              |
  </Accordion>

  <Accordion title="Change Organization Quotas">
    To change organization quotas, select **Organisations** from the left-side menu. From the organization list, select the organization you want to change. On the **Organisation Details** page, click <img alt="menu" /> and select **Change Organisation Quotas**.

    <Note>
      Setting a quota to 0 disables it, meaning there is no limit.
    </Note>

    | **Quota Name**           | **Description**                                                                                                                                                                                                                                                                                 |
    | ------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Max users                | The maximum number of users that can be created within the organization.                                                                                                                                                                                                                        |
    | Max projects             | The maximum number of Projects that can be created within the organization.                                                                                                                                                                                                                     |
    | Max channels per project | The maximum number of different Endpoints allowed in a Project. For example, if this is set to 1, you can create an unlimited number of Amazon Alexa Endpoints, but you will receive an error when trying to add a WhatsApp Endpoint. The restriction applies to Endpoint types, not instances. |
    | Max messages per day     | The maximum number of messages that can be processed per day. This quota isn't currently enforced but may be subject to change in the future.                                                                                                                                                   |
    | Max knowledge chunks     | The maximum number of Knowledge Chunks that can be created within the organization.                                                                                                                                                                                                             |
  </Accordion>

  <Accordion title="Change Password Policy">
    The password policy defines the rules that users must follow when creating or updating their passwords. It helps ensure that passwords are strong and secure.
    To change the password policy for an organization, select **Organisations** from the left-side menu. From the organization list, select the organization you want to change. On the **Organisation Details** page, click <img alt="menu" /> and select **Change Password Policy**.

    | **Rule**                               | **Description**                                                                                                                                                     |
    | -------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Minimum Length                         | The minimum number of characters required in the password.                                                                                                          |
    | Maximum Length                         | The maximum number of characters allowed in the password.                                                                                                           |
    | Minimum Lowercase Characters           | The minimum number of lowercase letters required in the password.                                                                                                   |
    | Minimum Uppercase Characters           | The minimum number of uppercase letters required in the password.                                                                                                   |
    | Minimum Special Characters             | The minimum number of special characters, for example, `!`, `@`, `#`, required in the password.                                                                     |
    | Maximum Identical Characters           | The maximum number of identical characters allowed in the password.                                                                                                 |
    | Failed Login Attempts Before Disabling | Number of failed login attempts allowed before the account is disabled. Set to `0` to disable this feature. Disabled users can be re-enabled via the Management UI. |

    Clicking **Save & Enforce** logs out all users in the organization. Upon re-login, they must change their password to comply with the new policy.
  </Accordion>

  <Accordion title="Change Expiration Values">
    The expiration values define how long certain sensitive data is retained in the system before it is deleted. To change the expiration values for an organization, select **Organisations** from the left-side menu. From the organization list, select the organization you want to change. On the **Organisation Details** page, click <img alt="menu" /> and select **Change Expiration Values**.

    | **Setting**                      | **Description**                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
    | -------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Contact Profiles Expiration      | The duration in minutes that a Contact Profile record is retained in the Analytics database.                                                                                                                                                                                                                                                                                                                                                                                                                      |
    | Session Expiration               | The duration in minutes that a session remains active in the Analytics database after the last message is sent. Each new message resets this timer. Alternatively, session expiration can be configured using the `MAX_SESSION_STATE_TTL_IN_MINUTES` environment variable or per [Endpoint](/ai/agents/deploy/endpoints/session-management). If both a global value (configured via the Management UI or environment variable) and an Endpoint-specific value are defined, the shorter duration takes precedence. |
    | Conversation Expiration          | The duration in minutes that a [Conversation](/ai/agents/analyze/collecting-data) record is retained in the Analytics database.                                                                                                                                                                                                                                                                                                                                                                                   |
    | Analytics Expiration             | The duration in minutes that an [Analytics](/ai/agents/analyze/collecting-data) record is retained in the Analytics database. This expiration also applies to data for [Goals](/ai/agents/analyze/goals-and-tasks/overview).                                                                                                                                                                                                                                                                                      |
    | Insights Session Data Expiration | The duration in minutes that an Insights Session record is retained in the Analytics database.                                                                                                                                                                                                                                                                                                                                                                                                                    |
    | Steps Data Expiration            | The duration in minutes that a [Step](/ai/agents/analyze/collecting-data#steps) record is retained in the Analytics database.                                                                                                                                                                                                                                                                                                                                                                                     |

    For more information about data expiration:

    * In Cognigy, refer to [Data Expiration](/ai/agents/analyze/data-expiration) in the Cognigy documentation.

    * In Insights, refer to [Data Expiration](/insights/data-management/data-expiration) in the Insights documentation.
  </Accordion>

  <Accordion title="Change Data Privacy Settings">
    The data privacy settings define how user data is handled within the organization. To change the data privacy settings for an organization, select **Organisations** from the left-side menu. From the list, select the organization you want to change. On the **Organisation Details** page, click <img alt="menu" /> and select **Change Data Privacy Settings**.

    | **Setting**                  | **Description**                                                                                                       |
    | ---------------------------- | --------------------------------------------------------------------------------------------------------------------- |
    | Enable Data Privacy Settings | The data privacy settings applied to all Projects within the organization.                                            |
    | Contact Profiles             | The creation of individual Contact Profiles to store persisted user data across sessions.                             |
    | Collect Analytics            | The collection of analytics data from all Endpoints.                                                                  |
    | Mask Sensitive Analytics     | The masking of all sensitive analytics data collected from Endpoints.                                                 |
    | Mask Sensitive Logging       | The masking of all sensitive logging information generated by Endpoints.                                              |
    | Collect Conversations        | The collection of conversation history from all Endpoints.                                                            |
    | Ignore List                  | The list of Projects that override organization-wide settings with their own Endpoint-specific data privacy settings. |
  </Accordion>

  <Accordion title="Change Organization Products">
    To change the products associated with an organization, select **Organisations** from the left-side menu. From the organization list, select the organization you want to change. On the **Organisation Details** page, click <img alt="menu" /> and select **Change Organisation Products**.

    Select or deselect Enable AI Ops Center if you want to activate or deactivate Ops Center for your organization, then click **Save**.
  </Accordion>
</AccordionGroup>

## Manage Users

Create and manage users within your Cognigy installation.
You can add users to organizations, view user details, impersonate users for troubleshooting, and enforce security settings like password policies or impersonation restrictions.

<AccordionGroup>
  <Accordion title="Create a User">
    1. In the Management UI, ensure that you have created an organization where you want to add a user.
    2. You can open a **Create new User** window by using one of the following options:
       * Select **Users** from the left-side menu, then click <img alt="menu" /> in the bottom-right corner.
       * Select **Organisations** from the left-side menu, select the organization where you want to add a user, then click **Create User** in the upper-right corner.
    3. In the **Create new User** window, fill in the following fields:
       * **Name** – enter the name of the user.
       * **Email** – enter the email address of the user. This is used for login and notifications.
       * **Organisation** – select the organization to which the user will be added. If you have opened the **Create new User** window from the **Organisations** page, the organization will be preselected.
       * **Password** – enter a password for the user. You can generate a random password by clicking <img alt="change password" />.
    4. Click **Create**. The user will be created with the `admin` role by default, which grants them full access to the organization. You can change the user's role later on the [Access Control](/ai/administer/access/admin-center/access-control) page for the organization where the user was added.

    You can add a user with the same email to multiple organizations. For details on how users access each organization, refer to the [Login for Users in Multiple Organizations](/ai/administer/installation/access-to-cognigy-ai#multiple-organizations) page.
  </Accordion>

  <Accordion title="View User Details">
    After creating a user, you can view them in the user list. Select an organization to view its details.

    | **Field**                 | **Description**                                                                                                                                               | **Example**                                         |
    | ------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------- |
    | **General Info**          |                                                                                                                                                               |                                                     |
    | Name                      | The user's display name.                                                                                                                                      | John Doe                                            |
    | Email                     | The user's login email address.                                                                                                                               | [john-doe@cognigy.com](mailto:john-doe@cognigy.com) |
    | Organisation              | The organization the user belongs to.                                                                                                                         | ACME                                                |
    | Roles                     | Roles assigned to the user that define their permissions.                                                                                                     | `admin`                                             |
    | **Meta Info**             |                                                                                                                                                               |                                                     |
    | Created At                | The date the account was created (day-month-year format).                                                                                                     | 10/01/2025                                          |
    | Created By                | The user or system account that created this user.                                                                                                            | system admin                                        |
    | Last Changed At           | The date the user record was last updated (day-month-year format).                                                                                            | 10/01/2025                                          |
    | Last Changed By           | The user or system account that last modified the user details.                                                                                               | system admin                                        |
    | **Status**                |                                                                                                                                                               |                                                     |
    | Accepted Terms of Service | Whether the user has accepted the Terms of Service.                                                                                                           | Yes                                                 |
    | User Enabled              | Whether the user account is currently active.                                                                                                                 | Yes                                                 |
    | Organisation Enabled      | Whether the user's organization is currently enabled in the system.                                                                                           | Yes                                                 |
    | **CXone Info**            | This section is optional and appears only if the user is assigned to a Cognigy organization linked to a NiCE CXone instance.                                  |                                                     |
    | CXone Email Address       | The user's email address (`cxoneEmail`) in the NiCE CXone instance. The value of this field is added automatically and can't be changed in the Management UI. |                                                     |
    | CXone ID                  | The user ID (`cxoneId`) in the NiCE CXone instance. The value of this field is added automatically and can't be changed in the Management UI.                 |                                                     |
  </Accordion>

  <Accordion title="Impersonate a User">
    When impersonating a user, you can log in to the system as the selected user without knowing their actual account password.

    To impersonate a user, select a user from the list, then click **Impersonate** in the upper-right corner of the **User Details** page.
    Make sure to log out after you have finished your work.
  </Accordion>

  <Accordion title="Block User Impersonation">
    To prevent unauthorized access and protect user privacy, blocking impersonation improves security and helps prevent account misuse.

    You can block user impersonation as follows:

    * For Cognigy SaaS installations, contact [Cognigy technical support](/help/get-help).
    * For on-premises installations, use the `BLOCK_IMPERSONATE_USER_ORG_WHITELIST` feature flag. You can block user impersonation for:
      * **One organization** — enter the `organisationId` in the `BLOCK_IMPERSONATE_USER_ORG_WHITELIST` feature flag.
      * **More than one organization** — enter the `organisationId` values separated by commas, for example, `BLOCK_IMPERSONATE_USER_ORG_WHITELIST=63c6af010aa7a0eadd88edbd,63c6af010aa7a0eadd88edbe`.
      * **All organizations** — set `BLOCK_IMPERSONATE_USER_ORG_WHITELIST=*`.
  </Accordion>

  <Accordion title="Other Operations">
    You can also disable or delete a user by selecting them from the user list, clicking <img alt="menu" />, and selecting the respective option on the **User Details** page.
  </Accordion>
</AccordionGroup>

## System Message

Use the System Message feature in the Management UI to notify users across all organizations about important events, such as scheduled maintenance or platform updates.
This feature is especially useful for notifying users in on-premises installations about planned downtimes or system-wide changes.

<Accordion title="Create a System Message">
  To create a system message:

  1. In the Management UI, select **System Message** in the left-side menu, then click <img alt="create system message" /> in the bottom-right corner.
  2. Specify the expiration date and time in the **Until when** field.
  3. Enter your message in the **Message** field. This message will be displayed to all users in the organization.
  4. *(Optional)* Activate **Show on Login Screen** to display the message before users log in.

  <Note>
    Only one system message can be active at a time.
  </Note>
</Accordion>


# Members
Source: https://docs.cognigy.com/ai/administer/access/members



<a href="/release-notes/4.75"><Badge>Updated in 4.75</Badge></a>

The *Members* interface in a Cognigy Project allows Project admins to assign and manage Project-specific roles. This interface allows fine-grained control over who can create, edit, view, or update different parts of a Project, complementing the [global roles](/ai/administer/access/admin-center/access-control) set in the Access Control.

## Project Roles

Project roles let you control what users can do within a specific Project. To manage permissions across the whole organization, use [global roles](/ai/administer/access/admin-center/access-control).

<Note>
  Project roles assigned to users determine what actions a user can perform in the UI and through API endpoints, ensuring consistent access control across both interfaces.
</Note>

Each permission works as follows:

* **Create** – create a new resource.
* **Read** – view or retrieve the resource and its details.
* **Update** – view and change an existing resource's data or settings.
* **Delete** – permanently remove the resource.

Having full permissions means a user can create, read, update, and delete the resource.

<ExpandableTable />

## Add a Member

1. In the left-side menu of the Project, go to **Manage > Members**, then click **+ Add new Member** in the upper-left corner.
2. In the **New Member** pane, select the user you want to add to the Project. Click **Add**.
3. In the **Project Roles** section, click **+** to select one or more roles to assign. You can assign multiple roles to a user.
4. *(Optional)* In the **Local restrictions** section, assign a user to specific locales by selecting the **User is restricted to the following locales** option and the locales.

## More Information

* [Access Control](/ai/administer/access/admin-center/access-control)


# Project Settings
Source: https://docs.cognigy.com/ai/administer/access/project-settings



<a href="/release-notes/2026.15"><Badge>Updated in 2026.15</Badge></a>

*Project Settings* are a set of configuration options that let you customize how your AI Agents operate. These settings apply to all resources within the Project. However, some settings can be overridden at the Flow level.

In Project settings, you can:

* Define default values for your Project, such as currency, time zone, and thresholds.
* Control analytics and data collection.
* Configure NLU behavior, including Intent recognition, Slot extraction, and confirmation logic.
* Connect your AI Agent to external services such as translation providers and speech providers.
* Select models for Knowledge AI and Generative AI features.
* Configure the Live Agent inbox.

To adjust Project settings, select a Project on the [Main page](/ai/overview/user-interface/main-page), then go to **Manage > Settings**.

## List of Settings

<AccordionGroup>
  <Accordion title="Generative AI Settings">
    | Setting                       | Description                                                                                                                                                                                                                                                                                                                                                                                             |
    | ----------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Enable Generative AI Features | Allows AI Agents to use Generative AI features. The setting is enabled by default.                                                                                                                                                                                                                                                                                                                      |
    | AI Agent Node                 | Allows you to select the [model](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) for [AI Agent Node](/ai/agents/develop/node-reference/ai/ai-agent) interactions.                                                                                                                                                                                                                          |
    | LLM Prompt Node               | Allows you to select the [model](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) for [LLM Prompt Node](/ai/agents/develop/node-reference/service/llm-prompt) responses.                                                                                                                                                                                                                    |
    | AI-Enhanced Outputs           | Allows you to select the [model](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) that [rephrases AI Agent outputs](/ai/agents/develop/gen-ai-and-llms/generative-ai#use-run-time-features).                                                                                                                                                                                                |
    | Design-Time LLM Features      | Allows you to select the [model](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) used for [design-time features](/ai/agents/develop/gen-ai-and-llms/generative-ai#design-time-generative-ai-features), such as Flow, Lexicon, Intent sentence generation. By default, the [Cognigy platform-provided LLM](/ai/agents/develop/gen-ai-and-llms/providers/cognigy-platform-provided) is used. |
    | Conversation Analyzer         | Allows you to select the [model](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) used by the Conversation Analyzer in [Cognigy Insights](/insights/overview) to analyze session transcripts. If no model is selected, the Project's default LLM is used.                                                                                                                                   |
  </Accordion>

  <Accordion title="Voice Preview Settings">
    | Setting           | Description                                                                                                                                                                                                                                                   |
    | ----------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Speech Provider   | Allows you to select a speech provider: Microsoft Azure Speech Services, Google Speech Service, Amazon Polly, Deepgram Speech Services, or ElevenLabs Speech Services. For more information, refer to [Voice Preview Feature](/ai/agents/test/voice-preview). |
    | Speech Connection | Allows you to configure the connection for the chosen provider. For more information, refer to [Voice Preview](/ai/agents/test/voice-preview).                                                                                                                |
  </Accordion>

  <Accordion title="Knowledge AI Settings">
    | Setting                                      | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
    | -------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Knowledge Search                             | Allows you to select the [model](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) for knowledge retrieval. This model is used to find the most relevant knowledge that match the user's request. For more information, refer to [Which Models to Choose?](/ai/agents/develop/knowledge-ai/overview#which-models-to-choose).                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
    | Answer Extraction                            | Allows you to select the [model](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) for answer generation. After relevant knowledge is retrieved, this model extracts or generates the most precise answer from the content. For more information, refer to [Which Models to Choose?](/ai/agents/develop/knowledge-ai/overview#which-models-to-choose).                                                                                                                                                                                                                                                                                                                                                                                                                              |
    | **Document Processing**                      |                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
    | Content Parser                               | Allows you to choose how documents are parsed and prepared for Knowledge AI. Select one of the following options:<ul><li>**Default** – advanced parser provided by Cognigy. For more information, refer to [Advanced Parser](/ai/agents/develop/knowledge-ai/knowledge-source/text-extraction/advanced-parser).</li><li>**Basic** – lightweight parser for simple documents. For more information, refer to [Basic Parser](/ai/agents/develop/knowledge-ai/knowledge-source/text-extraction/basic-parser).</li><li>**Azure AI Document Intelligence** – advanced parser that uses Microsoft Azure AI for structured document processing. For more information, refer to [Azure AI Document Intelligence](/ai/agents/develop/knowledge-ai/knowledge-source/text-extraction/azure-ai).</li></ul> |
    | Connection to Azure AI Document Intelligence | Allows you to configure a connection to Azure AI Document Intelligence if you selected **Azure AI Document Intelligence** as the content parser. Required for integrating with Azure services.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
  </Accordion>

  <Accordion title="Translation Settings">
    | Setting                       | Description                                                                                                                                                                                                                                       |
    | ----------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Translation Provider          | Allows you to select a translation provider: Microsoft Translator, Google Cloud Translation, or DeepL Translate Pro. For more information, refer to [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings). |
    | Translation Connection        | Allows you to configure a connection for the selected translation provider. For more information, refer to [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings).                                          |
    | Request Retries               | Sets how often Cognigy attempts to call the translation provider if an error occurs during the initial request. For example, if you set it to `3`, Cognigy tries three times before giving up on a translation request.                           |
    | Request Timeout               | Sets how long Cognigy waits for a response from the translation provider after making a request, in milliseconds. If the provider doesn't respond within this time frame, it's considered a timeout.                                              |
    | Sentence Cache Expiry Timeout | Sets how long translations are kept in the cache before they are considered outdated and removed. The default value is `86,400` seconds, which corresponds to 1 day. For no caching, set it to `0`.                                               |
  </Accordion>

  <Accordion title="Data Redaction Settings">
    Allows configuring data redaction settings for the Project. For more information, refer to [Data Redaction](/ai/platform-features/data-redaction).
  </Accordion>

  <Accordion title="Currency Settings">
    | Setting               | Description                                                                                                                                                          |
    | --------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Set Standard Currency | Defines the default currency for the [Goals](/ai/agents/analyze/goals-and-tasks/goals) feature when setting up the **Revenue Earnings** or **Cost Savings** metrics. |
  </Accordion>

  <Accordion title="NLU Settings">
    | Setting                              | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
    | ------------------------------------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | **Embedding Model**                  |                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | Embedding Model Provider             | Allows you to select **Cognigy** to use the Cognigy NLU model, or to connect a third-party embedding model provider for intent recognition. The default value is `Cognigy`. If you want to connect to a third-party embedding model provider, refer to [Intent Recognition with External Embedding Model](/ai/platform-features/nlu/external/external-nlu-intent-recognition).                                                                                                                                                                                                                                                                                                                                             |
    | Yes/No logic                         | Adds [Yes/No logic](/ai/platform-features/nlu/intents/yes-no-intents) in your Flows to confirm and reject an [Intent](/ai/platform-features/nlu/intents/overview), or the **Yes/No** type in [Question Nodes](/ai/agents/develop/node-reference/basic/question). The setting can be overridden at the [Flow](/ai/agents/develop/projects-and-flows/overview#settings) level.                                                                                                                                                                                                                                                                                                                                               |
    | Positive confirmation words          | Adds customized responses to confirm an [Intent](/ai/platform-features/nlu/intents/overview) and [Question Nodes](/ai/agents/develop/node-reference/basic/question). The setting can be overridden at the [Flow](/ai/agents/develop/projects-and-flows/overview#settings) level.                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | Negative confirmation words          | Adds customized responses to reject an [Intent](/ai/platform-features/nlu/intents/overview) and [Question Nodes](/ai/agents/develop/node-reference/basic/question). The setting can be overridden at the [Flow](/ai/agents/develop/projects-and-flows/overview#settings) level.                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | **Timezone**                         |                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | Select Timezone                      | Sets the time zone in which AI Agents operate.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | **Intent Mapper**                    |                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | Enable Case-Sensitive Intent Mapping | Allows the Intent Mapper to consider case sensitivity for all Flows. Disable the setting to ignore casing. After changing this setting, retrain all Flows via **Build > Flows > Train all Flows**. If **Train all Flows** isn't available, for version 2026.2 or earlier, set `FEATURE_TRAIN_ALL_PROJECT_FLOWS: "true"` in `values.yaml` for on-premises installations, or upgrade to version 2026.3 for SaaS. Note that case-sensitive mapping isn't supported for English locales (US, Indian, UK, Canadian, Australian), use the universal Locale instead.                                                                                                                                                              |
    | **Thresholds**                       |                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | Intent Threshold                     | Sets the confidence threshold and reconfirmation threshold:<ul><li>**Confidence Threshold** – the score from which an Intent is considered confirmed if a confirmation sentence is set. The default value is `0.4`.</li><li>**Reconfirmation Threshold** – the lower confidence bound at which an Intent is reconfirmed. The default value is `0.2`.</li></ul> For example, if a user input scores `0.3`, the AI Agent asks for reconfirmation; if the user input scores `0.5`, it will be confirmed automatically. For more information, refer to [Thresholds](/ai/platform-features/nlu/intents/thresholds). The setting can be overridden at the [Flow](/ai/agents/develop/projects-and-flows/overview#settings) level. |
    | **System Slots**                     |                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | Enable Age Slots                     | Enables the recognition of age-related terms. For more information, refer to [System-Defined Slots](/ai/platform-features/nlu/slots/system-defined). The setting can be overridden at the [Flow](/ai/agents/develop/projects-and-flows/overview#settings) level.                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | Enable Date Slots                    | Enables the recognition of date-related terms. For more information, refer to [System-Defined Slots](/ai/platform-features/nlu/slots/system-defined). The setting can be overridden at the [Flow](/ai/agents/develop/projects-and-flows/overview#settings) level.                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
    | Enable Distance Slots                | Enables the recognition of distance-related terms. For more information, refer to [System-Defined Slots](/ai/platform-features/nlu/slots/system-defined). The setting can be overridden at the [Flow](/ai/agents/develop/projects-and-flows/overview#settings) level.                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
    | Enable Duration Slots                | Enables the recognition of duration-related terms. For more information, refer to [System-Defined Slots](/ai/platform-features/nlu/slots/system-defined). The setting can be overridden at the [Flow](/ai/agents/develop/projects-and-flows/overview#settings) level.                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
    | Enable Email Slots                   | Enables the recognition of email addresses. For more information, refer to [System-Defined Slots](/ai/platform-features/nlu/slots/system-defined). The setting can be overridden at the [Flow](/ai/agents/develop/projects-and-flows/overview#settings) level.                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | Enable Money Slots                   | Enables the recognition of monetary values. For more information, refer to [System-Defined Slots](/ai/platform-features/nlu/slots/system-defined). The setting can be overridden at the [Flow](/ai/agents/develop/projects-and-flows/overview#settings) level.                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | Enable Number Slots                  | Enables the recognition of numerical values. For more information, refer to [System-Defined Slots](/ai/platform-features/nlu/slots/system-defined). The setting can be overridden at the [Flow](/ai/agents/develop/projects-and-flows/overview#settings) level.                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | Enable Percentage Slots              | Enables the recognition of percentage values. For more information, refer to [System-Defined Slots](/ai/platform-features/nlu/slots/system-defined). The setting can be overridden at the [Flow](/ai/agents/develop/projects-and-flows/overview#settings) level.                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | Enable Temperature Slots             | Enables the recognition of temperature values. For more information, refer to [System-Defined Slots](/ai/platform-features/nlu/slots/system-defined). The setting can be overridden at the [Flow](/ai/agents/develop/projects-and-flows/overview#settings) level.                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
    | Enable URL Slots                     | Enables the recognition of URLs. For more information, refer to [System-Defined Slots](/ai/platform-features/nlu/slots/system-defined). The setting can be overridden at the [Flow](/ai/agents/develop/projects-and-flows/overview#settings) level.                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
    | **Analytics**                        |                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | Collect                              | Enables collecting analytics from conversations.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | Allow Data Only                      | Enables collecting analytics for messages containing only JSON data, without text input.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
  </Accordion>

  <Accordion title="Live Agent Settings">
    | Setting                            | Description                                                           |
    | ---------------------------------- | --------------------------------------------------------------------- |
    | Overwrite Cognigy Live Agent Inbox | Allows you to enable overriding the default Cognigy Live Agent inbox. |
  </Accordion>
</AccordionGroup>

## More Information

* [Admin Center](/ai/administer/access/admin-center/overview)


# Billing
Source: https://docs.cognigy.com/ai/administer/billing



Cognigy supports two different billing models, depending on your license:

* [Cognigy billing](#cognigy-billing). Applies to customers using Cognigy products standalone.
* [NiCE CXone Cognigy billing](#nice-cxone-cognigy). Applies to customers using Cognigy products as part of NiCE CXone.

## Cognigy Billing

If you have a license agreement with Cognigy, your usage of Cognigy products is billed according to the Cognigy billing model.

Cognigy products have *three different ways* to charge customers:

* For *conversations* processed with Cognigy.
* For *concurrent lines* for calls handled with Cognigy Voice Gateway.
* For *knowledge queries* and *knowledge chunks* used in Knowledge AI.

### Billable Conversations

Cognigy charges customers based on the number of *billable conversations* processed through the platform.

A *billable conversation* is an *interaction between a user and an AI Agent or a human agent that flows through the Cognigy platform*. A conversation begins with the initial *input* via an Endpoint to the Cognigy platform and ends when the end user stops interacting. A billable conversation is limited to up to 50 user inputs within a 24-hour period.

Sending more than 50 inputs or interacting with the Cognigy platform for more than 24 hours will initiate subsequent billable conversations.

#### Common Scenarios

<AccordionGroup>
  <Accordion title="Scenario 1">
    **An end user sends exactly 50 user inputs within a period of less than 24 hours.**

    An end user begins interacting with a Cognigy AI Agent and sends 50 user inputs. The entire interaction takes 15 minutes. The system will record *1 billable conversation* as the end user had fewer than 51 user inputs and only interacted with the system for 15 minutes.
  </Accordion>

  <Accordion title="Scenario 2">
    **An end user sends more than 50 user inputs within a period of less than 24 hours.**

    An end user starts interacting with a Cognigy AI Agent and sends 101 user inputs within 2 hours. The system will record *3 billable conversations* as three multiples of 50 inputs have been processed.
  </Accordion>

  <Accordion title="Scenario 3">
    **An end user starts interacting by sending a message, doesn't respond for a day, and then continues the interaction.**

    An end user begins interacting with a Cognigy AI Agent by sending a single message. The user then pauses the interaction for more than 24 hours and returns the next day by sending another message. The system will record *2 billable conversations* since they extend over multiple days and exceed a duration of 24 hours.
  </Accordion>
</AccordionGroup>

#### xApps

Cognigy xApps offer multi-modality interactivity with your end users. Note that xApps are a separately licensed product feature that requires a license agreement with Cognigy. To learn more about xApps, refer to the [documentation](/xApps/overview).

When data is submitted from an xApp back into the Cognigy platform, it is handled the same as a standard end user interaction, such as sending a chat input or speaking an utterance in the case of a voice AI Agent.

<Accordion title="Scenario 1">
  **An end user books a cinema ticket with a bot using xApps.**

  An end user begins interacting with a Cognigy-powered ticket booking chatbot by sending a chat input. The bot responds by requesting the first name, last name, and ticket category. The user answers all three questions by providing the respective data and then gets a floor plan via an xApp, through which they can select a seat. The user selects a seat and submits the selection. The system will record *a single billable conversation* with the following explanation:

  * The end user greets the chatbot - **one user input**
  * The AI Agent asks for the first name
  * The end user provides the first name - **second user input**
  * The AI Agent asks for the last name
  * The end user provides the last name - **third user input**
  * The AI Agent asks for the ticket category
  * The end user provides the ticket category by clicking on a quick reply - **fourth user input**
  * The AI Agent presents a floor plan via an xApp
  * The end user selects a seat using the xApp and submits the selection - **fifth user input**
</Accordion>

#### Additional Billable Conversations

Additional charges apply if you use the Endpoint [Transformers](/ai/for-developers/transformers/overview), such as:

* Input Transformer
* Inject Transformer
* Notify Transformer

If an Endpoint Transformer execution results in a `falsy` value (`null`, `undefined`, or `false`), the message processing pipeline stops, and no Flow is executed. In such cases, Cognigy will charge an additional billable conversation for every 50 Endpoint Transformer executions.

Note that additional billing will only apply *if no Flow is executed* after your Endpoint Transformer runs.

### Calls

Cognigy offers the ability to build voice experiences and AI Agents that you can deploy to your telephony systems.
Note that you will need a separate license agreement with Cognigy for a product such as Cognigy Voice Gateway.
Customers can purchase different packages of *concurrent lines* to connect their voice-enabled AI Agents to their telephony systems. Reporting capabilities show the *peak number of concurrent lines* used on a specific day.

If a customer uses more concurrent lines than purchased on a given day, overage charges will apply for that day.

### Knowledge AI

Knowledge AI allows Cognigy customers to build RAG-based pipelines powered by Large Language Models.
Knowledge AI is a separately licensed product feature and requires a [license agreement](/ai/agents/develop/knowledge-ai/activate) with Cognigy.
To learn more about Knowledge AI, refer to the [documentation](/ai/agents/develop/knowledge-ai/overview).

Knowledge AI is licensed by *Knowledge Chunks* and *Knowledge Queries*. The former defines a *hard limit* on how much data can be ingested into the Knowledge AI solution, while the latter determines how often a query can be placed to retrieve a relevant Chunk of knowledge.

Cognigy offers various packages that include Knowledge Chunks and Queries. For Knowledge Queries, overages will apply if more queries are performed than what is included in the originally purchased package. The Knowledge Chunk limit can't be exceeded, as the product will display an error. Customers can purchase additional Knowledge Chunk packages if required.

Reporting capabilities show both the *number of Knowledge Chunks consumed* and the *number of Knowledge Queries performed*.

### Ops Center

Ops Center is a Cognigy product developed to manage the deployment of AI Agents at scale. It serves as the solution for gaining operational visibility and control over AI Agents.
Ops Center is a separately licensed product feature and requires a license agreement with Cognigy.
To learn more about Ops Center, refer to the [documentation](/ops-center/index).

## NiCE CXone Cognigy

If you are a NICE CXone customer, your usage of Cognigy products is billed per [conversation](#billable-conversations), similar to the Cognigy billing model. The key difference is that in the NICE CXone billing model, conversations are counted differently depending on whether they are digital or voice:

* **Digital** – up to 50 end-user messages within 24 hours count as 1 conversation. Only messages from the end user are counted; AI Agent responses don't count.
* **Voice** – up to 10 minutes per call count as 1 conversation.

Exceeding these limits results in additional conversations.

Products such as [xApps](#xapps), [Transformers](#additional-billable-conversations), [Knowledge AI](#knowledge-ai), and [Ops Center](#ops-center) are counted in NICE CXone billing the same way as in Cognigy billing.

### Common Scenarios

<AccordionGroup>
  <Accordion title="Scenario 1">
    **An end user sends up to 50 digital messages within 24 hours.**

    An end user interacts with an AI Agent and sends 50 messages in a single day.
    The system records *1 conversation* because the 50-message and 24-hour limits are not exceeded.
  </Accordion>

  <Accordion title="Scenario 2">
    **An end user sends more than 50 digital messages within 24 hours.**

    An end user sends 120 messages to an AI Agent in a single day.
    The system records *3 conversations* (50 + 50 + 20), as each set of up to 50 end-user messages counts as one conversation.
  </Accordion>

  <Accordion title="Scenario 3">
    **An end user engages in a voice call longer than 10 minutes.**

    An end user speaks with an AI Agent for 12 minutes.
    The system records *2 conversations* (10 minutes + 2 minutes) because each 10-minute interval counts as one conversation.
  </Accordion>

  <Accordion title="Scenario 4">
    **An end user starts a conversation that lasts more than 24 hours.**

    An end user begins a digital conversation with an AI Agent and continues the next day.
    The system records *2 conversations* (first 24 hours + remaining time) because the conversation exceeds the 24-hour period.
  </Accordion>
</AccordionGroup>


# Hotkeys
Source: https://docs.cognigy.com/ai/administer/hotkeys



This page lists the hotkeys for the Cognigy interface, covering the Flow editor, Interaction Panel, and Intents. Universal hotkeys should work everywhere the same way.
These hotkeys help you work faster and more efficiently with Cognigy.

## Limitations

* Any content stored in your clipboard expires after 10 minutes and becomes unavailable.

## Flow Editor

The hotkeys listed in the table apply to the Flow editor.

| Hotkey                                                                              | Description                            |
| ----------------------------------------------------------------------------------- | -------------------------------------- |
| <kbd>Ctrl+C</kbd><br /><kbd>Cmd+C</kbd><br /><kbd>Alt</kbd>+<kbd>\[drag/drop]</kbd> | Copies the selected Node.              |
| <kbd>Ctrl+V</kbd><br /><kbd>Cmd+V</kbd>                                             | Pastes the Node from the clipboard.    |
| <kbd>Ctrl+X</kbd><br /><kbd>Cmd+X</kbd>                                             | Cuts the selected Node.                |
| <kbd>Del</kbd>                                                                      | Deletes the selected Node.             |
| <kbd>Ctrl+S</kbd><br /><kbd>Cmd+S</kbd>                                             | Saves the Node currently being edited. |
| <kbd>Ctrl+Alt+F</kbd><br /><kbd>Cmd+Opt+F</kbd>                                     | Toggles the Node Search input field.   |
| <kbd>Ctrl+Alt+P</kbd><br /><kbd>Cmd+Opt+P</kbd>                                     | Opens the Voice Preview panel.         |

## Interaction Panel

The hotkeys listed in the table apply to the Interaction Panel.

| Hotkey                                  | Description                           |
| --------------------------------------- | ------------------------------------- |
| <kbd>Ctrl+B</kbd><br /><kbd>Cmd+B</kbd> | Resets a session on the **Chat** tab. |
| <kbd>Ctrl+Shift+F</kbd>                 | Toggles debug mode.                   |

## Intents

The hotkeys listed in the table apply to Intent example sentences. To use these hotkeys, activate the **Use extended keyboard accessibility features** setting in **My Profile**.

| Hotkey                                | Description                                                                                                          |
| ------------------------------------- | -------------------------------------------------------------------------------------------------------------------- |
| <kbd>Tab</kbd>                        | Allows you to navigate between sentences in Intent examples and view tooltips for each sentence.                     |
| <kbd>Tab</kbd>, then <kbd>Enter</kbd> | Allows you to navigate between sentences in Intent examples, view tooltips for each sentence, and delete a sentence. |


# Installation
Source: https://docs.cognigy.com/ai/administer/installation/about

Learn about Cognigy installation options, including Cognigy SaaS and on-premises deployments with Helm charts for Kubernetes environments.

Cognigy offers Cognigy and other Cognigy products as *Cognigy SaaS*.
Cognigy SaaS provides a fully managed cloud environment operated by Cognigy. The platform is maintained, secured, and updated by Cognigy, allowing organizations to focus on building and operating AI Agents without managing the underlying infrastructure.

In most cases, Cognigy SaaS runs in a shared cloud environment, where multiple organizations use the same Cognigy-managed infrastructure while their data and resources remain securely isolated.

<Warning>
  On-premises installations are no longer offered to new customers. However, current on-premises installations will keep operating as usual and will continue to receive updates.
</Warning>

## More Information

* [SaaS](/ai/administer/installation/ip-ranges-shared-environments)
* [Single Sign-on (SSO)](/ai/administer/installation/single-sign-on-saml2)


# Access to Cognigy
Source: https://docs.cognigy.com/ai/administer/installation/access-to-cognigy-ai



<a href="/release-notes/2026.18"><Badge>Updated in 2026.18</Badge></a>

Access to Cognigy allows you to log in to and log out of Cognigy.

## Prerequisites

* You need to have a Cognigy account.

## Login

### Limit Active Login Sessions

By default, there is no limit to the number of sessions you can start in Cognigy across different browsers. However, you can set a limit for the on-premises installations. Set the `REFRESH_TOKEN_MAX_AMOUNT_PER_USER` variable in `values.yml` to the number of active sessions you want to allow. For example, if you want to allow only one active session, set `REFRESH_TOKEN_MAX_AMOUNT_PER_USER=1`.

### Log in to Cognigy

You can access Cognigy's login page through the following login URL:

```txt theme={null}
https://<frontend-url>/login
```

`frontend-url` depends on your installation:

* For on-premises installations, `frontend-url` is a domain you defined to run Cognigy.
* For SaaS installations, `frontend-url` is provided to you during deployment.

For example, `frontend-url` for Cognigy's trial environment is `trial.cognigy.ai`, and the login URL is `https://trial.cognigy.ai/login`.

Select your preferred login method:

<Tabs>
  <Tab title="Log in directly to Cognigy">
    1. Go to the login page and enter your account credentials in the following fields:

       * **Email** — the email you provided to your Cognigy admin to register your Cognigy account.
       * **Password** — the password you received from your Cognigy admin. If you don't remember your password, you can [reset your password](#reset-password).

    2. *(Optional)* If you want to stay logged in after you close your session, activate the **Remember me** option.

    3. Click **Sign in**.
  </Tab>

  <Tab title="Log in to Cognigy via NiCE CXone Mpower">
    Depending on the Cognigy version, logging in via NiCE CXone Mpower works differently:

    <AccordionGroup>
      <Accordion title="Cognigy 2026.18.0 and later">
        NiCE CXone users can access Cognigy via their NiCE CXone accounts if the [Tenant ID](/ai/administer/access/management-ui#create-an-organization) from NiCE CXone Mpower is linked to a Cognigy organization.

        Linking the Tenant ID activates [single sign-on (SSO)](#log-in-with-single-sign-on), so users don't need a separate Cognigy username or password. Each Tenant ID is tied to a single Cognigy organization. When users log in via SSO, they can access only the organization associated with that Tenant ID.

        With a linked Tenant ID, NiCE CXone users can access Cognigy under the **Automation** category in the app selector of the NiCE CXone Mpower interface.
        This option is available only if the Cognigy application is activated in NiCE CXone Mpower.
        For more information, refer to the [NiCE CXone Mpower](https://help.nicecxone.com/content/cognigyai/enablecognigyaiandssoaccess.htm#Virtual) documentation.

        In CXone-integrated installations, opening a Cognigy URL directly also signs you in via SSO automatically:

        * With no active Cognigy session, you are redirected to the CXone login page instead of the Cognigy password form.
        * With an active CXone session, you are signed into Cognigy with no login page shown at all.

        You are then redirected back to the Cognigy page you originally requested.
      </Accordion>

      <Accordion title="Cognigy 2026.17.0 and earlier">
        NiCE CXone users can log in to Cognigy via their NiCE CXone accounts if the [Tenant ID](/ai/administer/access/management-ui#create-an-organization) from NiCE CXone Mpower is linked to a Cognigy organization.
        Linking the Tenant ID enables [single sign-on (SSO)](#log-in-with-single-sign-on), so users don't need a separate Cognigy username or password. Each Tenant ID is tied to a single Cognigy organization. When users log in via SSO, they can access only the organization associated with that Tenant ID.

        Specifying a Tenant ID allows NiCE CXone users to log in to Cognigy under the **Automation** category in the app selector of the NiCE CXone Mpower interface.
        This option is available only if the Cognigy application is activated in NiCE CXone Mpower.
        For more information, refer to the [NiCE CXone Mpower](https://help.nicecxone.com/content/cognigyai/enablecognigyaiandssoaccess.htm#Virtual) documentation.
      </Accordion>
    </AccordionGroup>
  </Tab>
</Tabs>

<Note>
  If you enter the wrong password, your account is temporarily locked. An error message appears on the screen and displays a countdown to unlock your account. The more times you enter the wrong password, the longer your account is locked.
</Note>

### Reset Password

1. Go to the login page and click **Forgot Password?**.
2. Enter your email in the **Email** field.
3. Click **Send link**.

You will receive an email with instructions to reset your password. You need to choose a password according to the [password policy](/ai/administer/access/management-ui#change-password-policy) your admin set up.

### Log in with Single Sign-on

In Cognigy, you can set up [single sign-on (SSO)](/ai/administer/installation/single-sign-on-saml2) and, if necessary, you can [enforce login with SSO](/ai/administer/installation/single-sign-on-saml2) for your users.

To log in with SSO, follow these steps:

<Tabs>
  <Tab title="Log in with Optional SSO">
    1. Go to the login page and click **Login with SSO**.
    2. Enter your email in the **Email** field.
    3. Click **Login with SSO**.
    4. Click **Sign in with SSO**.
  </Tab>

  <Tab title="Log in with Enforced SSO">
    1. Go to the login page and click **Login with SSO**.
    2. Enter your email in the **Email** field.
    3. Click **Sign in with SSO**.
  </Tab>
</Tabs>

### Multiple Organizations

In Cognigy, you can be part of more than one [organization](/ai/administer/access/management-ui#manage-organizations). Each organization has a different login URL. For each organization, the login URL follows this pattern:

```txt theme={null}
https://<frontend-url>/login?organisationId=<organization-id>
```

The `organization-id` parameter is a unique identifier with 24 characters that identifies an organization in Cognigy.

<Warning>
  If you are part of more than one organization, you need to include `organisationId` in the login URL. Otherwise, the login fails.
</Warning>

#### Email Notification for Failed Login

<a href="/release-notes/earlier-versions/cognigy-ai-pre-4.30"><Badge>Updated in 4.27.0</Badge></a>

If your login fails because you didn't include `organisationId` in the login URL, you receive an email with login links for each organization you belong to.

The login links in the email redirect you to the Cognigy login page with the `organisationId` parameter of the respective organization in the URL.

## Logout

### Log out of Cognigy

To log out of Cognigy, you have two options:

* In the upper-right corner of the **Main** page, select **User Menu > Logout**.
* In the bottom-left corner of the Project page, click **User Menu > Logout**.

### Automatic Logout

The **Automatic logout** feature allows you to set an inactivity period for users in your organization. If a user remains inactive longer than this period, they are automatically logged out.

By default, the inactivity period is 15 minutes. However, you can set a limit for on-premises installations. Set the following variables in `values.yml`:

* `FEATURE_ENABLE_AUTO_LOGOUT_ON_INACTIVITY=true`.
* `AUTO_LOGOUT_IDLE_TIME_IN_MIN` to the inactivity period you want in minutes.


# Cognigy Architecture
Source: https://docs.cognigy.com/ai/administer/installation/architecture

Explore Cognigy's cloud-native microservices architecture, with 30+ containerized services orchestrated by Kubernetes for scalability and reliability.

## Overview

Cognigy is a modern, cloud-native software solution based on a scalable microservice architecture. This modern architecture allows
us to leverage the compute-power of multiple servers for a single software product. Instead of running a single executable on a single server,
we have decomposed Cognigy into more than 30 separate executables. These smaller executables (hence the name microservice) despite being separated,
act as a single product — Cognigy — by interacting with each other through network calls.

<Frame>
  <img alt="Cognigy microservices architecture diagram" />
</Frame>

Each microservice runs as a containerized application on top of Kubernetes. Containers are a way of how processes can run in isolation
from each other and offer the ability to ship modern software products together with their dependencies, for instance, a runtime environment.
A program written in JavaScript might need a specific version of the Node.js runtime to function properly — containers allow packaging such
dependencies together with the actual application into one shippable unit. One of the more popular container runtimes and a set of other container-specific tools is [Docker](https://www.docker.com/).

Managing a high number of moving parts — the microservices mentioned above — is quite a challenge and requires an additional software product
that is called a container orchestrator.
[Kubernetes](https://kubernetes.io/) is such a container orchestrator dealing with the following tasks:

* Assigning containers to one of the available servers in the Kubernetes cluster respecting their individual hardware requirements and matching these against the available hardware on a server.
* Observing whether containers need to be restarted in case they crash.
* Updating containers to a newer version when a software update (for example, new version of Cognigy) has been shipped,
* Testing whether containerized processes still react by sending readiness- and liveness-probes.
* Scaling microservices according to load-indicators such as CPU load, memory profile or custom metrics by starting additional copies of services.

## Runtime & IDE

Cognigy itself is a product offering different types of functionality:

* IDE
* Runtime

While Cognigy Insights is actually also part of Cognigy, it plays a smaller role when talking about the Cognigy core product infrastructure. To learn more, read the [Insights](/insights/overview) documentation.

### IDE: Building your AI Agents

A service that allows you to create AI Agents and use all the components, such as Flows, Playbooks, Lexicons,
and Intents, is called the IDE (Integrated Development Environment) — a term often used in software development.
Once you are done with the first iteration of your AI Agent, a so-called Snapshot must be created to deploy your AI Agent into production.
A Snapshot is an immutable copy of your AI Agent containing all resources that make it up, such as your Flows, Lexicons, Intents, and Extensions.
Snapshots act as the connection to the second group of functionality: the Cognigy Runtime.

The following table shows the most relevant microservices forming the IDE:

| Microservice name      | Responsibilities                                                                                                                                                                                                                                                                                    |
| ---------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| service-api            | RESTful API to fully control Cognigy through HTTP calls. Full documentation in the form of an [OpenAPI spec](/ai/for-developers/developers/api-and-cli) is available. Also implements authentication & authorization. Integrations with Identity Providers via our Single Sign On integration.      |
| service-ui             | Serves the WebGUI for Cognigy                                                                                                                                                                                                                                                                       |
| service-webchat        | Serves a demo website which always contains the latest version of the Cognigy WebchatWidget embedded. Acts as a place for rapid prototyping.                                                                                                                                                        |
| service-resources      | Manages all resources within Cognigy - meaning Flows, Endpoints, Playbooks etc. by offering internal APIs which other Microservices can call.                                                                                                                                                       |
| service-custom-modules | Processes uploaded Cognigy Extensions and prepares them for usage.                                                                                                                                                                                                                                  |
| service-security       | Manages users, organizations and roles including permissions.                                                                                                                                                                                                                                       |
| service-task-manager   | Manages long running operations inside of Cognigy. Those operations include creating Cognigy Snapshots, Packaging assets for downloads, Processing your uploaded documents as part of our Knowledge AI solution. Our user interface exposes a task manager which shows all long-running operations. |

### Runtime: Running your AI Agents

Our Runtime is optimized for performance & throughput. All microservices belonging to the runtime are stress-tested on a regular basis and can scale horizontally. Cognigy's runtime is a highly distributed system and can leverage huge amounts of hardware if given. We run performance tests for every release of Cognigy in order to ensure that a new version is at least as fast & scalable as the last version.

The following table describes the responsibilities of key runtime microservices:

| Microservice name                               | Responsibilities                                                                                                                                                                                                                                 |
| ----------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| service-endpoint                                | Translates channel specific incoming requests to our internal representation, executes Transformers, counts billing relevant information and sends responses back to external channels.                                                          |
| service-app-session-manager                     | Acts as an API for Cognigy xApp sessions to connect. Distributes xApp relevant events to end-user devices and reacts on xApp data-inputs.                                                                                                        |
| service-ai                                      | Processes actual user messages, coordinates with our NLU microservices, waits for Extensions and HTTP Requests to external systems and processes Cognigy Script. This is the core of Cognigy.                                                    |
| service-nlp-ner                                 | Detects [System Slots](/ai/platform-features/nlu/slots/system-defined) in user utterances - these are things like dates, emails, currencies and numbers.                                                                                         |
| service-nlp-matcher                             | Detects [Lexicon Slots](/ai/platform-features/nlu/slots/system-defined) in user utterances using [Lexicons](/ai/platform-features/nlu/slots/user-defined/lexicon).                                                                               |
| service-nlp-classifier-score                    | Implement intent recognition based on [Intent models](/ai/platform-features/nlu/overview) our customers can train within our platform. There are variants for different languages (German, English, Korean, Japanese, Generic, XX).              |
| service-nlp-embedding                           | Runs our pre-trained word-embedding models which are part of our NLU system and drive [Intent models](/ai/platform-features/nlu/overview) our customers can create as part of our platform. Language dependent embedding services are available. |
| service-http & execution                        | Sends HTTP-Requests to external 3rd party systems when [HTTP Request Flow Node](/ai/agents/develop/node-reference/service/http-request) is used and executes Flow Nodes of our customers as part of [Extensions](/ai/for-developers/extensions). |
| service-profiles                                | Manages [Contact Profiles](/ai/agents/analyze/contact-profiles) and offers the ability to access them via Cognigy Script.                                                                                                                        |
| service-function-scheduler & function-execution | Schedules and executes the source-code of our customers in [Cognigy Functions](/ai/for-developers/functions).                                                                                                                                    |
| service-handover                                | Implements an interface for 3rd-party [handover providers](/ai/escalate/handover-reference/overview) like Ring Central Engage. Forwards end-users message to those external providers while conversations are in "handover".                     |


# Cognigy: Browser Requirements
Source: https://docs.cognigy.com/ai/administer/installation/browser-requirements



The Cognigy User Interface is compatible with the browsers listed below:

| Browser         | Versions                                      |
| --------------- | --------------------------------------------- |
| Google Chrome   | The newest version and two previous versions. |
| Microsoft Edge  | The newest version and two previous versions. |
| Mozilla Firefox | The newest version and two previous versions. |
| Safari          | The newest version and two previous versions. |

For security reasons, older versions of the listed web browsers are not supported by Cognigy.

## Troubleshooting

If you encounter any issues with browser compatibility, ensure the following before contacting the Cognigy support team:

* You are using the latest version of one of the Cognigy supported browsers.
* Your viewport is larger than 1280x800 pixels.
* For using Microsoft Japanese IME, your Edge or Chrome browser is on Windows 10 or higher, as there might be issues with other browsers and operating systems.


# Identity Providers
Source: https://docs.cognigy.com/ai/administer/installation/identity-providers

Explore the identity providers to set up single sign-on in Cognigy.

Cognigy integrates with popular identity providers (IdPs) to let users in your organization log in with [single sign-on (SSO)](/ai/administer/installation/single-sign-on-saml2) without the need for individual credentials.

You can use the following IdPs:

* [Auth0 with OpenID Connect](#auth0-openid-connect)
* [Auth0 with SAML 2.0](#auth0-saml-2-0)
* [Microsoft Entra ID](#microsoft-entra-id)
* [Google](#google)
* [Okta](#okta)
* [OneLogin](#onelogin)

For more information about about protocol-level details, see [Single Sign-on with SAML 2.0](/ai/administer/installation/single-sign-on-saml2).

## Prerequisites

* An account with the admin role in Cognigy.
* The organization ID of your Cognigy organization, referred to as `<organization-id>` in the examples. You can copy this ID from the [My Profile](/ai/administer/user-menu/my-profile) page by clicking <img alt="menu" /> **> Copy Organization ID**.
* The [API base URL](/ai/for-developers/developers/api-and-cli#api-base-url) of your Cognigy installation, referred to as `<api-base-url>` in the examples.
* Your Cognigy URL, referred to as `<cognigy-url>` in the examples. For example, for the trial environment, this URL is `https://trial.cognigy.ai/`.
* An [API key](/ai/administer/user-menu/my-profile) for sending configuration requests to the Cognigy API.
* Administrator access to the IdP tenant that you want to integrate with Cognigy.
* For Okta, you need an X.509 certificate. For more information, read Okta's [documentation about app certificate use](https://help.okta.com/oag/en-us/content/topics/access-gateway/about-application-certificate-use.htm).

## Limitations

* An organization can have only one SSO configuration. To replace an SSO configuration, delete it, then create another one. For more information, read [Change an SSO Configuration in Cognigy](/ai/administer/installation/single-sign-on-saml2#change-an-sso-configuration-in-cognigy).
* Only Microsoft Entra ID and OneLogin support single logout for Cognigy.

## Create an IdP App

To configure an IdP app, follow these steps:

<Tabs>
  <Tab title="Auth0 - OpenID Connect">
    1. Log in to the [Auth0 Dashboard](https://manage.auth0.com/) and select your tenant.

    2. In the left-side menu, go to **Applications > Applications** and click **+ Create Application**.

    3. Enter a name, for example, `Cognigy`, select **Regular Web Applications** as the app type, and click **Create**.

    4. Go to the **Settings** tab and copy the values from the fields in the **Basic Information** section. You will use them later to [configure the IdP in Cognigy](#configure-sso-in-cognigy-ai):

       * **Domain** — used in the `idpIssuer` parameter in the request payload.
       * **Client ID** — used in the `idpClientId` parameter in the request payload.
       * **Client Secret** — used in the `idpClientSecret` parameter in the request payload.

    5. On the **Settings** tab, configure the following using the API base URL and organization ID from the [Prerequisites](#prerequisites) section:

       * **Application Login URI** — enter `https://<api-base-url>/auth/oidc/callback/<organization-id>`.
       * **Allowed Callback URLs** — enter `https://<api-base-url>/auth/oidc/login/callback/<organization-id>`.
       * **Allowed Logout URLs** — enter `https://<api-base-url>/logout/<organization-id>`.
       * **Allowed Web Origins** — enter `https://*.cognigy.ai`.
       * **Allowed Origins (CORS)** — enter `https://*.cognigy.ai`.
  </Tab>

  <Tab title="Auth0 - SAML 2.0">
    <Steps>
      <Step title="Set Up an App">
        1. Log in to the [Auth0 Dashboard](https://manage.auth0.com/) and select your tenant.

        2. In the left-side menu, go to **Applications > Applications** and click **+ Create Application**.

        3. Enter a name, for example, `Cognigy`, select **Single Page Web Applications** as the app type, and click **Create**.

        4. Go to the **Settings** tab, configure the following using the API base URL and organization ID from the [Prerequisites](#prerequisites) section:

           * **Application Login URI** — enter `https://<api-base-url>/auth/saml/login/<organization-id>`.
           * **Allowed Callback URLs** — enter `https://<api-base-url>/auth/oidc/login/callback/<organization-id>`.

        5. On the **Addons** tab, activate the **SAML2 Web App** add-on. Auth0 opens the add-on settings dialog. From the **Usage** tab, copy and save the values from the following fields for later use in the API request payload to [configure the IdP in Cognigy](#configure-sso-in-cognigy-ai):

           * **Issuer** — used in the `idpIssuer` parameter in the request payload.
           * **Identity Provider Login URL** — used in the `idpLoginEndpoint` parameter in the request payload.
           * **Identity Provider Certificate** — used in the `idpCertificate` parameter in the request payload. Click **Download Auth0 certificate**. Base64-encode the certificate without line breaks. You can use the following terminal command:

             <CodeGroup>
               ```powershell PowerShell theme={null} theme={null}
               [Convert]::ToBase64String([IO.File]::ReadAllBytes(".\path-to-certificate"))
               ```

               ```bash Bash (macOS) theme={null} theme={null}
               base64 -i ./<path-to-certificate> | tr -d '\n'
               ```

               ```bash Bash (Linux) theme={null} theme={null}
               cat ./<path-to-certificate> | base64 -w0
               ```
             </CodeGroup>
      </Step>

      <Step title="Map Roles">
        For each user which you want to apply SSO to, set `user_metadata` to map the user's name and Cognigy role. To do so, follow these steps:

        1. Go to **User Management > Users**, select the user, scroll to the **Metadata** section, and paste the following JSON into the **user\_metadata** field:

           ```json theme={null}
           {
             "family_name": "<LAST NAME>",
             "given_name": "<FIRST NAME>",
             "role": "<COGNIGY_ROLE>"
           }
           ```

           The `role` value must match a [role in Cognigy](/ai/administer/access/admin-center/access-control), for example, `admin` or `base_role`.

        2. Go to **Actions > Library**, click **Create Action**, and configure the following:

        * **Name** — enter a name.
        * **Login / Post Login** — activate this option.
        * In the editor, paste the following code:

          ```javascript theme={null}
          exports.onExecutePostLogin = async (event, api) => {
            const userMetadata = event.user.user_metadata || {};

            api.samlResponse.setAttribute(
              "http://schemas.xmlsoap.org/ws/2005/05/identity/claims/nameidentifier",
              event.user.email
            );
            api.samlResponse.setAttribute("firstName", userMetadata.given_name || "");
            api.samlResponse.setAttribute("lastName", userMetadata.family_name || "");
            api.samlResponse.setAttribute("role", userMetadata.role || "");
          };
          ```

        3. Deploy the Action.
      </Step>
    </Steps>
  </Tab>

  <Tab title="Microsoft Entra ID">
    <Steps>
      <Step title="Set Up an App">
        1. Log in to the [Azure portal](https://portal.azure.com) with an administrator account and navigate to **Microsoft Entra ID**.

        2. In the top bar, click **+ Add** and select **Enterprise applications**. The **Browse Microsoft Entra Gallery** page opens.

        3. In the top bar, click **+ Create your own application**.

        4. Enter a name, for example, `Cognigy`, select **Integrate any other app you don't find in the gallery (Non-gallery)**, and click **Create**.

        5. On the new app page, open **Single sign-on** in the left navigation and select **SAML** as the sign-on method.

        6. In the **Basic SAML Configuration** section, click **Edit** and configure the following:

           * **Identifier (Entity ID)** — enter `https://<api-base-url>/auth/saml/login/<organization-id>`.
           * **Reply URL (Assertion Consumer Service URL)** — enter `https://<api-base-url>/auth/saml/login/<organization-id>`.
           * **Sign on URL** — enter `https://<api-base-url>/auth/saml/login/<organization-id>`.
           * *(Optional)* **Logout Url** — enter `https://<cognigy-url>/slo/<organization-id>`.

        7. Click **Save**.
      </Step>

      <Step title="Map User Attributes">
        1. In the **User Attributes and Claims** section, click **Edit** and confirm that the following claims are mapped to the user's profile attributes:

           | Claim name                                                           | Value            |
           | -------------------------------------------------------------------- | ---------------- |
           | `http://schemas.xmlsoap.org/ws/2005/05/identity/claims/emailaddress` | `user.mail`      |
           | `http://schemas.xmlsoap.org/ws/2005/05/identity/claims/givenname`    | `user.givenname` |
           | `http://schemas.xmlsoap.org/ws/2005/05/identity/claims/surname`      | `user.surname`   |

           Cognigy uses the email claim to identify the user and to create the Cognigy account on the first login.
      </Step>

      <Step title="Get Authentication Data">
        1. In the **SAML Signing Certificate** section, locate **Certificate (Base64)** and click **Download**. Save the file locally. Use this certificate as `idpCertificate` in the API request to [configure the IdP in Cognigy](#configure-sso-in-cognigy-ai). You need to encode the certificate as base64 without newlines beforehand. To do so, use the following terminal command:

           <CodeGroup>
             ```powershell PowerShell theme={null} theme={null}
             [Convert]::ToBase64String([IO.File]::ReadAllBytes(".\path-to-certificate"))
             ```

             ```bash Bash (macOS) theme={null} theme={null}
             base64 -i ./<path-to-certificate> | tr -d '\n'
             ```

             ```bash Bash (Linux) theme={null} theme={null}
             cat ./<path-to-certificate> | base64 -w0
             ```
           </CodeGroup>

        2. In the **Set up `<application name>`** section, copy the following values to be used in the API request to [configure the IdP in Cognigy](#configure-sso-in-cognigy-ai):

        * **Login URL** — used in the `idpLoginEndpoint` parameter in the request payload.
        * *(Optional)* **Logout URL** — used in the `idpLogoutUrl` parameter in the request payload when you enable single logout.
      </Step>

      <Step title="Assign Roles">
        1. In the top bar, search for `App registrations` and select this option.

        2. In the **All applications** tab, search for the app you created and select it.

        3. In the left-side menu, go to **Manage > Manifest** and deactivate the default `User` and `msiam_access` roles. To do so:

           1. Set `isEnabled` to `false` for the `User` and `msiam_access` roles, for example:
              ```json theme={null}
              "appRoles": [
                {
                  "allowedMemberTypes": [
                    "User"
                  ],
                  "description": "User",
                  "displayName": "User",
                  "id": "18d14569-c3bd-439b-9a66-3a2aee01d14f",
                  "isEnabled": false,
                  "origin": "Application",
                  "value": null
                },
                {
                  "allowedMemberTypes": [
                    "User"
                  ],
                  "description": "msiam_access",
                  "displayName": "msiam_access",
                  "id": "b9632174-c057-4f7e-951b-be3adc52bfe6",
                  "isEnabled": false,
                  "origin": "Application",
                  "value": null
                }
              ]
              ```
           2. In the left-side menu, go to **Manage > App roles**, select the **User** role, and click **Delete** on the right-side pane. Do the same for the `msiam_access` role.

        4. Paste the following JSON code in the `appRoles` array:
           <Accordion title="Roles Array">
             ```json theme={null}
             {
               "allowedMemberTypes": [
                 "User"
               ],
               "displayName": "Admin",
               "id": "8d17fe88-c0ca-4903-ae2a-a51098998bc1",
               "isEnabled": true,
               "description": "The Admin role in Cognigy",
               "value": "admin"
             },
             {
               "allowedMemberTypes": [
                 "User"
               ],
               "displayName": "API Keys",
               "id": "8d17fe88-c0ca-4903-ae2a-a51098998bc2",
               "isEnabled": true,
               "description": "The apiKeys role in Cognigy",
               "value": "apiKeys"
             },
             {
               "allowedMemberTypes": [
                 "User"
               ],
               "displayName": "Base",
               "id": "8d17fe88-c0ca-4903-ae2a-a51098998bc3",
               "isEnabled": true,
               "description": "The base role in Cognigy",
               "value": "base_role"
             },
             {
               "allowedMemberTypes": [
                 "User"
               ],
               "displayName": "Full Support User",
               "id": "8d17fe88-c0ca-4903-ae2a-a51098998bc4",
               "isEnabled": true,
               "description": "Admin privileges, no user assignments in Cognigy",
               "value": "fullSupportUser"
             },
             {
               "allowedMemberTypes": [
                 "User"
               ],
               "displayName": "OData",
               "id": "8d17fe88-c0ca-4903-ae2a-a51098998bc5",
               "isEnabled": true,
               "description": "The OData role in Cognigy",
               "value": "odata"
             },
             {
               "allowedMemberTypes": [
                 "User"
               ],
               "displayName": "Project Manager",
               "id": "8d17fe88-c0ca-4903-ae2a-a51098998bc6",
               "isEnabled": true,
               "description": "The Project Manager role in Cognigy",
               "value": "projectManager"
             },
             {
               "allowedMemberTypes": [
                 "User"
               ],
               "displayName": "User Manager",
               "id": "8d17fe88-c0ca-4903-ae2a-a51098998bc7",
               "isEnabled": true,
               "description": "The User Manager role in Cognigy",
               "value": "userManager"
             },
             {
               "allowedMemberTypes": [
                 "User"
               ],
               "displayName": "Administrator",
               "id": "8d17fe88-c0ca-4903-ae2a-a51098998bc8",
               "isEnabled": true,
               "description": "The Administrator role in Live Agent",
               "value": "liveAgentAdmin"
             },
             {
               "allowedMemberTypes": [
                 "User"
               ],
               "displayName": "Agent",
               "id": "8d17fe88-c0ca-4903-ae2a-a51098998bc9",
               "isEnabled": true,
               "description": "The Agent role in Live Agent",
               "value": "liveAgentAgent"
             },
             {
               "allowedMemberTypes": [
                 "User"
               ],
               "displayName": "Supervisor",
               "id": "8d17fe88-c0ca-4903-ae2a-a51098998bd1",
               "isEnabled": true,
               "description": "The Supervisor role in Live Agent",
               "value": "liveagentSupervisor"
             },
             {
               "allowedMemberTypes": [
                 "User"
               ],
               "displayName": "View user details",
               "id": "8d17fe88-c0ca-4903-ae2a-a51098998bd2",
               "isEnabled": true,
               "description": "The role to view user details in Cognigy",
               "value": "userDetailsViewer"
             },
             {
               "allowedMemberTypes": [
                 "User"
               ],
               "displayName": "Voice Gateway User",
               "id": "8d17fe88-c0ca-4903-ae2a-a51098998bd3",
               "isEnabled": true,
               "description": "The Account scope in Voice Gateway",
               "value": "voiceGatewayUser"
             },
             {
               "allowedMemberTypes": [
                 "User"
               ],
               "displayName": "Basic Support User",
               "id": "8d17fe88-c0ca-4903-ae2a-a51098998bd4",
               "isEnabled": true,
               "description": "Partial Admin, read-only, no assignments, no OData/API, no Knowledge AI in Cognigy",
               "value": "basicSupportUser"
             },
             {
               "allowedMemberTypes": [
                 "User"
               ],
               "displayName": "Project Assigner",
               "id": "8d17fe88-c0ca-4903-ae2a-a51098998bd5",
               "isEnabled": true,
               "description": "Assigns Agents, read-only access, no global roles, limited features in Cognigy",
               "value": "projectAssigner"
             }
             ```
           </Accordion>

        5. In the left-side menu, go to **Users and groups** and assign the app to the users or groups that must have access to Cognigy through SSO.
      </Step>
    </Steps>
  </Tab>

  <Tab title="Google">
    <Steps>
      <Step title="Add a Custom Attribute">
        You need to add a custom attribute that Cognigy uses when the user logs in for the first time. To do so, follow these steps:

        1. Log in to the [Google Admin console](https://admin.google.com/).

        2. In the left-side menu, go to **Directory > Users**.

        3. At the top of the **Users** list, click **More options > Manage user attributes**.

        4. Click the **Add Custom Attribute** button and configure the following:

           * **Category** — enter a unique name to identify the custom attribute, for example, `Cognigy SSO`.
           * **Description** — enter a relevant description of the custom attribute.
           * In the **Custom fields** section, configure three custom fields with the following values:

             | Name       | Info type | Visibility                | Number of values |
             | ---------- | --------- | ------------------------- | ---------------- |
             | First Name | Text      | Visible to user and admin | Single value     |
             | Last Name  | Text      | Visible to user and admin | Single value     |
             | Role       | Text      | Visible to user and admin | Single value     |

           <Note>
             If a user doesn't have a role assigned, Cognigy assigns `base_role` to the user.
           </Note>

        5. Click **Add**.
      </Step>

      <Step title="Set Up an App">
        1. In the left-side menu, go to **Apps > SAML Apps**.

        2. Click the plus button in the lower-right corner and select **Setup My Own Custom App**.

        3. In the **Google IdP Information** dialog, copy the **SSO URL** value and download the certificate. You'll use the SSO URL as `idpLoginEndpoint` and the certificate as `idpCertificate` in the API request to [configure the IdP in Cognigy](#configure-sso-in-cognigy-ai). You need to encode the certificate as base64 without newlines beforehand. To do so, use the following terminal command:

           <CodeGroup>
             ```powershell PowerShell theme={null} theme={null}
             [Convert]::ToBase64String([IO.File]::ReadAllBytes(".\path-to-certificate"))
             ```

             ```bash Bash (macOS) theme={null} theme={null}
             base64 -i ./<path-to-certificate> | tr -d '\n'
             ```

             ```bash Bash (Linux) theme={null} theme={null}
             cat ./<path-to-certificate> | base64 -w0
             ```
           </CodeGroup>

           Click **Next**.

        4. Enter a name for the app, for example, `Cognigy`, and click **Next**.

        5. On the service provider details page, configure the following:

           * **ACS URL** — enter `https://<api-base-url>/auth/saml/login/<organization-id>`.
           * **Entity ID** — enter `https://<api-base-url>/auth/saml/login/<organization-id>`.
           * **Signed Response** — select the checkbox.
           * **Name ID Format** — set to `EMAIL`.

           Click **Next**.

        6. On the **Attribute Mapping** page, add the following attribute mappings so that Cognigy receives the user's first name, last name, and role from Google:

        | Service Provider Attribute | Custom Attribute | Field      |
        | -------------------------- | ---------------- | ---------- |
        | `firstName`                | Cognigy SSO      | First Name |
        | `lastName`                 | Cognigy SSO      | Last Name  |
        | `role`                     | Cognigy SSO      | Role       |

        7. Click **Finish**, then activate the app for the users or organizational units that must have access to Cognigy through Google SSO.
      </Step>
    </Steps>
  </Tab>

  <Tab title="Okta">
    <Steps>
      <Step title="Set Up an App">
        1. Log in to your [Okta Admin Console](https://login.okta.com/) with an administrator account.
        2. In the left-side menu, go to **Applications > Applications** and click **Create New App**.
        3. Select **SAML 2.0** as the sign-on method and click **Create**.
        4. On the **General Settings** tab, enter an app name, for example, `Cognigy SSO`, optionally upload a logo, and click **Next**.
        5. On the **Configure SAML** tab, configure the following:

           * **Single sign-on URL** — enter `https://<api-base-url>/auth/saml/login/<organization-id>`.
           * **Use this for Recipient URL and Destination URL** — keep selected.
           * **Audience URI (SP Entity ID)** — enter `https://<api-base-url>/auth/saml/login/<organization-id>`.
           * **Name ID format** — select **Unspecified**.
           * **Application username** — select **Email**.
           * **Update application username on** — select **Create and update**.
        6. *(Optional)* To encrypt SAML requests, click **Show Advanced Settings**, set **Assertion Encryption** to **Encrypted**, and [upload an X.509 certificate](#prerequisites). You will need the private key of this certificate to register the IdP in Cognigy.
      </Step>

      <Step title="Map User Attributes">
        1. In the **Attribute Statements** section, add the following attributes so that Cognigy receives the user profile data it needs to create the account:

           | Name        | Name format | Value            |
           | ----------- | ----------- | ---------------- |
           | `email`     | Unspecified | `user.email`     |
           | `firstName` | Unspecified | `user.firstName` |
           | `lastName`  | Unspecified | `user.lastName`  |

           Click **Next**.

        2. On the **Feedback** tab, select **I'm an Okta customer adding an internal app** and click **Finish**.
      </Step>

      <Step title="Get Authentication Data">
        1. On the **Applications** page, select the app you created, then open the **Sign On** tab.
        2. In the **SAML Signing Certificates** section, click **View Setup Instructions** or **Identity Provider metadata** and copy the following values:

           * **Identity Provider Single Sign-On URL** — used in the `idpLoginEndpoint` parameter in the request payload.
           * **Identity Provider Issuer** — used in the `idpIssuer` parameter in the request payload.
           * **X.509 Certificate** — used in the `idpCertificate` parameter in the request payload. You need to encode the certificate as base64 without newlines beforehand. If you enabled SAML request encryption, you also need to encode the private key as base64. You can use the following terminal command:

             <CodeGroup>
               ```powershell PowerShell theme={null} theme={null}
               [Convert]::ToBase64String([IO.File]::ReadAllBytes(".\path-to-certificate-or-private-key"))
               ```

               ```bash Bash (macOS) theme={null} theme={null}
               base64 -i ./<path-to-certificate-or-private-key> | tr -d '\n'
               ```

               ```bash Bash (Linux) theme={null} theme={null}
               cat ./<path-to-certificate-or-private-key> | base64 -w0
               ```
             </CodeGroup>
      </Step>

      <Step title="Assign Users">
        1. In the left-side menu of the Admin Console, go to **Directory > Profile Editor**, and click **Profile** next to the profile of the app you created.

        2. Click **Add Attribute** and configure the following:

           * **Data type** — select **string**.
           * **Display name** — enter `Role`.
           * **Variable name** — enter `role`.
           * *(Optional)* **Description** — enter a relevant description, for example, `The role of the user in Cognigy`.
           * **Enum** — activate **Define enumerated list of values**.
           * **Attribute member** — enter a display name and the value of the role to which SSO applies, for example, `Base Role` and `base_role`. The value must match a [role in Cognigy](/ai/administer/access/admin-center/access-control).
           * **Attribute required** — select this option.\
             Click **Save**.

        3. Go to **Applications > Applications** and select the app you created.

        4. On the **Assignments** tab, click the Edit button next to the user's name and assign the role they should have in Cognigy. This user can log in via SSO.
      </Step>
    </Steps>
  </Tab>

  <Tab title="OneLogin">
    <Steps>
      <Step title="Set Up an App">
        1. Log in to your OneLogin Administration portal with an administrator account.

        2. In the top bar, click **Applications**, then click **Add App** in the upper-right corner.

        3. Search for `SAML Custom Connector (Advanced)` in the search field and select this option.

        4. In the **Display Name** field, enter a name that identifies the integration, for example, `Cognigy SSO`. Optionally upload icons. Click **Save**.

        5. On the new app page, click the **Configuration** tab in the left-side menu and configure the following:

           * **ACS (Consumer) URL** — enter `https://<api-base-url>/auth/saml/login/<organization-id>`.
           * **ACS (Consumer) URL Validator** — enter `https://<api-base-url>/auth/saml/login/<organization-id>`.
           * *(Optional)* **Single Logout URL** — enter `https://<cognigy-url>/slo/<organization-id>`.

           Click **Save**.

        6. In the left-side menu, go to the **Parameters** tab, configure the following parameter and value pairs, and select the **Include in SAML assertion** option for each pair:

           | Parameter   | Value      |
           | ----------- | ---------- |
           | `email`     | Email      |
           | `firstName` | First Name |
           | `lastName`  | Last Name  |
           | `role`      | User Role  |

           Click **Save** to apply the attribute mapping.
      </Step>

      <Step title="Get Authentication Data">
        1. In the left-side menu, go to the **SSO** tab and copy the following values:

           * **X.509 Certificate** — click **View Details** and download the certificate. You will use its contents as `idpCertificate`. You need to encode the certificate as base64 without newlines beforehand. You can use the following terminal command:

             <CodeGroup>
               ```powershell PowerShell theme={null} theme={null}
               [Convert]::ToBase64String([IO.File]::ReadAllBytes(".\path-to-certificate"))
               ```

               ```bash Bash (macOS) theme={null} theme={null}
               base64 -i ./<path-to-certificate> | tr -d '\n'
               ```

               ```bash Bash (Linux) theme={null} theme={null}
               cat ./<path-to-certificate> | base64 -w0
               ```
             </CodeGroup>

           * **Issuer URL** — used in the `idpIssuer` parameter in the request payload.

           * **SAML 2.0 Endpoint (HTTP)** — used in the `idpLoginEndpoint` parameter in the request payload.

           * *(Optional)* **SLO Endpoint** — used in the `idpLogoutEndpoint` parameter in the request payload.

        2. In the left-side menu, navigate to **Users > Roles** and click **New Role**.

        3. Add the following Cognigy roles one by one, select the app you created, and click **Save**:

           * `admin`
           * `apiKeys`
           * `base_role`
           * `livechat`
           * `odata`
           * `projectManager`
           * `userManager`

        4. In the left-side menu, go to the **Users** tab and select a user for SSO. On the user page, enter the role they have in Cognigy in the **role** field. The role must match the roles you configured in step 3.

        5. Click **Save**. The user can log in via SSO.
      </Step>
    </Steps>
  </Tab>
</Tabs>

## Configure SSO in Cognigy

After the SSO app is ready, use the [`POST /v2.0/identityprovider/configure`](/api-reference/identity-providers/set-up-an-identity-provider) method to register the SSO configuration in Cognigy.

<Tabs>
  <Tab title="Auth0 - OpenID Connect">
    Send the API request with the following parameters:

    * `X-API-Key` header — the API key from the [Prerequisites](#prerequisites) section.
    * `idpIssuer` — the Domain value you copied earlier from [the **Settings** tab](#create-an-idp-app).
    * `idpClientId` — the Client ID value you copied earlier from [the **Settings** tab](#create-an-idp-app).
    * `idpClientSecret` — the Client Secret value you copied earlier from [the **Settings** tab](#create-an-idp-app).

    <Accordion title="API Request Example">
      <CodeGroup>
        ```bash -X POST "https://<api-base-url>/v2.0/identityprovider/configure" \ theme={null}
          -H "Content-Type: application/json" \
          -H "X-API-Key: <your-api-token>" \
          -d '{
            "idpType": "oidc",
            "idpIssuer": "<DOMAIN>",
            "idpClientId": "<CLIENT ID>",
            "idpClientSecret": "<CLIENT SECRET>",
            "idpIdTokenSignedResponseAlg": "RS256",
            "idpTokenEndpointAuthMethod": "client_secret_basic"
          }'
        ```

        ```javascript JavaScript theme={null}
        const response = await fetch("https://<api-base-url>/v2.0/identityprovider/configure", {
          method: "POST",
          headers: {
            "Content-Type": "application/json",
            "X-API-Key": "<your-api-token>"
          },
          body: JSON.stringify({
            idpType: "oidc",
            idpIssuer: "<DOMAIN>",              // Domain from the Auth0 app Settings tab
            idpClientId: "<CLIENT ID>",         // Client ID from the Auth0 app Settings tab
            idpClientSecret: "<CLIENT SECRET>", // Client Secret from the Auth0 app Settings tab
            idpIdTokenSignedResponseAlg: "RS256",
            idpTokenEndpointAuthMethod: "client_secret_basic"
          })
        });
        ```

        ```python Python theme={null}
        import requests

        response = requests.post(
            "https://<api-base-url>/v2.0/identityprovider/configure",
            headers={
                "Content-Type": "application/json",
                "X-API-Key": "<your-api-token>"
            },
            json={
                "idpType": "oidc",
                "idpIssuer": "<DOMAIN>",                # Domain from the Auth0 app Settings tab
                "idpClientId": "<CLIENT ID>",           # Client ID from the Auth0 app Settings tab
                "idpClientSecret": "<CLIENT SECRET>",   # Client Secret from the Auth0 app Settings tab
                "idpIdTokenSignedResponseAlg": "RS256",
                "idpTokenEndpointAuthMethod": "client_secret_basic"
            }
        )
        ```
      </CodeGroup>
    </Accordion>
  </Tab>

  <Tab title="Auth0 - SAML 2.0">
    Send the API request with the following data:

    * `X-API-Key` header — the API key from the [Prerequisites](#prerequisites) section.
    * `idpIssuer` — the Issuer value you copied earlier from [the **Usage** tab of the **SAML2 Web App** add-on](#create-an-idp-app).
    * `idpLoginEndpoint` — the Identity Provider Login URL you copied earlier from [the **Usage** tab of the **SAML2 Web App** add-on](#create-an-idp-app).
    * `idpCertificate` — the Identity Provider Certificate you copied earlier from [the **Usage** tab of the **SAML2 Web App** add-on](#create-an-idp-app), including the `BEGIN`/`END` markers, with real line breaks replaced by `\n`.
    * `idpIdentifierFormat` — the NameID format. Use `urn:oasis:names:tc:SAML:1.1:nameid-format:emailAddress` to match the add-on settings.

    <Accordion title="API Request Example">
      <CodeGroup>
        ```bash -X POST "https://<api-base-url>/v2.0/identityprovider/configure" \ theme={null}
          -H "Content-Type: application/json" \
          -H "X-API-Key: <your-api-token>" \
          -d '{
            "idpType": "saml",
            "idpIssuer": "urn:<your-tenant>.auth0.com",
            "idpLoginEndpoint": "https://<your-tenant>.auth0.com/samlp/<client-id>",
            "idpCertificate": "-----BEGIN CERTIFICATE-----\n<certificate-body>\n-----END CERTIFICATE-----",
            "idpIdentifierFormat": "urn:oasis:names:tc:SAML:1.1:nameid-format:emailAddress"
          }'
        ```

        ```javascript JavaScript theme={null}
        const response = await fetch("https://<api-base-url>/v2.0/identityprovider/configure", {
          method: "POST",
          headers: {
            "Content-Type": "application/json",
            "X-API-Key": "<your-api-token>"
          },
          body: JSON.stringify({
            idpType: "saml",
            idpLoginEndpoint: "https://<your-tenant>.auth0.com/samlp/<client-id>",                          // Identity Provider Login URL from the Auth0 add-on Usage tab
            idpIssuer: "urn:<your-tenant>.auth0.com",                                                    // Identity Provider Issuer from the Auth0 add-on Usage tab
            idpCertificate: "-----BEGIN CERTIFICATE-----\n<certificate-body>\n-----END CERTIFICATE-----",      // Identity Provider Certificate from the Auth0 add-on Usage tab
            idpIdentifierFormat: "urn:oasis:names:tc:SAML:1.1:nameid-format:emailAddress"
          })
        });
        ```

        ```python Python theme={null}
        import requests

        response = requests.post(
            "https://<api-base-url>/v2.0/identityprovider/configure",
            headers={
                "Content-Type": "application/json",
                "X-API-Key": "<your-api-token>"
            },
            json={
                "idpType": "saml",
                "idpLoginEndpoint": "https://<your-tenant>.auth0.com/samlp/<client-id>",                          # Identity Provider Login URL from the Auth0 add-on Usage tab
                "idpIssuer": "urn:<your-tenant>.auth0.com",                                                    # Identity Provider Issuer from the Auth0 add-on Usage tab
                "idpCertificate": "-----BEGIN CERTIFICATE-----\n<certificate-body>\n-----END CERTIFICATE-----",      # Identity Provider Certificate from the Auth0 add-on Usage tab
                "idpIdentifierFormat": "urn:oasis:names:tc:SAML:1.1:nameid-format:emailAddress"
            }
        )
        ```
      </CodeGroup>
    </Accordion>
  </Tab>

  <Tab title="Microsoft Entra ID">
    Send the API request with the following data:

    * `X-API-Key` header — the API key from the [Prerequisites](#prerequisites) section.
    * `idpIssuer` — use `https://<api-base-url>/auth/saml/login/<organization-id>`.
    * `idpLoginEndpoint` — the Login URL you copied earlier from the **Set up `<application name>`** section.
    * `idpCertificate` — the certificate you downloaded earlier from the **SAML Signing Certificate** section, base64-encoded as a single continuous string with no line breaks.
    * *(Optional)* `idpLogoutUrl` — the Logout URL you copied earlier from the **Set up `<application name>`** section.

    <Accordion title="API Request Example">
      <CodeGroup>
        ```bash cURL theme={null}
        curl -X POST "https://<api-base-url>/v2.0/identityprovider/configure" \
          -H "Content-Type: application/json" \
          -H "X-API-Key: <your-api-token>" \
          -d '{
            "idpType": "saml",
            "idpIssuer": "https://<api-base-url>/auth/saml/login/<organization-id>",
            "idpLoginEndpoint": "https://login.microsoftonline.com/4a7853bd-xxxx-xxxx-xxxx-xxxxxxxxxxxx/saml2",
            "idpCertificate": "LS0tLS1CRUdJTiBDRVJUSUZJQ0FURS0tLS0t...",
            "idpLogoutUrl": "https://login.microsoftonline.com/4a7853bd-xxxx-xxxx-xxxx-xxxxxxxxxxxx/saml2"
          }'
        ```

        ```javascript JavaScript theme={null}
        const response = await fetch("https://<api-base-url>/v2.0/identityprovider/configure", {
          method: "POST",
          headers: {
            "Content-Type": "application/json",
            "X-API-Key": "<your-api-token>"
          },
          body: JSON.stringify({
            idpType: "saml",
            idpIssuer: "https://<api-base-url>/auth/saml/login/<organization-id>",                                // Cognigy SSO URL
            idpLoginEndpoint: "https://login.microsoftonline.com/4a7853bd-xxxx-xxxx-xxxx-xxxxxxxxxxxx/saml2",    // Login URL from the Set up `<application name>` section
            idpCertificate: "LS0tLS1CRUdJTiBDRVJUSUZJQ0FURS0tLS0t...",                                            // Base64-encoded contents of the downloaded .crt file
            idpLogoutUrl: "https://login.microsoftonline.com/4a7853bd-xxxx-xxxx-xxxx-xxxxxxxxxxxx/saml2"          // Logout URL from the Set up `<application name>` section (optional, for SP-initiated SLO)
          })
        });
        ```

        ```python Python theme={null}
        import requests

        response = requests.post(
            "https://<api-base-url>/v2.0/identityprovider/configure",
            headers={
                "Content-Type": "application/json",
                "X-API-Key": "<your-api-token>"
            },
            json={
                "idpType": "saml",
                "idpIssuer": "https://<api-base-url>/auth/saml/login/<organization-id>",                                # Cognigy SSO URL
                "idpLoginEndpoint": "https://login.microsoftonline.com/4a7853bd-xxxx-xxxx-xxxx-xxxxxxxxxxxx/saml2",    # Login URL from the Set up `<application name>` section
                "idpCertificate": "LS0tLS1CRUdJTiBDRVJUSUZJQ0FURS0tLS0t...",                                            # Base64-encoded contents of the downloaded .crt file
                "idpLogoutUrl": "https://login.microsoftonline.com/4a7853bd-xxxx-xxxx-xxxx-xxxxxxxxxxxx/saml2"          # Logout URL from the Set up `<application name>` section (optional, for SP-initiated SLO)
            }
        )
        ```
      </CodeGroup>
    </Accordion>
  </Tab>

  <Tab title="Google">
    <Warning>
      Don't include the PEM header, the PEM footer, or any newline characters in the certificate you submit to Cognigy. A malformed certificate string causes the SSO configuration request to fail.
    </Warning>

    Send the API request with the following data:

    * `X-API-Key` header — the API key from the [Prerequisites](#prerequisites) section.
    * `idpIssuer` — use `https://<api-base-url>/auth/saml/login/<organization-id>`.
    * `idpLoginEndpoint` — the **SSO URL** value you copied earlier from the **Google IdP Information** dialog.
    * `idpCertificate` — the certificate you downloaded earlier from the **Google IdP Information** dialog, base64-encoded as a single continuous string with no line breaks.

    <Accordion title="API Request Example">
      <CodeGroup>
        ```bash cURL theme={null}
        curl -X POST "https://<api-base-url>/v2.0/identityprovider/configure" \
          -H "Content-Type: application/json" \
          -H "X-API-Key: <your-api-token>" \
          -d '{
            "idpType": "saml",
            "idpIssuer": "https://<api-base-url>/auth/saml/login/<organization-id>",
            "idpLoginEndpoint": "https://accounts.google.com/o/saml2/idp?idpid=XXXX",
            "idpCertificate": "MIIDdTCCAl2gAwIBAgIJAKx..."
          }'
        ```

        ```javascript JavaScript theme={null}
        const response = await fetch("https://<api-base-url>/v2.0/identityprovider/configure", {
          method: "POST",
          headers: {
            "Content-Type": "application/json",
            "X-API-Key": "<your-api-token>"
          },
          body: JSON.stringify({
            idpType: "saml",
            idpIssuer: "https://<api-base-url>/auth/saml/login/<organization-id>",   // Cognigy SSO URL
            idpLoginEndpoint: "https://accounts.google.com/o/saml2/idp?idpid=XXXX",  // SSO URL from Google IdP Information dialog in the Google Admin console
            idpCertificate: "MIIDdTCCAl2gAwIBAgIJAKx..."                              // Certificate from Google IdP Information dialog, base64-encoded on a single line
          })
        });
        ```

        ```python Python theme={null}
        import requests

        response = requests.post(
            "https://<api-base-url>/v2.0/identityprovider/configure",
            headers={
                "Content-Type": "application/json",
                "X-API-Key": "<your-api-token>"
            },
            json={
                "idpType": "saml",
                "idpIssuer": "https://<api-base-url>/auth/saml/login/<organization-id>",    # Cognigy SSO URL
                "idpLoginEndpoint": "https://accounts.google.com/o/saml2/idp?idpid=XXXX",   # SSO URL from Google IdP Information dialog in the Google Admin console
                "idpCertificate": "MIIDdTCCAl2gAwIBAgIJAKx..."                               # Certificate from Google IdP Information dialog, base64-encoded on a single line
            }
        )
        ```
      </CodeGroup>
    </Accordion>
  </Tab>

  <Tab title="Okta">
    Send the API request with the following data:

    * `X-API-Key` header — the API key from the [Prerequisites](#prerequisites) section.
    * `idpIssuer` — the Identity Provider Issuer value you copied [from the SAML setup instructions page](#create-an-idp-app).
    * `idpLoginEndpoint` — the Identity Provider Single Sign-On URL value you copied [from the SAML setup instructions page](#create-an-idp-app).
    * For encrypted SAML requests:
      * `idpCertificate` — X.509 Certificate you uploaded under **Show Advanced Settings > Assertion Encryption**, base64-encoded as a single continuous string with no line breaks.
      * `decryptionPrivateKey` — the private key matching the X.509 certificate you [uploaded](#create-an-idp-app) under **Show Advanced Settings > Assertion Encryption** and base64-encoded as a single continuous string.

    <AccordionGroup>
      <Accordion title="API Request Example (Without Encryption)">
        <CodeGroup>
          ```bash cURL theme={null}
          curl -X POST "https://<api-base-url>/v2.0/identityprovider/configure" \
            -H "Content-Type: application/json" \
            -H "X-API-Key: <your-api-token>" \
            -d '{
              "idpType": "saml",
              "idpIssuer": "<OKTA_IDENTITY_PROVIDER_ISSUER>",
              "idpLoginEndpoint": "<OKTA_SINGLE_SIGN_ON_URL>"
            }'
          ```

          ```javascript JavaScript theme={null}
          const response = await fetch("https://<api-base-url>/v2.0/identityprovider/configure", {
            method: "POST",
            headers: {
              "Content-Type": "application/json",
              "X-API-Key": "<your-api-token>"
            },
            body: JSON.stringify({
              idpType: "saml",
              idpIssuer: "<OKTA_IDENTITY_PROVIDER_ISSUER>",     // Identity Provider Issuer from the Okta Sign On tab
              idpLoginEndpoint: "<OKTA_SINGLE_SIGN_ON_URL>"        // Identity Provider Single Sign-On URL from the Okta Sign On tab
            })
          });
          ```

          ```python Python theme={null}
          import requests

          response = requests.post(
              "https://<api-base-url>/v2.0/identityprovider/configure",
              headers={
                  "Content-Type": "application/json",
                  "X-API-Key": "<your-api-token>"
              },
              json={
                  "idpType": "saml",
                  "idpIssuer": "<OKTA_IDENTITY_PROVIDER_ISSUER>",     # Identity Provider Issuer from the Okta Sign On tab
                  "idpLoginEndpoint": "<OKTA_SINGLE_SIGN_ON_URL>"        # Identity Provider Single Sign-On URL from the Okta Sign On tab
              }
          )
          ```
        </CodeGroup>
      </Accordion>

      <Accordion title="API Request Example (With Encryption)">
        <CodeGroup>
          ```bash cURL theme={null}
          curl -X POST "https://<api-base-url>/v2.0/identityprovider/configure" \
            -H "Content-Type: application/json" \
            -H "X-API-Key: <your-api-token>" \
            -d '{
              "idpType": "saml",
              "idpIssuer": "<OKTA_IDENTITY_PROVIDER_ISSUER>",
              "idpLoginEndpoint": "<OKTA_SINGLE_SIGN_ON_URL>",
              "idpCertificate": "<OKTA_X509_CERTIFICATE>",
              "decryptionPrivateKey": "<OKTA_PRIVATE_KEY>"
            }'
          ```

          ```javascript JavaScript theme={null}
          const response = await fetch("https://<api-base-url>/v2.0/identityprovider/configure", {
            method: "POST",
            headers: {
              "Content-Type": "application/json",
              "X-API-Key": "<your-api-token>"
            },
            body: JSON.stringify({
              idpType: "saml",
              idpIssuer: "<OKTA_IDENTITY_PROVIDER_ISSUER>",     // Identity Provider Issuer from the Okta Sign On tab
              idpLoginEndpoint: "<OKTA_SINGLE_SIGN_ON_URL>",    // Identity Provider Single Sign-On URL from the Okta Sign On tab
              idpCertificate: "<OKTA_X509_CERTIFICATE>",        // X.509 Certificate from the Okta Sign On tab, base64-encoded on a single line
              decryptionPrivateKey: "<OKTA_PRIVATE_KEY>"        // Private key matching the certificate uploaded to Okta, base64-encoded on a single line
            })
          });
          ```

          ```python Python theme={null}
          import requests

          response = requests.post(
              "https://<api-base-url>/v2.0/identityprovider/configure",
              headers={
                  "Content-Type": "application/json",
                  "X-API-Key": "<your-api-token>"
              },
              json={
                  "idpType": "saml",
                  "idpIssuer": "<OKTA_IDENTITY_PROVIDER_ISSUER>",     # Identity Provider Issuer from the Okta Sign On tab
                  "idpLoginEndpoint": "<OKTA_SINGLE_SIGN_ON_URL>",       # Identity Provider Single Sign-On URL from the Okta Sign On tab
                  "idpCertificate": "<OKTA_X509_CERTIFICATE>",        # X.509 Certificate from the Okta Sign On tab, base64-encoded on a single line
                  "decryptionPrivateKey": "<OKTA_PRIVATE_KEY>"        # Private key matching the certificate uploaded to Okta, base64-encoded on a single line
              }
          )
          ```
        </CodeGroup>
      </Accordion>
    </AccordionGroup>
  </Tab>

  <Tab title="OneLogin">
    Send the API request with the following data:

    * `X-API-Key` header — the API key from the [Prerequisites](#prerequisites) section.
    * `idpIssuer` — the Issuer URL you copied earlier from [the **SSO** tab of the OneLogin app](#create-an-idp-app).
    * `idpLoginEndpoint` — the SAML 2.0 Endpoint (HTTP) you copied earlier from [the **SSO** tab of the OneLogin app](#create-an-idp-app).
    * `idpCertificate` — the X.509 Certificate you encoded earlier from [the **SSO** tab of the OneLogin app](#create-an-idp-app), base64-encoded as a single continuous string with no line breaks.
    * *(Optional)* `idpLogoutEndpoint` — the SLO Endpoint you copied earlier from [the **SSO** tab of the OneLogin app](#create-an-idp-app).

    <Accordion title="API Request Example">
      <CodeGroup>
        ```bash cURL theme={null}
        curl -X POST "https://<api-base-url>/v2.0/identityprovider/configure" \
          -H "Content-Type: application/json" \
          -H "X-API-Key: <your-api-token>" \
          -d '{
            "idpType": "saml",
            "idpIssuer": "https://app.onelogin.com/saml/metadata/<id>",
            "idpLoginEndpoint": "https://<subdomain>.onelogin.com/trust/saml2/http-post/sso/<id>",
            "idpCertificate": "<ONELOGIN_X509_CERTIFICATE>",
            "idpLogoutEndpoint": "https://<subdomain>.onelogin.com/trust/saml2/http-redirect/slo/<id>"
          }'
        ```

        ```javascript JavaScript theme={null}
        const response = await fetch("https://<api-base-url>/v2.0/identityprovider/configure", {
          method: "POST",
          headers: {
            "Content-Type": "application/json",
            "X-API-Key": "<your-api-token>"
          },
          body: JSON.stringify({
            idpType: "saml",
            idpIssuer: "https://app.onelogin.com/saml/metadata/<id>",                                      // Issuer URL from the OneLogin SSO tab
            idpLoginEndpoint: "https://<subdomain>.onelogin.com/trust/saml2/http-post/sso/<id>",           // SAML 2.0 Endpoint (HTTP) from the OneLogin SSO tab
            idpCertificate: "<ONELOGIN_X509_CERTIFICATE>",                                                 // X.509 Certificate from the OneLogin SSO tab, base64-encoded on a single line
            idpLogoutEndpoint: "https://<subdomain>.onelogin.com/trust/saml2/http-redirect/slo/<id>"       // SLO Endpoint from the OneLogin SSO tab (optional)
          })
        });
        ```

        ```python Python theme={null}
        import requests

        response = requests.post(
            "https://<api-base-url>/v2.0/identityprovider/configure",
            headers={
                "Content-Type": "application/json",
                "X-API-Key": "<your-api-token>"
            },
            json={
                "idpType": "saml",
                "idpIssuer": "https://app.onelogin.com/saml/metadata/<id>",                                      # Issuer URL from the OneLogin SSO tab
                "idpLoginEndpoint": "https://<subdomain>.onelogin.com/trust/saml2/http-post/sso/<id>",           # SAML 2.0 Endpoint (HTTP) from the OneLogin SSO tab
                "idpCertificate": "<ONELOGIN_X509_CERTIFICATE>",                                                 # X.509 Certificate from the OneLogin SSO tab, base64-encoded on a single line
                "idpLogoutEndpoint": "https://<subdomain>.onelogin.com/trust/saml2/http-redirect/slo/<id>"       # SLO Endpoint from the OneLogin SSO tab (optional)
            }
        )
        ```
      </CodeGroup>
    </Accordion>
  </Tab>
</Tabs>

A successful request returns a confirmation that the identity provider configuration has been saved. The Cognigy login page now displays a **Log in with SSO** button for users in your organization.

## Test the SSO Login

1. On the Cognigy login page, enter the email address of a user assigned to the IdP app and click **Log in with SSO**. Cognigy redirects you to the IdP and prompts you for your IdP credentials. After authentication, the IdP redirects back to Cognigy and logs you in.
2. Click the user icon in the lower-left corner and select **Logout**.
3. Enter your email address again and click **Log in with SSO**. This time, you are logged in directly without re-entering your IdP credentials.

## More Information

* [Single Sign-on with SAML 2.0](/ai/administer/installation/single-sign-on-saml2)
* [Auth0: OpenID Connect Protocol](https://auth0.com/docs/authenticate/protocols/openid-connect-protocol)
* [Auth0: SAML Configuration](https://auth0.com/docs/authenticate/protocols/saml/saml-configuration)
* [Microsoft Entra ID: Enable SAML single sign-on for an enterprise application](https://learn.microsoft.com/en-us/entra/identity/enterprise-apps/add-application-portal-setup-sso)
* [Google Workspace: Set up your own custom SAML app](https://knowledge.workspace.google.com/admin/apps/set-up-your-own-custom-saml-app)
* [Okta: Create SAML app integrations](https://help.okta.com/oie/en-us/content/topics/apps/apps_app_integration_wizard_saml.htm)
* [OneLogin: SAML Custom Connector (Advanced)](https://support.onelogin.com/kb/4266907/saml-custom-connector-advanced)


# IP Address Ranges for SaaS
Source: https://docs.cognigy.com/ai/administer/installation/ip-ranges-shared-environments

List of public IP address ranges for NiCE Cognigy and NiCE CXone SaaS installations used for SIP signaling and outbound requests, supporting IP whitelisting.

This page lists the public IP addresses used by NiCE Cognigy and NiCE CXone SaaS installations to perform SIP signaling or send outbound requests.

Both Cognigy and Cognigy Voice Gateway use multiple public *IP addresses* per installation. *IP addresses* in this document refer to the network addresses used for outbound requests (Cognigy) and SIP signaling (Voice Gateway).

If you implement IP whitelisting on your API servers for added security, inform Cognigy Support or your Customer Success Manager.

<Note>
  The list of IP addresses may change over time. NiCE Cognigy will promptly notify you of any updates.
</Note>

<Warning>
  The IP addresses for SIP signaling aren't used for transmitting audio data. Separate IP addresses handle audio/media traffic and are subject to change due to automatic scaling, so NiCE Cognigy doesn't provide those addresses. The required IP addresses and ports for audio/media will be negotiated dynamically when establishing calls.
</Warning>

Find the relevant public IP addresses according to your installation type:

* [Cognigy SaaS](#nice-cognigy-saas-clusters)
* [NiCE CXone Shared SaaS](#nice-cxone-shared-saas-clusters)

## NiCE Cognigy SaaS Clusters

Standard Cognigy environments in multiple regions.

### 🇪🇺 Europe SaaS - app.cognigy.ai

#### Cognigy

Cognigy uses the following IP addresses for outbound requests:

* 3.70.39.176
* 3.69.130.45
* 3.124.74.126

#### Cognigy Voice Gateway

Cognigy Voice Gateway uses the following IP addresses for SIP signaling:

* 3.68.22.26
* 3.73.70.70
* 3.74.34.237

### 🇺🇸 US SaaS - app-us.cognigy.ai

#### Cognigy

Cognigy uses the following IP addresses for outbound requests:

* 3.229.129.99
* 34.236.251.178
* 54.236.86.4

#### Cognigy Voice Gateway

Cognigy Voice Gateway uses the following IP addresses for SIP signaling:

* 18.210.218.95
* 34.233.59.163

### 🇬🇧 United Kingdom SaaS - app-uk.cognigy.ai

#### Cognigy

Cognigy uses the following IP addresses for outbound requests:

* 13.42.121.72
* 3.10.92.32
* 35.177.65.3

#### Cognigy Voice Gateway

Cognigy Voice Gateway will use the following IP addresses for SIP signaling:

* 13.43.21.62
* 18.170.191.49

### 🇦🇺 Australia SaaS - au-02.cognigy.ai

#### Cognigy

Cognigy uses the following IP addresses for outbound requests:

* 13.238.252.9
* 13.55.38.237
* 3.105.214.205

#### Cognigy Voice Gateway

Cognigy Voice Gateway uses the following IP addresses for SIP signaling:

* 54.66.80.76
* 52.65.113.43

### 🇪🇺 Trial Europe - trial.cognigy.ai

#### Cognigy

Cognigy uses the following IP addresses for outbound requests:

* 3.69.16.128
* 18.158.151.80
* 3.64.180.26

#### Cognigy Voice Gateway

Cognigy Voice Gateway uses the following IP addresses for SIP signaling:

* 3.77.106.237
* 18.197.64.96

### 🇺🇸 Trial US - trial-us.cognigy.ai

#### Cognigy

Cognigy uses the following IP addresses for outbound requests:

* 50.17.209.2
* 54.158.99.91
* 34.225.133.252

#### Cognigy Voice Gateway

Cognigy Voice Gateway uses the following IP addresses for SIP signaling:

* 3.218.190.67
* 44.223.130.33

## NiCE CXone Shared SaaS Clusters

Shared NiCE CXone clusters hosted on AWS in various regions.

### 🇦🇺 AU1 \[Asia Pacific (Sydney)] - cognigy-au1.nicecxone.com

#### Cognigy

NiCE CXone uses the following IP addresses for outbound requests:

* 15.134.174.233
* 54.79.247.44
* 13.237.226.156

#### Cognigy Voice Gateway

NiCE CXone uses the following IP addresses for SIP signaling:

* 54.253.194.206
* 15.134.158.214
* 3.106.31.243

### 🇯🇵 JP1 \[Asia Pacific (Tokyo)] - cognigy-jp1.nicecxone.com

#### Cognigy

NiCE CXone uses the following IP addresses for outbound requests:

* 52.68.126.161
* 54.95.9.113
* 52.68.154.59

#### Cognigy Voice Gateway

NiCE CXone uses the following IP addresses for SIP signaling:

* 54.64.108.183
* 176.34.59.56
* 18.179.170.87

### 🇺🇸 NA3 \[US East (N. Virginia)] - cognigy-na3.nicecxone.com

#### Cognigy

NiCE CXone uses the following IP addresses for outbound requests:

* 100.49.18.38
* 3.209.7.189
* 54.87.214.223

#### Cognigy Voice Gateway

NiCE CXone uses the following IP addresses for SIP signaling:

* 100.49.58.109
* 13.223.123.139
* 3.232.173.71

### 🇨🇦 CA1 \[Canada (Central)] - cognigy-ca1.nicecxone.com

#### Cognigy

NiCE CXone uses the following IP addresses for outbound requests:

* 15.156.226.173
* 15.157.85.122
* 15.223.137.91

#### Cognigy Voice Gateway

NiCE CXone uses the following IP addresses for SIP signaling:

* 3.97.48.74
* 16.52.187.121
* 16.52.13.133

### 🇪🇺 EU1 \[Europe (Frankfurt)] - cognigy-eu1.nicecxone.com

#### Cognigy

NiCE CXone uses the following IP addresses for outbound requests:

* 63.180.145.39
* 18.153.113.137
* 3.120.154.140

#### Cognigy Voice Gateway

NiCE CXone uses the following IP addresses for SIP signaling:

* 63.176.23.224
* 18.156.88.20
* 63.180.111.41

### 🇺🇸 NA1 \[US West (Oregon)] - cognigy-na1.nicecxone.com

#### Cognigy

NiCE CXone uses the following IP addresses for outbound requests:

* 44.252.142.51
* 44.254.240.46
* 44.254.21.124

#### Cognigy Voice Gateway

NiCE CXone uses the following IP addresses for SIP signaling:

* 44.231.41.51
* 52.38.13.255
* 16.144.112.10

### 🇬🇧 UK1 \[Europe (London)] - cognigy-uk1.nicecxone.com

#### Cognigy

NiCE CXone uses the following IP addresses for outbound requests:

* 18.135.62.239
* 13.43.245.110
* 51.24.4.142

#### Cognigy Voice Gateway

NiCE CXone uses the following IP addresses for SIP signaling:

* 13.43.5.8
* 18.171.45.134
* 13.41.33.70

### 🇬🇧 UK3 \[Europe (London)] - cognigy-uk3.nicecxone-sov2.uk

#### Cognigy

NiCE CXone uses the following IP addresses for outbound requests:

* 18.133.25.173
* 13.134.88.22
* 13.42.195.17

#### Cognigy Voice Gateway

NiCE CXone uses the following IP addresses for SIP signaling:

* 3.11.88.93
* 52.56.213.183

### 🇮🇳 IN1 \[Asia Pacific (Mumbai)] - cognigy-in1.nicecxone.com

#### Cognigy

NiCE CXone uses the following IP addresses for outbound requests:

* 15.207.158.222
* 3.6.187.58
* 13.205.10.135

#### Cognigy Voice Gateway

NiCE CXone uses the following IP addresses for SIP signaling:

* 3.108.85.21
* 13.232.154.152
* 13.201.208.232

### 🇿🇦 ZA1 \[Africa (Cape Town)] - cognigy-za1.nicecxone.com

#### Cognigy

NiCE CXone uses the following IP addresses for outbound requests:

* 16.28.7.47
* 15.240.7.206
* 13.246.150.193

#### Cognigy Voice Gateway

NiCE CXone uses the following IP addresses for SIP signaling:

* 13.245.83.99
* 13.244.217.225
* 13.244.158.63


# Cognigy SaaS Service Health
Source: https://docs.cognigy.com/ai/administer/installation/saas-service-health

Monitor real-time and historical Cognigy SaaS system performance, view past incidents, and subscribe to updates via the SaaS Service Health platform.

The *Cognigy SaaS Service Health* platform lets Cognigy SaaS customers access real-time and historical data on system performance.

Within the platform, you can do the following:

* **Check Statuses by Product**. View the current status and daily performance metrics for each specific product offered by Cognigy. This information enables you to monitor the health of the services you rely on.
* **View Past Incidents**. Access detailed incident history to review past issues and their resolutions. This information helps you track the system reliability over time.
* **Subscribe to Updates**. Stay informed by subscribing to updates. You will receive notifications about system performance and incidents through your preferred communication channels.

## Availability

* Cognigy SaaS. The platform is available at [status.cognigy.ai](https://status.cognigy.ai/).

## Subscribe to Updates

To subscribe to updates, follow these steps:

1. Open the Cognigy SaaS Service Health platform.
2. In the upper-right corner, click **Subscribe to Updates**.
3. Select one of the following channels:
   * **Email** — get email notifications whenever an incident occurs or is resolved.
   * **Phone** — get text message notifications whenever an incident occurs or is resolved.
   * **Slack** — get incident updates and maintenance status messages in Slack.
   * **Atom or RSS Feed** — access incident updates via Atom Feed or RSS Feed.
4. Follow the on-screen instructions to complete your subscription.

## More Information

* [Overview](/ai/administer/installation/overview)


# Single Sign-on with SAML 2.0
Source: https://docs.cognigy.com/ai/administer/installation/single-sign-on-saml2

Configure SAML 2.0 single sign-on (SSO) for Cognigy to enable secure user authentication with your organization's identity provider.

Cognigy integrates with popular single sign-on (SSO) [identity providers (IDPs)](/ai/administer/installation/identity-providers), such as Microsoft Entra, OneLogin, and Okta. This integration allows users within your organization to log in to Cognigy without separate credentials.

Using SSO automatically grants access that ensures a streamlined authentication process. Additionally, you can use the robust Access Control tool in Cognigy to effectively manage the access rights of individual users and fulfill security standards. If you need SSO to comply with security standards across your organization, you can [enforce login with SSO](#enforce-login-with-sso).

## Get the SSO URL

To configure an SSO in your IDP, you will need the URL that is used during the SAML authentication process. The SAML requests are sent to the API service, so you should use the API domain that you have configured for your installation. The SSO URL will have the following format:

```txt theme={null}
https://<api-url>/auth/saml/login/<organization-id>
```

For example, the api-url for the app environment is `api-app.cognigy.ai`. The `organization-id` is the `id` of your organization within Cognigy. You can get your `organisationId` on the [My Profile](/ai/agents/develop/ai-agent-memory/profile) page by clicking <img alt="menu" /> **> Copy Organization ID**.

You will need the SSO URL when configuring your IDP using one of the setup guides provided.

To configure the SSO Provider with Cognigy, refer to one of the specific guides for the [supported IDPs](#more-information). These guides provide examples of the required API requests. If your IDP is not listed, we recommend following the [OneLogin guide](https://support.cognigy.com/hc/en-us/articles/360016310699-OneLogin#introduction-0-0) as an example. The API request for configuring SSO with Cognigy is the same, but the configuration values may vary depending on the provider.

## Get the SLO URL

<Warning>
  Single Logout is only supported with OneLogin and Microsoft Entra ID:

  * Service Provider initiated Single Logout is only supported with Microsoft Entra ID.
  * IDP initiated Single Logout is only supported with OneLogin.
</Warning>

To configure Single Logout for your IDP, you will need the URL that is used to process the logout request from the IDP. During the Single Logout process, the IDP will redirect to the frontend of Cognigy. Therefore, you should use the frontend domain that you have configured for your installation. The SLO URL will have the following format:

```txt theme={null}
https://<frontend-url>/slo/<organization-id>
```

For instance, the frontend URL could be `app.cognigy.ai`.

Learn more about connecting to your preferred SSO provider in [Cognigy Help Center](https://support.cognigy.com/hc/en-us/sections/360004563679-Single-Sign-on).

## Change an SSO Configuration in Cognigy

You can have only one SSO configuration for your organization. If you want to change the configuration, you first need to delete it and create a new one. To delete an SSO configuration, send a `POST` request to:

```txt theme={null}
https://<api-url>/v2.0/identityprovider/reset
```

Read more about using the Cognigy API on the [API Reference Page](https://api-trial.cognigy.ai/openapi#post-/v2.0/identityprovider/reset).

## Enforce Login with SSO

Enforcing login with SSO can be useful to standardize the authentication process and meet security standards within your Cognigy installation. By default, you can log in to Cognigy with your credentials or with SSO. However, you can enforce login with SSO for all users across either all organizations or specific ones within your installation. In this case, users can only log in with SSO.

To enforce login with SSO, proceed as follows:

* For shared SaaS installations, contact [Cognigy technical support](/help/get-help). If you use your credentials to log in, you receive an error message. To log in, enter your email address and click **Sign in with SSO**.
* For dedicated SaaS installations, contact [Cognigy technical support](/help/get-help). You can enforce login with SSO for:
  * The entire Cognigy installation — the login with credentials is deactivated for all users. To log in, enter your email address on the login page and click **Sign in with SSO**.
  * Specific organizations — if you use your credentials to log in, you receive an error message. To log in, enter your email address on the login page and click **Sign in with SSO**.
* For on-premises installations, you can enforce login with SSO for:
  * The entire Cognigy installation — set the `DEACTIVATE_NON_SSO_LOGIN` feature flag to `true`. The login with credentials is deactivated for all users. To log in, enter your email address on the login page and click **Sign in with SSO**.
  * Specific organizations — set the `DEACTIVATE_NON_SSO_LOGIN` feature flag to `false` and assign the `organization-id` to the `DEACTIVATE_NON_SSO_LOGIN_ORG_WHITELIST` environment variable in `values.yml`. If you use your credentials to log in to the organizations listed in `DEACTIVATE_NON_SSO_LOGIN_ORG_WHITELIST`, you receive an error message. To log in, enter your email address on the login page and click **Sign in with SSO**.

## Log in with SSO

<Accordion title="First time login">
  During the first login with SSO, you need to log in using your IDP credentials. Cognigy redirects you to your configured IDP. You need to enter your IDP credentials to complete the login.

  <Tabs>
    <Tab title="Log in with Optional SSO">
      1. Go to the login page and click **Login with SSO**.
      2. Enter your email in the **Email** field.
      3. Click **Login with SSO**.
      4. Click **Sign in with SSO**.
    </Tab>

    <Tab title="Log in with Enforced SSO">
      1. Go to the login page and click **Login with SSO**.
      2. Enter your email in the **Email** field.
      3. Click **Sign in with SSO**.
    </Tab>
  </Tabs>
</Accordion>

## More Information

* [Identity Providers](/ai/administer/installation/identity-providers)


# Quotas and Limits
Source: https://docs.cognigy.com/ai/administer/limitations



This page describes the default quotas and limits that apply to Cognigy installations.

* *Limits* are technical or architectural caps that can't be changed (or can only be changed in rare special cases). For example, the maximum size is 2 GB.
* *Quotas* are configurable usage allocations. You can set higher quotas for your on-premises installations.

## Default Quotas and Limits for Cognigy

The table compares values across different installation and environment types:

* **Default** – the baseline value used in both SaaS and on-premises environments. It serves as the reference point for all other installations and is applied as the initial configuration unless explicitly overridden.
* **Trial** – overrides applied in trial environments.
* **App** – overrides applied in shared SaaS installations.

If you have an on-premises Cognigy installation, you can set TTL values system-wide via the `values.yaml` (config map) or organization-wide via [Management UI](/ai/administer/access/management-ui). If you have a dedicated SaaS installation, you can set these values in the [Management UI](/ai/administer/access/management-ui) or by contacting [Cognigy technical support](#submit-a-support-request).

| Quota / Limit Name                                                                                                              | Default                             | Trial  | App    |
| ------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------- | ------ | ------ |
| **Network**                                                                                                                     |                                     |        |        |
| Rate limit<sup>[1](#footnote1)</sup>                                                                                            | -                                   |        |        |
| Waiting period before the `user disconnected` event is triggered                                                                | 3 seconds<sup>[2](#footnote2)</sup> |        |        |
| **Database and HTTP Request Node Calls**                                                                                        |                                     |        |        |
| Maximum query result from database Nodes                                                                                        | 500 kB                              |        |        |
| Maximum [Input](/ai/agents/develop/ai-agent-memory/input) size                                                                  | 65 kB<sup>[3](#footnote3)</sup>     |        | 1.5 MB |
| Maximum [Context](/ai/agents/develop/ai-agent-memory/context) size                                                              | 256 kB<sup>[3](#footnote3)</sup>    |        | 1.5 MB |
| Maximum [Profile](/ai/agents/develop/ai-agent-memory/profile) size                                                              | 65 kB<sup>[3](#footnote3)</sup>     |        | 1.5 MB |
| Maximum HTTP payload size for HTTP Request Node                                                                                 | 65 kB                               |        |        |
| Maximum HTTP response size for HTTP Request Node                                                                                | 524 kB                              | 2.6 MB | 1.5 MB |
| HTTP Request Node timeout<sup>[4](#footnote4)</sup>                                                                             | 8 s                                 |        | 15 s   |
| Maximum JSON-encoded payload size for requests to Cognigy REST API                                                              | 60 kB                               |        |        |
| Minimum HTTP text body (used in CSV)                                                                                            | 100 kB                              |        |        |
| Maximum text-encoded (raw) payload size for requests to Cognigy REST API (for example, used to upload CSV files)                | 2 MB                                |        |        |
| **Projects**                                                                                                                    |                                     |        |        |
| Maximum number of invitations to Projects per user per day<sup>[5](#footnote5)</sup>                                            | 10                                  |        |        |
| **NLU**                                                                                                                         |                                     |        |        |
| Maximum number of example sentences per Intent in Flows                                                                         | 200                                 |        |        |
| Maximum total number of example sentences in Flows, including sentences in attached Flows                                       | 10,000                              | 15,000 |        |
| Maximum total number of Intents in Flows, including Intents in attached Flows                                                   | 2,500                               | 10,000 |        |
| Maximum time an Intent training task is allowed to run                                                                          | 10 m                                |        |        |
| **Intent Trainer**                                                                                                              |                                     |        |        |
| Maximum file size for uploading Intent Trainer records                                                                          | 150 MB                              |        |        |
| **Functions**                                                                                                                   |                                     |        |        |
| Maximum number of outbound HTTP requests per Function execution                                                                 | 5                                   |        |        |
| Maximum HTTP response size in Functions                                                                                         | 1 MB                                |        |        |
| Maximum time a Function can run                                                                                                 | 15 m                                |        |        |
| Maximum time Cognigy Function instance data is stored                                                                           | 30 days                             |        |        |
| Maximum number of active and running Cognigy Functions per organization                                                         | 10                                  |        |        |
| Maximum size of the parameters object used to start a Function instance                                                         | 130 kB                              |        |        |
| **Snapshots**                                                                                                                   |                                     |        |        |
| Maximum size of a Snapshot created or uploaded via the Cognigy UI                                                               | 8 GB<sup>[7](#footnote7)</sup>      |        |        |
| Maximum size of a Snapshot created or uploaded via the Cognigy API                                                              | 2 GB<sup>[8](#footnote8)</sup>      |        |        |
| Maximum number of Snapshots per Project                                                                                         | 10                                  |        |        |
| Lifetime of a downloadable Snapshot before it needs to be re-created                                                            | 1 day                               |        |        |
| **Packages**                                                                                                                    |                                     |        |        |
| Maximum size of a Package created or uploaded via the Cognigy UI                                                                | 8 GB<sup>[7](#footnote7)</sup>      |        |        |
| Maximum size of a Package created or uploaded via the Cognigy API                                                               | 2 GB<sup>[9](#footnote9)</sup>      |        |        |
| **Playbooks**                                                                                                                   |                                     |        |        |
| Maximum number of concurrent Playbook executions                                                                                | 10                                  |        |        |
| **Locales**                                                                                                                     |                                     |        |        |
| Maximum number of Locales per Project                                                                                           | 10                                  |        |        |
| **Extensions**                                                                                                                  |                                     |        |        |
| Extension actions                                                                                                               | 10                                  |        |        |
| Maximum execution time for Extension Nodes                                                                                      | 20 s                                |        |        |
| Maximum size of an individual Extension package                                                                                 | 128 MB                              |        |        |
| Maximum size of file-system cache used for executing Extension Nodes                                                            | 512 MB                              |        |        |
| Maximum number of HTTP requests an Options Resolver can send                                                                    | 5                                   |        |        |
| Maximum HTTP response size in an Options Resolver                                                                               | 1 MB                                |        |        |
| Maximum time an Options Resolver can run                                                                                        | 5 s (max HTTP calls \* 1 s)         |        |        |
| Maximum message size for [untrusted Extensions](/ai/for-developers/extensions)                                                  | 300 KB                              |        |        |
| **Execution and Operations**                                                                                                    |                                     |        |        |
| Maximum number of times a Flow can execute the same path for a single user message before triggering `Infinite Loop Protection` | 4                                   |        |        |
| Maximum number of undo/redo operations to be stored per user in the Flow editor                                                 | 5                                   |        |        |
| **Transformers**                                                                                                                |                                     |        |        |
| Maximum number of HTTP requests that can be sent from an Endpoint transformer                                                   | 1                                   | 2      |        |
| Maximum amount of time an Endpoint transformer can run before it is stopped                                                     | 5 s                                 |        |        |
| **SMTP Node**                                                                                                                   |                                     |        |        |
| Maximum number of retries for sending an outbound email                                                                         | 3                                   |        |        |
| Maximum size of an email attachment                                                                                             | 10 MB                               |        |        |
| **Code Node**                                                                                                                   |                                     |        |        |
| Maximum code execution time                                                                                                     | 1 s                                 |        |        |
| Maximum number of characters allowed in the code editor                                                                         | 200K                                |        |        |
| Maximum API calls per Code Node                                                                                                 | 100                                 |        |        |

## Retention of Records

<Note>
  [Cognigy Insights](/insights/overview) and the Cognigy [OData endpoint](/ai/agents/analyze/odata) rely on the records listed in the **Data** section in the table.
</Note>

| Record Type                                                                                                                        | Default  |
| ---------------------------------------------------------------------------------------------------------------------------------- | -------- |
| **Logs**                                                                                                                           |          |
| Maximum time until a log entry on the Logs page is removed                                                                         | 24 hours |
| **Sessions**                                                                                                                       |          |
| Maximum time until information about a Flow execution session is removed<sup>[6](#footnote6)</sup>                                 | 7 days   |
| **Data**                                                                                                                           |          |
| Maximum time until a Contact Profile that hasn't been read or modified is removed                                                  | 30 days  |
| Maximum time until a [simulation Contact Profile](/ai/agents/test/simulator/overview) that hasn't been read or modified is removed | 5 days   |
| Maximum time until conversation transcripts are deleted                                                                            | 30 days  |
| **Intent Trainer Records**                                                                                                         |          |
| Maximum time Intent Trainer records are stored in the database                                                                     | 10 days  |
| **Audit Events**                                                                                                                   |          |
| Maximum time until audit events are removed                                                                                        | 30 days  |
| **xApp Session**                                                                                                                   |          |
| Maximum time until an xApp Session expires                                                                                         | 30 days  |
| **PCAP File in the Voice Gateway Self-Service Portal**                                                                             |          |
| Maximum time until a PCAP File is removed                                                                                          | 14 days  |

## Submit a Support Request

Change requests can be reviewed only for dedicated SaaS installations. Changes to shared SaaS installations affect functionality and performance for all customers.

If you want to request a change for dedicated SaaS, use this template:

<Accordion title="Quota Change Request Template">
  ```txt theme={null}
  Hello Cognigy Support,

  I would like to raise the quota “<Description>” to <x> value.
  We need this because <short explanation of your use case>.

  Best regards,
  Your Name
  ```
</Accordion>

***

<sup>1</sup>: A rate limit is set to protect the system stability and ensure reliable performance across all Cognigy SaaS installations.

<sup>2</sup>: If you have an on-premises installation, you can change this limit. The `SOCKET_ENDPOINT_DISCONNECT_GRACE_PERIOD` variable lets you set a waiting period (in seconds) before marking a user as disconnected. This delay ensures temporary network issues or brief interruptions don't immediately trigger the `user disconnected` event.

<sup>3</sup>: The limit for the Input, Context, and Profile objects is shared and can be configured for on-premises installations using the `MAX_MEMORY_OBJECT_SIZE` variable. Increasing the object size could lead to performance issues.

<sup>4</sup>: You can change this limit if you have an on-premises installation. For more information, refer to the [HTTP Request Node](/ai/agents/develop/node-reference/service/http-request) article.

<sup>5</sup>: If you have an on-premises installation, you can change this limit by setting the `MAX_ADD_PROJECT_MEMBER_REQUESTS_PER_EMAIL_PER_DAY` environment variable.

<sup>6</sup>: If you have an on-premises
installation, you can change this limit in the [Management
UI](/ai/administer/access/management-ui#change-expiration-values)
or by setting the `MAX_SESSION_STATE_TTL_IN_MINUTES` environment variable. For
some Endpoints, you can also configure the limit in the [Session
Management](/ai/agents/deploy/endpoints/session-management) section. The
session TTL is determined by the lowest value among these three settings.

<sup>7</sup>: Snapshots and Packages use the same maximum size limit. For on-premises installations, you can increase this limit by setting the `RESUMABLE_UPLOAD_MAX_FILE_SIZE` environment variable.

<sup>8</sup>: If you have an on-premises installation, you can change this limit by setting the `SNAPSHOT_MAX_FILE_SIZE` environment variable.

<sup>9</sup>: If you have an on-premises installation, you can change this limit by setting the `PACKAGE_MAX_FILE_SIZE` environment variable.


# My Profile
Source: https://docs.cognigy.com/ai/administer/user-menu/my-profile

Manage your Cognigy profile on the My Profile page, including your preferences, API keys, login history, connected accounts, and personal information.

<a href="/release-notes/2026.8"><Badge>Updated in 2026.8</Badge></a>

*My Profile* lets you manage your user profile, preferences, API keys, and connected accounts.

The **My Profile** page includes:

* [Preferences](#preferences)
* [Global Roles](#global-roles)
* [Login History](#login-history)
* [API Keys](#api-keys)
* [Connected Accounts](#connected-accounts)

## Working with My Profile

Access the **My Profile** page by clicking your user icon in the upper-right corner of the Main page or the bottom-left corner of the Project page.

On the **My Profile** page, you can update preferences, view login history, manage API keys, and more. Click <img alt="menu" /> in the upper-right corner to access the extended menu, where you can select the following actions:

* **Edit Profile** — opens the profile editing page where you can update your personal information.
* **Copy Organization ID** — copies the organization ID associated with your user account. This ID is used for Single Sign-On (SSO) and other integrations.
* **Delete Account** — permanently deletes your user profile from Cognigy, including all API keys and connected accounts. This action is irreversible.

### Preferences

Customize your user experience in Cognigy by setting your preferred language, time zone for Insights, and other interface preferences.

| **Preference**                      | **Description**                                                                                                                                                                                           | **Default Value**         |
| ----------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------- |
| Select Language                     | Sets the language for the user interface. Cognigy supports the following user interface languages:<ul><li>English</li><li>French</li><li>German</li><li>Japanese</li><li>Korean</li><li>Spanish</li></ul> | English                   |
| Select Insights Time Zone           | Defines the time zone for viewing Insights and analytics data.                                                                                                                                            | (UTC+02:00) Europe/Berlin |
| Show Node Count Warnings            | Displays warnings if the number of Nodes in a Flow exceeds best practice limits.                                                                                                                          | Enabled                   |
| Use Contrast Colors for Node Label  | Improves the readability of Node labels by applying higher contrast colors.                                                                                                                               | Enabled                   |
| Use Extended Keyboard Accessibility | Enables additional keyboard shortcuts and focus options to support accessible navigation.                                                                                                                 | Enabled                   |

### Global Roles

View user's [global roles](/ai/administer/access/admin-center/access-control) that apply across multiple Projects. You can configure more granular roles for each individual Project.

### Login History

View a list of previous login sessions and their location (if available). You can use this information for auditing purposes.

### API Keys

<Warning>
  API keys listed with the **Legacy** tag are stored in an outdated storage format and don't comply with the current security standards. If you use legacy API keys, migrate to the new API keys as soon as possible and revoke the legacy API keys.
</Warning>

API keys allow third-party services to access Cognigy in the same way a human user would by entering their password to log in to the platform.

API keys inherit the same permissions as the associated user profile, such as Admin, Project Manager, or OData.

API keys provide authorization to perform requests made available via the [Cognigy Open API](/ai/for-developers/developers/api-and-cli) or to return conversation data from the [Cognigy OData Analytics Endpoint](/ai/agents/analyze/odata).

Only users with [global roles](/ai/administer/access/admin-center/access-control) can manage their API keys.

<AccordionGroup>
  <Accordion title="Create an API Key">
    To create an API key, follow these steps:

    1. On the **My Profile** page, go to the **API Keys** section and click **+**.
    2. Enter a unique name for the API key and click **Confirm**.
    3. Copy the API key by clicking its value and save it securely. For security reasons, the key is shown only once in this window; if not saved, you won't be able to view or copy it again.

    Once the API key is created, its name and creation date will be displayed in the list of API keys.
  </Accordion>

  <Accordion title="Revoke an API Key">
    To revoke an API key, follow these steps:

    1. On the **My Profile** page, go to the **API Keys** section.
    2. Select <img alt="menu" /> **Revoke API Key** next to the API key you want to revoke.

    The API key will be revoked and can no longer be used for authentication in API requests.
  </Accordion>
</AccordionGroup>

### Connected Accounts

Manage third-party accounts that Cognigy connects to, simplifying deployment to supported platforms. Currently, only Amazon accounts are supported.


# User Menu
Source: https://docs.cognigy.com/ai/administer/user-menu/overview

Explore the Cognigy User Menu. Manage preferences, billing, access control, audit events, and quickly access Cognigy products and documentation.

The *User Menu* provides access to various user-specific features.
From the User menu, you can customize preferences, monitor billing and audit events,
manage roles with ACL, access Cognigy products and documentation quickly, and log out.

To open the User Menu, follow these steps:

1. Go to the Cognigy interface.
2. Access the User Menu using one of the following options:
   * Click the **User Menu** icon on the **Main** page in the upper-right corner.
   * Select the User Menu in the bottom-left corner of the **Project** page.

## User Menu Options

The following options are available in the User Menu:

| Option                                                      | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
| ----------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| [My Profile](/ai/administer/user-menu/my-profile)           | The interface where you can manage and customize personal account settings.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
| [Admin Center](/ai/administer/access/admin-center/overview) | The interface for users with the `admin` role, allowing them to manage the organization by monitoring metrics for key resources, controlling access, and tracking system events. It includes the following sections:<ul><li>[Dashboard](/ai/administer/access/admin-center/dashboard) — track billable conversations, Knowledge Chunks, and Queries for your organization. You can also activate additional Cognigy services, such as Live Agent and Voice Gateway, or delete the organization.</li><li>[Access Control](/ai/administer/access/admin-center/access-control) — monitor actions taken by other users within your organization's account.</li><li>[Audit Events](/ai/administer/access/admin-center/audit-events) — the page where users with the Admin role can monitor actions performed by other users within your organization's account.</li></ul> |
| Documentation                                               | The link to access the Cognigy technical documentation.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |
| [Insights](/insights/overview)                              | The link to access Cognigy Insights.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
| [Live Agent](/live-agent/assistants/overview)               | The link to access Cognigy Live Agent.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
| [Voice Gateway](/voice-gateway/overview)                    | The link to access Cognigy Voice Gateway.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
| Logout                                                      | The option allows users to log out of the interface and also provides the ability to view the current product version.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |


# Analytics Steps
Source: https://docs.cognigy.com/ai/agents/analyze/analytics-steps



In Cognigy, *Analytics Steps* are progress markers that help you monitor the Flow paths users take during interactions with AI Agents. These steps are recorded in the Analytics database and can be analyzed in [Insights](/ai/agents/analyze/insights) or retrieved through the [OData Endpoint](/ai/agents/analyze/odata).

There are two types of Analytics Steps:

* [Node Steps](#node-steps)
* [Intent Steps](#intent-steps)

## Working with Analytics Steps

You can create Analytics Steps in your Flows in **Build > Flows**.

### Node Steps

Node Steps are Analytics Steps that record Node executions in the Analytics database:

1. In the Flow editor, click the Node for which you want to add an Analytics Step.
2. In **Settings > Analytics Step** in the Node editor, enter a name for the step in the **Analytics Step** field and save the Node. The step name can't:
   * Contain the following characters: `/`, `:`, `*`, `?`, `"`, `<`, `>`, `|`, `¥`, `!`, `&`, `$`, `§`, `%`, `{`, `}`, `=`, `»`, `,`, `.`, `;`, `+`, `~`, `#`, `` ` ``, `'`, `°`, `µ`, `€`.
   * Start or end with a whitespace.
   * Be longer than 128 characters.

### Intent Steps

Intent Steps are Analytics Steps that record Intent recognition in the Analytics database:

1. In the Flow editor, navigate to **NLU > Intents** and select an Intent.
2. In **Advanced > Analytics step**, enter a name for the step in the **Analytics step** field. The step name can't:
   * Contain the following characters: `/`, `:`, `*`, `?`, `"`, `<`, `>`, `|`, `¥`, `!`, `&`, `$`, `§`, `%`, `{`, `}`, `=`, `»`, `,`, `.`, `;`, `+`, `~`, `#`, `` ` ``, `'`, `°`, `µ`, `€`.
   * Start or end with a whitespace.
   * Be longer than 128 characters.

<Note>
  Analytics Steps exist in the OData records only after a conversation has triggered them at least once.
</Note>

## More Information

* [Collecting Data](/ai/agents/analyze/collecting-data)
* [Insights](/ai/agents/analyze/insights)
* [OData Endpoint](/ai/agents/analyze/odata)
* [Flows](/ai/agents/develop/projects-and-flows/overview)


# Call Tracing
Source: https://docs.cognigy.com/ai/agents/analyze/call-tracing

Record and debug voice AI Agent calls with the Call Tracing feature.

If you create a voice AI Agent and use Voice Gateway as a voice provider, you can record calls while setting up your voice AI Agent for debugging purposes. This approach lets you troubleshoot voice AI Agents, identify what is causing latency, and listen to the voice AI Agents' outputs to ensure they work as expected.

For more information, read the [Recent Calls](/voice-gateway/webapp/recent-calls) article in the Voice Gateway documentation.

## More Information

* [Recent Calls](/voice-gateway/webapp/recent-calls)


# Collecting Data
Source: https://docs.cognigy.com/ai/agents/analyze/collecting-data

Cognigy stores AI Agent data for Insights, OData reporting, and BI tools like Power BI and Tableau.

<a href="/release-notes/2026.1"><Badge>Updated in 2026.1</Badge></a>

Cognigy collects data from conversations with chat and voice AI Agents in PostgreSQL and MongoDB databases. The collected data is used in Insights to display reports and explorers, and by the OData Endpoint to create customized reports and charts in business intelligence systems, such as Power BI, Tableau, and others.

## Data Collections

Analytics data is accessed by connecting to the [OData Analytics Endpoint](/ai/agents/analyze/odata). The set of available collections depends on the OData endpoint version you are using.

The following collections are available in OData 2.3 and 2.4:

<AccordionGroup>
  <Accordion title="Analytics">
    This collection contains one record per user input, with detailed data about the user message, such as NLU results, tasks achieved, and input text. Records are finalized and saved at the end of the session. This collection is different from the others because you can adjust its records during the session using the [Overwrite Analytics](/ai/agents/develop/node-reference/analytics/overwrite-analytics) or [Code](/ai/for-developers/code/overview) Nodes.
  </Accordion>

  <Accordion title="Conversations">
    This collection contains one record per user input, one record per AI Agent output, and one record per human agent input. Conversation records are written immediately after a message arrives from the respective source. These records are commonly used for generating conversation transcripts.
  </Accordion>

  <Accordion title="Steps">
    This collection contains one record per [Analytics Step](/ai/agents/analyze/analytics-steps) in a Flow. You can add Analytics Steps to Nodes or Intents. Both [Question Nodes](/ai/agents/develop/node-reference/basic/question) and [Intents](/ai/platform-features/nlu/intents/overview) contain an Analytics Step by default. The primary function of this record is to keep a list of unique IDs to retrieve Analytics Steps independently of any name changes.
  </Accordion>

  <Accordion title="ExecutedSteps">
    This collection contains one record per executed Analytics Step for every session. You can use this collection to identify the step that occurred prior to the current step.
  </Accordion>

  <Accordion title="Sessions">
    This collection contains one record per session. The main purpose of this record is to provide a list of ordered steps that occurred in each session.
  </Accordion>

  <Accordion title="LiveAgentEscalations">
    This collection contains one record per handover from an AI Agent to a human agent in [Cognigy Live Agent](/live-agent/overview). It includes data about the escalation, such as the session, inbox, team, and assigned agent. Use this collection to analyze handover patterns and human agent workload.
  </Accordion>
</AccordionGroup>

The following collections are available in OData 2.4 only:

<AccordionGroup>
  <Accordion title="Goals">
    This collection contains one record per [goal](/ai/agents/analyze/goals-and-tasks/goals) defined in a Project. Each record includes the goal configuration, including the goal name, description, and version. Use this collection to retrieve the reference data for all goals and link it with the Goal Steps, Goal Step Metrics, and Goal Events collections.
  </Accordion>

  <Accordion title="GoalSteps">
    This collection contains one record per step within a [goal](/ai/agents/analyze/goals-and-tasks/goals). Each Goal consists of an initial step and one or more subsequent steps, at least one of which must be marked as a completion step. Use this collection to retrieve the structure of each goal and understand which steps lead to goal completion.
  </Accordion>

  <Accordion title="GoalStepMetrics">
    This collection contains one record per business metric at goal step level, such as money saved, revenue earned, or time saved. These metrics are configured per goal step and measure the business impact of completing each step. Use this collection to analyze the value delivered at each stage of a goal.
  </Accordion>

  <Accordion title="GoalEvents">
    This collection contains one record per goal-related event that occurs during a session. An event is written each time an end user triggers a goal step in a conversation via a [Track Goal Node](/ai/agents/develop/node-reference/analytics/track-goal). Use this collection to track goal progress and completion rates across sessions.
  </Accordion>
</AccordionGroup>

## Analytics Data Flow

Analytics data is generated throughout the life cycle of a user input in Cognigy. The diagram shows when each type of data is generated and written to the respective collection.

<Frame>
  <img alt="Analytics data flow from user input through NLU processing to database storage" />
</Frame>

1. After the user input is processed by the NLU engine, an Analytics record is generated.
2. While the Flow processes the user input, an [Overwrite Analytics](/ai/agents/develop/node-reference/analytics/overwrite-analytics) or [Code](/ai/for-developers/code/overview) Node can overwrite the originally generated data.
3. When the Flow execution is complete, Cognigy adds the final analytics information, for example, execution time, to the Analytics record before storing it in the database.

## More Information

* [Analytics Steps](/ai/agents/analyze/analytics-steps)
* [Goals and Tasks](/ai/agents/analyze/goals-and-tasks/overview)
* [OData Endpoint](/ai/agents/analyze/odata)


# Contact Profiles
Source: https://docs.cognigy.com/ai/agents/analyze/contact-profiles



<a href="/release-notes/2026.16"><Badge>Updated in 2026.16</Badge></a>

*Contact Profiles* contain information about users who interact with your AI Agents. In the Contact Profiles, you can view user data that was collected during conversations with AI Agents, for example, first name and email address, and navigate through the conversation transcript. With this data, AI Agents can provide personalized interactions and contact management across different channels, such as Webchat, WhatsApp, Slack, and more. You can access the Contact Profiles from a Flow through the [Profile](/ai/agents/develop/ai-agent-memory/profile) object and manage them with specific Nodes.

Contact Profiles offer the following benefits:

* **Data Collection**. You can gather user data during interactions using predefined schemas and custom fields, and store this data in the Profile object.
* **Management Flexibility**. You can manage Contact Profiles flexibly to navigate conversation history and set expiration dates for data.
* **GDPR Compliance**. Cognigy provides tools to comply with the General Data Protection Regulation (GDPR), including exporting user data, deleting Contact Profiles, and managing data collection preferences for individual contacts.

## Limitations

* A Contact Profile is removed after 30 days if not read or modified. If you have an on-premises installation, you can configure an expiration date for Contact Profiles in the [Management UI](/ai/administer/access/management-ui).
* You can't update data in an inactive Contact Profile. If you try to do so, the AI Agent throws an error. To update an inactive Contact Profile, activate it first with the [Activate Profile](/ai/agents/develop/node-reference/analytics/activate-profile) Node, then update the data.
* A [simulated Contact Profile](/ai/agents/test/simulator/overview) is removed after 5 days if not read or modified. If you have an on-premises installation, you can configure an expiration date for simulated Contact Profiles via the `SIMULATION_PROFILE_TTL_IN_MINUTES` variable.

## Profile Schema

The *Profile Schema* organizes the Contact Profile information into fields that you can customize to meet your needs. By default, the Profile Schema has the following fields:

<Accordion title="Default Profile Schema">
  | Field                   | Profile object Key | Description                                                                                                                                                                                                                                          |
  | ----------------------- | ------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
  | User ID                 | profileId          | A unique identifier assigned to each end user.                                                                                                                                                                                                       |
  | First Name              | firstname          | The first name of the end user.                                                                                                                                                                                                                      |
  | Last Name               | lastname           | The last name of the end user.                                                                                                                                                                                                                       |
  | Accepted Privacy Policy | privacy\_policy    | Indicates whether the end user has agreed to the privacy policy terms and specifies the time when the agreement was made. This field is applicable to Webchat v3 when an end user accepts the [Privacy Notice](/webchat/v3/features#privacy-notice). |
  | Accepted GDPR           | accepted\_gdpr     | Indicates whether the end user has consented to the General Data Protection Regulation (GDPR) compliance terms.                                                                                                                                      |
  | Tasks Completed         | tasks              | Refers to all the [tasks](/ai/agents/analyze/goals-and-tasks/tasks) that were completed by the end user during conversations with AI or human agents.                                                                                                |
  | Email                   | email              | The email address of the end user.                                                                                                                                                                                                                   |
  | Gender                  | gender             | The gender identity of the end user, if provided.                                                                                                                                                                                                    |
  | Age                     | age                | The age of the end user, in years.                                                                                                                                                                                                                   |
  | Birthday                | birthday           | The date of birth of the end user.                                                                                                                                                                                                                   |
  | Location                | location           | The geographical location or address information associated with the end user.                                                                                                                                                                       |
  | Memories                | memories           | The data recorded via [Add Memory Nodes](/ai/agents/develop/node-reference/analytics/add-memory).                                                                                                                                                    |
</Accordion>

## Working with Contact Profiles

<Tabs>
  <Tab title="GUI">
    You can view, export, and delete Contact Profiles in **Manage > Contact Profiles**, as well as copy the Contact Profile ID and access the **Contact Profile** page. Use the **Hide Inactive Profiles** option to hide or display inactive Contact Profiles. This option is activated by default.

    You can also use the **Simulation Profiles** option to view only personas created in the [simulation scenarios](/ai/agents/test/simulator/overview).

    The **Contact Profile** page displays a list of conversation transcripts, with the stored Contact Profile data available in the right pane. To view more details about each session, select a session from the list. You are redirected to the [Conversations explorer](/insights/explorers/conversations) in Insights.

    **Flow Editor**

    In the Flow editor, you can manage Contact Profiles using the following Nodes:

    * [Activate Profile](/ai/agents/develop/node-reference/analytics/activate-profile)
    * [Deactivate Profile](/ai/agents/develop/node-reference/analytics/deactivate-profile)
    * [Delete Profile](/ai/agents/develop/node-reference/analytics/delete-profile)
    * [Update Profile](/ai/agents/develop/node-reference/analytics/update-profile)
    * [Merge Profile](/ai/agents/develop/node-reference/analytics/merge-profile)

    With these Nodes, you can create personalized experiences. For example, you can deactivate a Contact Profile if a user requests to stop data collection or update a Contact Profile with new information.
  </Tab>

  <Tab title="API">
    You can view, edit, and delete Contact Profiles using the [Cognigy API](/api-reference/contact-profiles/get-contact-profiles-performance).
  </Tab>
</Tabs>

### Customize the Profile Schema

<Tabs>
  <Tab title="GUI">
    You can add custom fields to the Profile Schema. To add a custom field, follow these steps:

    1. In **Manage > Contact Profiles**, click **Edit Schema**.
    2. In the **Profile Schema** window, click **Add** and configure the following:

    * **Field Name** — enter a name for the custom field.
    * **Internal Name** — enter a system name for the custom field. The internal name:
      * Must consist only of uppercase and lowercase alphanumeric characters and underscores.
      * Is case-insensitive.
      * Must be unique and not duplicate internal names of other parameters.
    * **Type**:
      * **Text** — for string values. This type accepts any combination of characters, including letters, numbers, symbols, and spaces.
      * **Number** — for integers. This type accepts whole numbers within the range of `-9,007,199,254,740,991` to `9,007,199,254,740,991`. Numbers outside this range can't be accurately represented in JavaScript.
      * **YesNo** — for boolean values.
      * **Complex** — for objects. This type accepts nested structures containing other data types.

    3. Click **Save** to confirm your changes.
  </Tab>

  <Tab title="API">
    You can add custom fields to the Profile Schema using the [Cognigy API](https://api-trial.cognigy.ai/openapi#put-/v2.0/profiles/schema/-projectId-).
  </Tab>
</Tabs>

The custom field appears in both the Profile Schema and each Contact Profile.

To delete a custom field, select <img alt="menu" /> **> Delete** next to the field. When you delete a custom field, any data related to this field is removed from all Contact Profiles.

### Deactivate Data Collection for Contact Profiles

User data is stored in all Contact Profiles by default. You can deactivate data collection for specific Contact Profiles. To deactivate data collection for a Contact Profile, follow these steps:

1. In **Manage > Contact Profiles**, select a Contact Profile from the list and deactivate the **Data collection active** setting.
2. Confirm your decision by selecting:

* **Keep** — keeps previously collected data.
* **Delete Data** — deletes all data, including past recordings.

## More Information

* [Analytics](/ai/agents/analyze/overview)
* [Profile Object](/ai/agents/develop/ai-agent-memory/profile)
* [Conversations Explorer](/insights/explorers/conversations)


# Data Expiration in Cognigy
Source: https://docs.cognigy.com/ai/agents/analyze/data-expiration



In Cognigy, data retention for Analytics data is controlled by time-to-live (TTL) settings. TTL defines how long data is stored before automatic deletion.

This article outlines the data expiration logic and provides examples of how different TTL settings affect data retention.

## Data Expiration Logic

Cognigy uses the smallest positive TTL value from the following sources:

* **Management UI Settings**. The TTL is set at the organization level via the Management UI.
* **Infrastructure Variables**. This TTL is set via a server-side environment variable starting with `INSIGHTS_`, for example, `INSIGHTS_MAX_RAW_ANALYTICS_TTL_IN_MINUTES`.
* **Additional Infrastructure Variables**. This TTL is set via server-side environment variables, such as`MAX_CONVERSATION_TTL_IN_MINUTES`.

If no TTL values are positive or valid, the data doesn't expire.

<Accordion title="Example">
  Configuration:

  * Management UI TTL = 6 months (262,800 minutes)
  * Infrastructure Flag TTL = 1 year (525,600 minutes)

  Result: the data will expire after 6 months.
</Accordion>

### TTL for Analytics Data

Analytics data includes information from the Analytics database: Analytics, Sessions, Step Events, Steps, Conversations collections, and Contact Profiles.
The expiration of this data is controlled by environment variables or settings configured in the Management UI.
Each value defines how long a specific type of data is stored.
Once the configured time is reached, the data is automatically deleted.

<AccordionGroup>
  <Accordion title="List of Environment Variables">
    | Data                     | Environment Variable                        | Description                                                                                                                                                                                                                        | Default Value     |
    | ------------------------ | ------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------- |
    | Contact Profiles         | `MAX_CONTACT_PROFILE_TTL_IN_MINUTES`        | Maximum time in minutes for Contact Profiles.                                                                                                                                                                                      | `43200` (30 days) |
    | Conversations            | `MAX_CONVERSATION_TTL_IN_MINUTES`           | Maximum time in minutes for conversations. This variable can be overridden by `INSIGHTS_MAX_CONVERSATIONS_TTL_IN_MINUTES`. For more information, read [TTL for Conversations](#ttl-for-conversations).                             | `43200` (30 days) |
    | Raw Analytics            | `INSIGHTS_MAX_RAW_ANALYTICS_TTL_IN_MINUTES` | Time in minutes after which raw analytics in Insights data expires. `0` means no expiration.                                                                                                                                       | `0`               |
    | Conversations (Insights) | `INSIGHTS_MAX_CONVERSATIONS_TTL_IN_MINUTES` | Time in minutes after which conversation data in Insights expires. `0` means no expiration. This variable overrides `MAX_CONVERSATION_TTL_IN_MINUTES`. For more information, read [TTL for Conversations](#ttl-for-conversations). | `0`               |
    | Step Events              | `INSIGHTS_MAX_STEP_EVENTS_TTL_IN_MINUTES`   | Time in minutes after which step event data in Insights expires. `0` means no expiration.                                                                                                                                          | `0`               |
    | Steps                    | `INSIGHTS_MAX_STEPS_TTL_IN_MINUTES`         | Time in minutes after which step data in Insights expires. `0` means no expiration.                                                                                                                                                | `0`               |
    | Sessions                 | `INSIGHTS_MAX_SESSIONS_TTL_IN_MINUTES`      | Time in minutes after which session data in Insights expires. `0` means no expiration.                                                                                                                                             | `0`               |
  </Accordion>

  <Accordion title="Examples">
    | Management UI TTL (minutes) | `INSIGHTS_MAX_RAW_ANALYTICS_TTL_IN_MINUTES` | Analytics Expires After |
    | --------------------------- | ------------------------------------------- | ----------------------- |
    | 525,600 (1 year)            | 262,800 (6 months)                          | 262,800 (6 months)      |
    | 0                           | 525,600 (1 year)                            | 525,600 (1 year)        |
    | 262,800 (6 months)          | 525,600 (1 year)                            | 262,800 (6 months)      |
    | 525,600 (1 year)            | 0                                           | 525,600 (1 year)        |
    | 0                           | 0                                           | Doesn't expire          |
  </Accordion>
</AccordionGroup>

#### TTL for Conversations

Conversation data in Cognigy expires according to the following data expiration logic:

1. The system first checks the environment variable: `INSIGHTS_MAX_CONVERSATIONS_TTL_IN_MINUTES`. If this variable is set and has a positive value, it takes precedence.
2. If the `INSIGHTS_MAX_CONVERSATIONS_TTL_IN_MINUTES` variable isn't set or invalid, it falls back to `MAX_CONVERSATION_TTL_IN_MINUTES`.
3. The TTL set in the Management UI is always included in the final check.
4. The system picks the smallest positive TTL from the three sources to decide when conversation data expires.

<Accordion title="Examples">
  | Management UI TTL (min) | `INSIGHTS_MAX_CONVERSATIONS_TTL_IN_MINUTES` | `MAX_CONVERSATION_TTL_IN_MINUTES` | Conversation Expires After |
  | ----------------------- | ------------------------------------------- | --------------------------------- | -------------------------- |
  | 0                       | 0                                           | 0                                 | Doesn't expire             |
  | 40                      | –                                           | 60                                | 40 minutes                 |
  | 0                       | –                                           | 30                                | 30 minutes                 |
  | 40                      | –                                           | 30                                | 30 minutes                 |
  | 90                      | 60                                          | 120                               | 60 minutes                 |
  | 30                      | 45                                          | 25                                | 25 minutes                 |
  | –                       | 20                                          | 60                                | 20 minutes                 |
  | –                       | –                                           | 10                                | 10 minutes                 |
</Accordion>

### TTL for Simulator Results

Simulation results have a fixed retention period that isn't configurable through the Management UI or environment variables.

| Data               | Retention Period | Configurable |
| ------------------ | ---------------- | ------------ |
| Simulation Results | 90 days          | No           |

For more information, read [Simulator Limitations](/ai/agents/test/simulator/overview#limitations).

## More Information

* [Quotas and Limits](/ai/administer/limitations)
* [Data Expiration in Insights](/insights/data-management/data-expiration)


# Goals
Source: https://docs.cognigy.com/ai/agents/analyze/goals-and-tasks/goals



<Note>
  Goals is available only in the new [Insights framework](/insights/data-management/pre-aggregation) and is being released through a phased rollout. Once it becomes available for your environment, your Cognigy business representative will notify you via email.
  You can also explore the Goals Report in the [Cognigy trial environment](https://trial.cognigy.ai/). We encourage you to try this feature out and provide us with feedback.
</Note>

*Goals* are structured objectives that require specific steps to be completed in AI Agent interactions. You can use goal steps to track user actions with [Track Goal Nodes](/ai/agents/develop/node-reference/analytics/track-goal) and tie these actions to measurable business metrics. This setup ensures alignment of your AI Agents with organizational objectives.

With goals, you can evaluate and enhance the business impact of your AI Agent:

* **Track user interactions**. Monitor specific user interactions during conversations to ensure they align with your business objectives.
* **Monitor different goals and their steps**. Track different goals and their corresponding steps to get a comprehensive view of the user journey in conversations.
* **Measure business value**. Goals are linked to business metrics, allowing you to assess the value of AI Agents in achieving key business objectives. Metrics, such as time and money for each completed goal and step, provide a detailed view of the AI Agent's performance.
* **Gain valuable insights**. Measure how effectively your AI Agent conversations achieve business objectives and optimize AI Agents to improve performance and user satisfaction.

## Limitations

* By default, the maximum number of steps per goal is 10. You can increase this number to 20 by using the `FEATURE_USE_GOALS__MAX_STEPS_PER_GOAL` variable in the `values.yaml` file.

## Working with Goals

You can create, edit, and delete goals and their steps, as well as export them as a Package in **Build > Goals**.

When you delete a goal or a step, all associated analytics data is deleted and the Track Goal Node linked to the goal stops working. Before deleting the goal or step, either replace it with another goal or step in the Flow, or remove the linked Node if it's no longer needed.

### Determine What You Want to Measure

Decide whether you want to track money, time, or both:

* **Money Saved**. Track cost savings. To set the currency, go to **Manage > Settings > Currency Settings** and select a currency type from the **Set Standard Currency** list. You can select only one type of currency per Project.
* **Revenue Earned**. Monitor the additional revenue you earn by automating the process with the AI Agent. To set the currency, go to **Manage > Settings > Currency Settings** and select a currency type from the **Set Standard Currency** list. You can select only one type of currency per Project.
* **Time Saved**. Measure the amount of time a person spends on the task. This metric shows the time human agents save when the goal is achieved.

### Implement Goals in Flows

To implement goals in your Flows, follow these steps:

<AccordionGroup>
  <Accordion title="1. Create Goals">
    1. In **Build > Goals**, click **+ New Goal** and configure the parameters of a goal:

    * **Name** — this field is mandatory. Enter a name that reflects the goal. For example, `Reduce Call Handling Time` or `Increase Automation Efficiency`.
    * **Description** — this field is optional. Provide a brief description that reflects the goal. For example, `This goal aims to reduce average call handling time by automating common customer requests`.

    2. In the **Initial Step** section, complete the following fields:

    * **Name** — this field is mandatory. Provide a descriptive name for the starting step. For example, `Initiate AI Agent Conversation`. The initial step is required, as the goal activates when the Flow triggers this step.
    * **Description** — this field is optional. Add a detailed description of the starting step. For example, `This step triggers when a user begins an AI Agent session`.

    3. In the **Steps** section, click **+ New Step** and add at least one step to complete the goal configuration. If you add only one step, it must be marked as a Completion Point.
    4. Configure the step parameters:

    * **Name** — this field is mandatory. Enter a name that reflects the step. For example, `Customer Request Handled`.
    * **Description** — this field is optional. Describe the step. For example, `This step captures when the AI Agent successfully resolves a customer request`.
    * **Mark this step as a completion step** — activate this setting to mark the current step as the point where the goal is considered fully completed. For example, this step can reflect the key result when the AI Agent has resolved all user requests. At least one step in the goal must be marked as completed. Multiple steps can be marked as completed.

    5. In the **Success Metrics** field, measure the impact of your goal by tracking the following performance metrics:

    * **Cost Savings** — track saved money by entering values in the format `xx.xx`. For example, `150.00` represents `$150`. Make sure your currency is configured in the **Currency Settings** section in the [Project Settings](#determine-what-you-want-to-measure).
    * **Revenue Earnings** — track earnings by entering values in the format `xx.xx`. For example, `150.00` represents `$150`. Make sure your currency is configured in the **Currency Settings** section in the [Project Settings](#determine-what-you-want-to-measure).
    * **Time Saved** — track saved time by entering values in the format `00:00:00` (hours:minutes:seconds). For example, `02:45:00` represents 2 hours and 45 minutes.
  </Accordion>

  <Accordion title="2. Add Goals to Flows">
    1. In **Build > Flows**, select a Flow or click **+ New Flow** to create one.
    2. In the Flow editor, add a Track Goal Node to monitor goal progress in the Flow.
    3. Configure the Node settings:

    * **Select Goal** — choose the goal you recently created, such as `Reduce Call Handling Time`.
    * **Select Steps** — choose the goal steps that you want to track. The first selected step must be the initial step. Set the initial step at the beginning of the process you want to track in your Flow. Otherwise, the goal isn't tracked.

    You can add multiple Track Goal Nodes to the same Flow to track different steps and KPIs.

    By defining multiple steps and associating them with predefined metrics, goals provide a more comprehensive way to evaluate success compared to [Tasks](/ai/agents/analyze/goals-and-tasks/tasks). While using multiple steps, you must still define the initial and completion steps. Without a completion step, your goal is partially completed or incomplete if only the initial step is triggered.
  </Accordion>
</AccordionGroup>

### Track Goals in Insights

After having added goals and their steps to Flows, you can monitor metrics in the Insights [Goals report](/insights/reports/goals).

## Example

In this example, you will configure goals and steps in a Flow, such as entering passenger information and completing payment. This setup lets you track the pre-booking process and monitor and analyze each step of the booking process. Follow these instructions:

<AccordionGroup>
  <Accordion title="1. Configure Goals">
    1. [Create a Goal](#implement-goals-in-flows) with the name `Pre-book a Flight`.
    2. Set up these steps in this Goal:

    * **Initial Step** — navigate to the **Initial Step** section and enter `Select booking options` in the **Name** field.
    * **Further Steps** — use the **+ New Step** option to add two further steps with the following names:
      * **Passenger Details** — at this step, set the time savings metric to `3 minutes` in the **Time Saved** field.
      * **Confirm Booking** — at this step, set the cost savings metric to `$15` in the **Cost Savings** field, then activate the **Mark this step as a completion step** option.
  </Accordion>

  <Accordion title="2. Create Flows and Configure the Initial Step">
    1. [Create a Flow](/ai/agents/develop/projects-and-flows/overview#working-with-flows) and add a Question Node to the Flow.
    2. Open the Question Node and select the **Text with Buttons** output type.
    3. In the **Text** field, enter the following text:

    ```txt theme={null}
    Hello, I’m Mr. Traveler, and I can help you find discounts on tours to Europe.  
    At the moment, there are no discounted tours available. However, you can leave a request for pre-booking, and I’ll assist you in placing the order. Alternatively, you can contact an operator, though that process may take longer.  

    What works best for you?
    ```

    4. Click **Add Button** and configure the first button:
    5. In the **Button Title** field, enter `pre-booking`.
    6. From the **Select Button Type** list, select **Postback Value**.
    7. In the **Postback Value** field, enter `pre-booking` and save.
    8. Click **Add Button** and configure the second button:
    9. In the **Button Title** field, enter `contact an operator`.
    10. From the **Select Button Type** list, select **Postback Value**.
    11. In the **Postback Value** field, enter `contact an operator` and save.
    12. Below the Question Node, add a Track Goal Node. In the Node editor, select the initial step you created earlier and save. This Node will initiate the goal tracking process.
    13. Below the Track Goal Node, add an If Node. In the Node editor, configure the `Answer contains pre-booking` condition:
    14. In the **Condition** field, click <img alt="token" /> and select the **Answer** Token.
    15. From the dropdown list, select **contains**.
    16. Specify `pre-booking` below the list and save.

    <figure>
      <img alt="Flow with Question, Track Goal, and If Nodes for pre-booking" />
    </figure>
  </Accordion>

  <Accordion title="3. Configure the Passenger Details Step">
    1. Below your Then child Node, add a Question Node. Below the Else child Node, add a Handover to Human Agent Node.
    2. Add a Question Node.
       In the Node editor, configure the following fields:
    3. From the **Question Type** list, select **Text**.
    4. From the **Output Type** list, select **Text** and enter the following text:

    ```txt theme={null}
    I'll just need your full name and contact information to finalize the booking.
    ```

    3. Add a new Track Goal Node. In the Node editor, select **Passenger Details Step** from the **Steps** list and save. This Node will record how much time is saved when the user submits the data manually.
  </Accordion>

  <Accordion title="4. Configure the Confirm Booking Step">
    1. Below the Track Goal Node, add a Question Node.
    2. Add a Question Node.
    3. Open the Node editor and select **Email** from the **Question Type** list.
    4. Select the **Text with Buttons** output type.
    5. In the **Text** field, enter the following text:

    ```txt theme={null}
    As soon as new offers become available, you will receive an email about your pre-booking.  

    Please confirm your booking by entering your email.
    ```

    1. In the **Button Title** field, enter `pre-booking`.
    2. From the **Select Button Type** list, select **Postback Value**.
    3. In the **Postback Value** field, enter `pre-booking` and save.
    4. Below the Question Node, add a Track Goal Node. In the Node editor, select **Confirm Booking** from the **Steps** list and save. This Node will record how much money is saved when the user submits the data manually.
    5. Below the Track Goal Node, add a Say Node. In the Node editor, enter the following text in the **Text** field:

    ```txt theme={null}
    Your email: <Answer>  

    Thank you for the confirmation!
    ```

    Replace `<Answer>` with the **Answer** Token by clicking <img alt="token" />.

    <figure>
      <img alt="Flow with Track Goal and Say Nodes for booking confirmation" />
    </figure>
  </Accordion>

  <Accordion title="5. Test your Flow">
    Test your Flow via the [Interaction Panel](/ai/agents/test/interaction-panel/overview). Activate [debug mode](/ai/agents/test/interaction-panel/chat#debug-mode) to see when Track Goals are triggered. To view recorded metrics, go to [Insights](/insights/index) and select [Goals](/insights/reports/goals) from the left-side menu.
  </Accordion>
</AccordionGroup>

## More Information

* [Track Goal](/ai/agents/develop/node-reference/analytics/track-goal)
* [Insights Goals Report](/insights/reports/goals)


# Goals and Tasks
Source: https://docs.cognigy.com/ai/agents/analyze/goals-and-tasks/overview



*Goals* and *Tasks* are objectives you can use to track and measure specific outcomes or achievements in a Flow. These objectives help you monitor your AI Agents' performance.

A goal or a task represents a specific target in a Flow that you want to achieve. For example, you can set a goal or a task to track whether a user has completed a registration process or provided the necessary information.

Both goals and tasks offer the following benefits:

* **Enhanced Performance Monitoring**. By tracking how often certain goals or tasks are achieved, you can evaluate the success rate of different interactions or processes.
* **Advanced Reporting and Analytics**. Achieved goals and tasks are available in various reports in Insights and OData. You can analyze this data to track trends and improve your conversational strategies.
* **Better Decision Making**. With detailed data on goal and task achievements, you can make informed decisions about optimizing Flows and improving the AI Agent's performance.
* **Focused Analysis**. If you need a detailed analysis of specific interactions, you can combine goals and tasks with [Analytics steps](/ai/agents/analyze/analytics-steps) to get a comprehensive view of user interactions.

## Life Cycle of Goals and Tasks

<Frame>
  <img alt="Life cycle of goals and tasks from definition to analysis" />
</Frame>

Goals and tasks have the following life cycle:

1. The conversation designer defines a goal or task.
2. The conversation designer configures the goal or task settings.
3. The end user triggers the goal or task in the conversation with an AI Agent.
4. The system records the goal or task completion and stores it in the Analytics database.
5. The system processes the result to be available for retrieval through the OData endpoint and renders new data on the Insights reports.
6. The conversation designer accesses and analyzes data to review the results.
7. The conversation designer improves the Flow.

## Comparison Table

The table shows the differences between the following analytics tools: Analytics steps, tasks, and goals.

| Property                 | Analytics Steps                                                                                                                                      | Tasks                                                                                                                                                                                                                                                                                                                                 | Goals                                                                                                                                                                        |
| ------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Purpose                  | Collects detailed data about each event or interaction.                                                                                              | Records when defined tasks are achieved. Measures the success of predefined objectives or tasks.                                                                                                                                                                                                                                      | Records when defined goals are achieved. Offers advanced analysis tools and more detailed reporting.                                                                         |
| Configuration            | Add an Analytics step to a Node. If the Node is executed, the step is recorded in the Analytics database.                                            | Add a [Complete Task](/ai/agents/develop/node-reference/analytics/complete-task) Node to the Flow.                                                                                                                                                                                                                                    | Create a goal and include steps in the goal. Add [Track Goal Nodes](/ai/agents/develop/node-reference/analytics/track-goal) in the Flow and set the goals you want to track. |
| Integration              | Integration with the OData endpoint [Steps](/ai/agents/analyze/odata#steps) and [ExecutedSteps](/ai/agents/analyze/odata#executedsteps) collections. | Integration with the [Contact Profile](/ai/agents/analyze/contact-profiles), the Insights [Overview](/insights/reports/overview) and [Engagement](/insights/reports/engagement) reports, as well as the OData endpoint [Sessions](/ai/agents/analyze/odata#sessions) and [Analytics](/ai/agents/analyze/odata#analytics) collections. | Improved integration with Insights through the [Goals report](/insights/reports/goals) and OData, and enhanced data reporting features.                                      |
| Contact Profiles records | Not recorded in the Contact Profile.                                                                                                                 | Recorded in the Contact Profile.                                                                                                                                                                                                                                                                                                      | Recorded in the Contact Profile.                                                                                                                                             |

## More Information

* [Analytics steps](/ai/agents/analyze/analytics-steps)
* [OData](/ai/agents/analyze/odata)
* [Insights](/insights/overview)


# Tasks
Source: https://docs.cognigy.com/ai/agents/analyze/goals-and-tasks/tasks



<a href="/release-notes/4.100"><Badge>Updated in 4.100</Badge></a>

*Tasks* let you monitor when specific events occur in a Flow.

To track tasks in a Flow, use a [Complete Task](/ai/agents/develop/node-reference/analytics/complete-task) Node.

Once the Node is triggered,
the task is recorded in the Analytics database and can be monitored in the [Contact Profile](/ai/agents/analyze/contact-profiles),
viewed in Insights [reports](/insights/reports/overview) and [explorers](/insights/explorers/conversations#session-transcript-details), and accessed through the [OData endpoint](/ai/agents/analyze/odata) for further analysis.

For more information, read the [Complete Task Node](/ai/agents/develop/node-reference/analytics/complete-task) article.

## Compatibility Alias to `profile.tasks`

In the [Profile](/ai/agents/develop/ai-agent-memory/profile) object, `profile.goals` has been renamed to `profile.tasks`.
To ensure your existing code still works in Flows, `profile.goals` now functions as a compatibility alias that points to `profile.tasks`. This means you can still use `profile.goals`. Its value is stored in the `profile.tasks` object. Accessing `profile.goals` provides the value from `profile.tasks`, for example:

* When using the `{{ profile.goals }}` or `{{ cp.goals }}` CognigyScript expressions.
* In Code Nodes, Extension Nodes, and Functions.

The `profile.goals` alias is available for existing implementations to maintain continuity. However, for new Flows, use the new `profile.tasks` naming convention in your CognigyScript expressions and code. This approach ensures that your Flows comply with the latest platform standards and best practices.

### Restrictions of the Alias

This compatibility alias for `profile.goals` isn't a standard property of the Profile object.
This alias limits how you can use or inspect it:

* **Property Listing** – `profile.goals` doesn't appear when you list the object's properties, for example, when using `Object.keys(profile)`.
* **JSON Serialization** – `profile.goals` doesn't appear when you turn the Profile object into JSON, for example, when using `JSON.stringify(profile)`.

## More Information

* [Goals](/ai/agents/analyze/goals-and-tasks/goals)
* [Overview](/ai/agents/analyze/goals-and-tasks/overview)


# Insights
Source: https://docs.cognigy.com/ai/agents/analyze/insights

Use Cognigy Insights to view analytics and visualizations of your AI Agent interactions, including reports and metrics.

# Insights

Insights provides you with advanced analytics and visualizations for each [Project](/ai/agents/develop/projects-and-flows/projects) based on the data collected from interactions with chat and voice AI Agents.

## Key Features

* **Data Visualization**. Insights offer various visualization tools such as reports and explorers to represent data in a visually appealing and easily understandable format. This way, you can quickly identify trends, patterns, and anomalies within their conversational data.
* **Metric Tracking**. Insights provide metrics related to user interactions, engagement levels, conversation flow, and more, helping track the performance of their chat and voice AI Agents over time.
* **Performance Analysis**. You can analyze the effectiveness of their conversational AI implementations by examining metrics such as response times, completion rates, user satisfaction scores, and goal achievements.
* **Data Export**. Insights offer the ability to export reports for further analysis in CSV format.

To learn more about Insights, go to the [Insights](/insights/overview) documentation.

## Limitations

Customizing reports in Cognigy Insights isn't possible. Use the [OData Analytics Endpoint](/ai/agents/analyze/odata) to access the data used to populate the reports.


# Cognigy OData Analytics Endpoint
Source: https://docs.cognigy.com/ai/agents/analyze/odata

Cognigy provides an OData analytics Endpoint for retrieving raw data via GET requests. It supports enterprise analytics needs, enabling detailed queries for spreadsheets and dashboard creation with preferred data visualization tools.

<a href="/release-notes/2026.18"><Badge>Updated in 2026.18</Badge></a>

[Open Data Protocol (OData)](https://www.odata.org/) is a standard for building and using RESTful APIs. This protocol simplifies data access for web clients by providing a consistent way to describe data models and perform CRUD operations (create, read, update, delete) using standard HTTP methods and URLs.

Based on the OData protocol version 4,
Cognigy offers an *OData Analytics endpoint* to retrieve raw data using `GET` requests.
This endpoint meets all your enterprise analytics needs,
enabling you to make fine-grained queries in your spreadsheets
or build rich dashboards for AI Agents with your favorite data visualization tool.

Watch this video to get a quick overview of what the OData Analytics endpoint is, how it works, and in which scenarios you can use it:

<iframe title="Cognigy Sessions: Cognigy Insights & OData" />

To analyze your AI Agent's performance, you can use the [built-in Insights reports](/insights/overview) as an alternative to OData or in addition to it.

## Limitations

<Accordion title="OData Endpoint Field Size and Length Limitations">
  * By default, the OData endpoint fields have the following limitations by data type:

  | Type                    | Limitation                                                                              |
  | ----------------------- | --------------------------------------------------------------------------------------- |
  | String                  | The maximum size of a string field is 16 MB.                                            |
  | Number (double)         | The maximum and minimum values for a double are approximately ±1.7976931348623157E+308. |
  | Number (32-bit integer) | The maximum and minimum values for a 32-bit integer are ±2,147,483,647.                 |
  | Number (64-bit integer) | The maximum and minimum values for a 64-bit integer are ±9,223,372,036,854,775,807.     |

  * The OData endpoint fields in the following table have a maximum length that differs from the default limitations:

  | Fields                                                                                    | Maximum length |
  | ----------------------------------------------------------------------------------------- | -------------- |
  | endpointName, localeName, projectName, snapshotName, flowReferenceId                      | 200 characters |
  | state, userType, channel, flowLanguage, intent, intentFlow, flowName                      | 256 characters |
  | custom1, custom2, custom3, custom4, custom5, custom6, custom7, custom8, custom9, custom10 | 512 characters |
  | label, stepLabel                                                                          | 128 characters |
  | entityReferenceId                                                                         | 36 characters  |
</Accordion>

## Restrictions

* The Cognigy OData endpoint supports only the [OData protocol version 4](https://www.odata.org/documentation/).
* The Cognigy OData endpoint only supports `GET` requests and doesn't support any other request types, such as `PATCH`, `DELETE`, or `POST`.
* The data retrieved through the Cognigy OData endpoint isn't updated in real time and may not reflect the current state.
* The Cognigy OData endpoint supports a limited set of system query options, operators, and query functions. For detailed information about the query syntax, refer to the [OData Query Syntax](#odata-query-syntax) section.

## Supported Versions

| Cognigy version  | Cognigy OData endpoint version | Status                                   |
| ---------------- | ------------------------------ | ---------------------------------------- |
| 4.65 and later   | 2.4                            | Supported                                |
| 4.48 and later   | 2.3                            | Supported; deprecation in February, 2026 |
| 4.31-4.47        | 2.0–2.3                        | Not supported                            |
| 4.30 and earlier | 2.0–2.2                        | Not supported                            |

## Data Protection

You can control the data available through the OData endpoint at the [Project](/ai/administer/access/project-settings), [Endpoint](/ai/agents/deploy/endpoints/overview), and Node levels:

* Project — use [Data Redaction](/ai/platform-features/data-redaction) to detect and redact data, including personally identifiable information (PII), from the OData endpoint.
* Endpoint — go to the [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics) section in the Endpoint settings:
  * If you disable the **Collect Analytics** setting, no analytics data is logged or available in OData.
  * If you enable the **Mask Sensitive Analytics** setting, the `inputText` and `inputData` fields are masked.
* Node — use the **Mask and keep analytics data** parameter in the [Delete Profile](/ai/agents/develop/node-reference/analytics/delete-profile) Node and the **Keep analytics records, mask personal data** option in the [Deactivate Profile](/ai/agents/develop/node-reference/analytics/deactivate-profile) Node. These Nodes can mask personal data while including analytics data across the Analytics, Conversations, and Sessions collections. By default, rows with masked personal data in these collections are included in OData results. To exclude rows where personal data has been masked:
  * For SaaS installations, contact [Cognigy technical support](/help/get-help).
  * For on-premises installations, set `ODATA_FILTER_MASKED_ROWS=true` in `values.yaml`.

## OData Endpoint Access

You can access the OData endpoint through a different domain than the one from Cognigy:

<Tabs>
  <Tab title="Cognigy">
    Depending on the type of Cognigy installation you have, the OData endpoint domain is as follows:

    * Trial — `https://odata-trial.cognigy.ai`.
    * Shared SaaS — `https://odata-<your-domain>.cognigy.ai`.
    * Dedicated SaaS — contact your Customer Success Manager or Cognigy Support to get the OData endpoint URL.
    * On-premises — `https://odata-<backend-url>`, where `<backend-url>` refer to the URL assigned to the `BACKEND_BASE_URL_WITH_PROTOCOL` in `values.yaml`.
  </Tab>

  <Tab title="NiCE CXone">
    Use the following pattern: `https://cognigy-odata-<region>.nicecxone.com`, where `<region>` matches your NiCE CXone cluster. For example:

    * `https://cognigy-odata-na1.nicecxone.com`
    * `https://cognigy-odata-na3.nicecxone.com`
    * `https://cognigy-odata-eu1.nicecxone.com`

    For a full list of available regions, refer to [NiCE CXone Shared SaaS](/ai/administer/installation/ip-ranges-shared-environments#nice-cxone-shared-saas-clusters).
  </Tab>
</Tabs>

## Authentication

To connect to the OData endpoint, you need to have the `odata` [global role](/ai/administer/access/admin-center/access-control#before-you-begin) and an [API Key](/ai/administer/user-menu/my-profile#api-keys).

To authenticate your API requests to the OData endpoint, you need the following parameters:

* The Cognigy OData base domain. It is the root URL where your OData service is hosted.
* The API version of the Cognigy OData endpoint. Currently, Cognigy supports OData 2.3 and 2.4.
* The [collection](#collections) in the OData service you want to access.
* The API key that authenticates and authorizes the API access. Generate the API key in your [profile](/ai/administer/user-menu/my-profile#api-keys).

### Authentication Methods

You can use the following methods to authenticate your requests to the OData endpoint:

* [HTTP Headers](#http-headers-examples) — this method provides more security to access the OData endpoint. HTTP header authentication prevents API keys from exposure in the URL, server access logs, and other security risks. This method doesn't support Microsoft Excel and Power BI integrations.
* [Query Parameters](#query-parameters-example) — allows for [Microsoft Excel and Power BI integrations](#authentication-for-microsoft-excel-and-power-bi). For this method, ensure appropriate access controls and security measures to prevent security risks.

#### Examples

The following examples use the following parameters:

* Cognigy OData base domain — `odata-trial.cognigy.ai`
* Cognigy OData endpoint version — `v2.4`
* OData collection — `Analytics`
* API key — `123abc`

<Tabs>
  <Tab title="HTTP Headers">
    <CodeGroup>
      ```curl cURL theme={null}
      curl -H "apikey: 123abc" \
      "https://odata-trial.cognigy.ai/v2.4/Analytics"
      ```

      ```javascript JavaScript theme={null}
      fetch('https://odata-trial.cognigy.ai/v2.4/Analytics', {
        headers: {
          'apikey': '123abc'
        }
      });
      ```

      ```python Python theme={null}
      import requests
      requests.get('https://odata-trial.cognigy.ai/v2.4/Analytics', headers={'apikey': '123abc'})
      ```

      ```nodejs Node.js theme={null}
      const axios = require('axios');
      axios.get('https://odata-trial.cognigy.ai/v2.4/Analytics', { headers: { 'apikey': '123abc' } });
      ```
    </CodeGroup>
  </Tab>

  <Tab title="Query Parameters">
    `https://odata-trial.cognigy.ai/v2.4/Analytics?apikey=123abc`
  </Tab>
</Tabs>

If you are using the query parameters method to authenticate your requests to the OData endpoint, you can migrate to the HTTP headers method without breaking changes as follows:

<Accordion title="Migrate from Query Parameters to HTTP Headers">
  If you currently use query parameters authentication with `https://odata-trial.cognigy.ai/v2.4/Analytics?apikey=123abc`, you can migrate to HTTP headers authentication by using the following HTTP headers:

  <CodeGroup>
    ```curl cURL theme={null}
    curl -H "apikey: 123abc" \
    "https://odata-trial.cognigy.ai/v2.4/Analytics"
    ```

    ```javascript JavaScript theme={null}
    fetch('https://odata-trial.cognigy.ai/v2.4/Analytics', {
      headers: {
        'apikey': '123abc'
      }
    });
    ```

    ```python Python theme={null}
    import requests
    requests.get('https://odata-trial.cognigy.ai/v2.4/Analytics', headers={'apikey': '123abc'})
    ```

    ```nodejs Node.js theme={null}
    const axios = require('axios');
    axios.get('https://odata-trial.cognigy.ai/v2.4/Analytics', { headers: { 'apikey': '123abc' } });
    ```
  </CodeGroup>
</Accordion>

### Authentication for Microsoft Excel and Power BI

When using Microsoft Excel and Power BI, you might be prompted to authenticate. In this case, select **Anonymous Authentication**.

## OData Query Syntax

The following rules define how OData queries must be structured when using system query options, operators, and string functions.

These syntax requirements ensure that requests are valid, predictable, and compatible with the OData v4 protocol.
They apply to [all Cognigy OData endpoint Collections](#cognigy-ai-odata-collections) and must be followed when creating query expressions.

<Accordion title="List of Requirements">
  | Topic            | Details                                                                                                                                                                                                                                       |
  | ---------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
  | Date Format      | Dates must use ISO 8601 format. Base format: `YYYY-MM-DDThh:mm:ssZ`. Milliseconds are supported using `YYYY-MM-DDThh:mm:ss.sssZ`. Example: `2024-01-01T00:00:00.000Z`. `.sss` represents milliseconds (3 digits), and `Z` indicates UTC time. |
  | Case Sensitivity | OData system query options (`$filter`, `$select`) and operators (`eq`, `gt`, `and`) are case-sensitive.                                                                                                                                       |
  | String Literals  | String values inside `$filter` must be enclosed in single quotes. Example: `userId eq '12345'`.                                                                                                                                               |
</Accordion>

### System Query Options

The Cognigy OData endpoint supports a limited set of [system query options](https://docs.oasis-open.org/odata/odata/v4.01/odata-v4.01-part1-protocol.html#sec_SystemQueryOptions):

| Query Option                       | Purpose                                                                                                                                                                                                                                     |
| ---------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `$filter`                          | Filters results based on conditions.                                                                                                                                                                                                        |
| `$skip`                            | Skips a number of records. The `$skip` query at high values, such as `10000`, may impact performance and cause API timeouts. In these cases, use [keyset pagination](#keyset-pagination), for example, with the `timestamp` or `_id` field. |
| `$top`                             | Limits the number of returned records.                                                                                                                                                                                                      |
| `$orderby`                         | Sorts the returned results by one or more fields. You can specify ascending (`asc`) or descending (`desc`) order. If no direction is provided, ascending order is used by default.                                                          |
| `$select`                          | Selects specific fields.                                                                                                                                                                                                                    |
| `$count`<sup>[1](#footnote1)</sup> | Returns total count.                                                                                                                                                                                                                        |

### Supported Operators and Query Functions

The Cognigy OData endpoint supports a limited set of [operators](https://docs.oasis-open.org/odata/odata/v4.01/odata-v4.01-part1-protocol.html#sec_BuiltinFilterOperations) and [query functions](https://docs.oasis-open.org/odata/odata/v4.01/odata-v4.01-part1-protocol.html#sec_BuiltinQueryFunctions) that you can use in the `$filter` system query option to create more complex queries.

<AccordionGroup>
  <Accordion title="Comparison Operators">
    These operators allow you to compare field values against specific criteria to filter your data effectively.

    | Operator | Meaning          | Example                             |
    | -------- | ---------------- | ----------------------------------- |
    | `eq`     | Equal            | `userId eq '12345'`                 |
    | `ne`     | Not equal        | `status ne 'active'`                |
    | `gt`     | Greater than     | `timestamp gt 2024-01-01T00:00:00Z` |
    | `ge`     | Greater or equal | `value ge 10`                       |
    | `lt`     | Less than        | `value lt 100`                      |
    | `le`     | Less or equal    | `value le 100`                      |
  </Accordion>

  <Accordion title="Logical Operators">
    These operators allow you to combine multiple conditions for creating complex queries.

    <Warning>
      The `not` logical operator isn't supported in the Cognigy OData endpoint. Expressions such as `not startswith(field, 'value')`, `not contains(field, 'value')`, or `startswith(field, 'value') eq false` won't work.
    </Warning>

    | Operator | Meaning     | Example                                     |
    | -------- | ----------- | ------------------------------------------- |
    | `and`    | Logical AND | `status eq 'active' and channel eq 'web'`   |
    | `or`     | Logical OR  | `status eq 'active' or status eq 'pending'` |
  </Accordion>

  <Accordion title="Query Functions">
    These functions allow you to perform complex filtering based on string patterns.

    <Warning>
      * Query functions may affect the performance of your requests. Use them with caution, especially on large datasets.
      * Query functions don't work in fields of type `text[]` (array of text values). Use exact match filtering (`eq`) or filter in a different supported field.
    </Warning>

    | Function     | Purpose                             | Example                      |
    | ------------ | ----------------------------------- | ---------------------------- |
    | `contains`   | Checks if string contains value.    | `contains(userId,'admin')`   |
    | `startswith` | Checks if string starts with value. | `startswith(userId,'user_')` |
    | `endswith`   | Checks if string ends with value.   | `endswith(email,'.com')`     |
  </Accordion>
</AccordionGroup>

### Examples

The tables contain examples of requests using system query options and operators.
Select the OData endpoint version [compatible with your Cognigy version](#supported-versions).

<Tabs>
  <Tab title="OData 2.4">
    | Request Description                                                                                                                                                                                   | OData 2.4 Request                                                                                                                                                  |
    | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Returns total count of Analytics records (not supported in Microsoft Excel or Power BI).                                                                                                              | `https://odata-trial.cognigy.ai/v2.4/Analytics/$count?apikey=<your-api-key>`                                                                                       |
    | Returns all Analytics records for the given API Key.                                                                                                                                                  | `https://odata-trial.cognigy.ai/v2.4/Analytics?apikey=<your-api-key>`                                                                                              |
    | Returns the first 10 Analytics records.                                                                                                                                                               | `https://odata-trial.cognigy.ai/v2.4/Analytics/?$top=10&apikey=<your-api-key>`                                                                                     |
    | Returns the top 5 Analytics records where `executionTime` is less than 50ms, ordered by `executionTime`.                                                                                              | `https://odata-trial.cognigy.ai/v2.4/Analytics/?$filter=executionTime lt 50&$top=5&$orderby=executionTime&apikey=<your-api-key>`                                   |
    | Returns all Analytics records for a specific Cognigy Project filtered by `projectId`.                                                                                                                 | `https://odata-trial.cognigy.ai/v2.4/Analytics/?$filter=projectId eq '<project-id>'&apikey=<your-api-key>`                                                         |
    | Returns Project and session IDs for a specific Cognigy project. The request filters the data based on the `projectId` field and selects only the `projectId` and `sessionId` for each matching entry. | `https://odata-trial.cognigy.ai/v2.4/Analytics/?$filter=projectId eq '<project-id>'&$select=projectId,sessionId&apikey=<your-api-key>`                             |
    | Returns all Analytics records between two dates, for example, 1st Jan 2021 and 1st July 2021.                                                                                                         | `https://odata-trial.cognigy.ai/v2.4/Analytics/?$filter=timestamp gt '2021-01-01T00:00:00.000Z' and timestamp lt '2021-07-01T00:00:00.000Z'&apikey=<your-api-key>` |
    | Returns Analytics records where a field contains a specific string value.                                                                                                                             | `https://odata-trial.cognigy.ai/v2.4/Analytics/?$filter=contains(userId,'admin')&apikey=<your-api-key>`                                                            |
    | Returns Analytics records where a field starts with a specific string value.                                                                                                                          | `https://odata-trial.cognigy.ai/v2.4/Analytics/?$filter=startswith(userId,'user_')&apikey=<your-api-key>`                                                          |
    | Returns Analytics records where a field ends with a specific string value.                                                                                                                            | `https://odata-trial.cognigy.ai/v2.4/Analytics/?$filter=endswith(userId,'_admin')&apikey=<your-api-key>`                                                           |
  </Tab>

  <Tab title="OData 2.3">
    | Request Description                                                                                                                                                                                       | OData 2.3 Request                                                                                                                                                  |
    | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Returns total count of Analytics records (not supported in Microsoft Excel or Power BI).                                                                                                                  | `https://odata-trial.cognigy.ai/v2.3/Analytics/$count?apikey=<your-api-key>`                                                                                       |
    | Returns all Analytics records for the given API Key.                                                                                                                                                      | `https://odata-trial.cognigy.ai/v2.3/Analytics?apikey=<your-api-key>`                                                                                              |
    | Returns the first 10 Analytics records.                                                                                                                                                                   | `https://odata-trial.cognigy.ai/v2.3/Analytics/?$top=10&apikey=<your-api-key>`                                                                                     |
    | Returns the top 5 Analytics records where `executionTime` is less than 50ms, ordered by `executionTime`.                                                                                                  | `https://odata-trial.cognigy.ai/v2.3/Analytics/?$filter=executionTime lt 50&$top=5&$orderby=executionTime&apikey=<your-api-key>`                                   |
    | Returns all Analytics records for a specific Cognigy Project filtered by `projectId`                                                                                                                      | `https://odata-trial.cognigy.ai/v2.3/Analytics/?$filter=projectId eq '<project-id>'&apikey=<your-api-key>`                                                         |
    | Returns Project IDs and session IDs for a specific Cognigy project. The request filters the data based on the `projectId` field and selects only the `projectId` and `sessionId` for each matching entry. | `https://odata-trial.cognigy.ai/v2.3/Analytics/?$filter=projectId eq '<project-id>'&$select=projectId,sessionId&apikey=<your-api-key>`                             |
    | Returns all Analytics records between two dates, for example, 1st Jan 2021 and 1st July 2021.                                                                                                             | `https://odata-trial.cognigy.ai/v2.3/Analytics/?$filter=timestamp gt '2021-01-01T00:00:00.000Z' and timestamp lt '2021-07-01T00:00:00.000Z'&apikey=<your-api-key>` |
  </Tab>
</Tabs>

For a full reference,
refer to the extensive collection of resources at [OData.org](https://www.odata.org/) and the [Oasis OData URL Convention](http://docs.oasis-open.org/odata/odata/v4.0/odata-v4.0-part2-url-conventions.html) documentation.

## Cognigy OData Collections

The OData endpoint provides access to several data *collections* containing information about your interactions and workflows. A *collection* is similar to a database table or a set of entities containing data.
The set of collections varies depending on which version you are using:

<Tabs>
  <Tab title="Version 2.4">
    The current version of the OData endpoint is `v2.4`. In this version, the following OData collections are available:

    * [Analytics](#analytics) (`/Analytics`)
    * [Conversations](#conversations) (`/Conversations`)
    * [Steps](#steps) (`/Steps`)
    * [ExecutedSteps](#executedsteps) (`/ExecutedSteps`)
    * [Sessions](#sessions) (`/Sessions`)
    * [LiveAgentEscalations](#liveagentescalations) (`/LiveAgentEscalations`)
    * [Goals](#goals) (`/Goals`)
    * [GoalSteps](#goalsteps) (`/GoalSteps`)
    * [GoalStepMetrics](#goalstepmetrics) (`/GoalStepMetrics`)
    * [GoalEvents](#goalevents) (`/GoalEvents`)

    The URL for accessing the OData endpoint 2.4 is as follows:
    `https://<odata-domain>/v2.4/<collection>?apikey=<your-api-key>`
  </Tab>

  <Tab title="Version 2.3">
    The previous version of the OData endpoint is `2.3`. In this version, the following OData collections are available:

    * [Analytics](#analytics) (`/Analytics`)
    * [Conversations](#conversations) (`/Conversations`)
    * [Steps](#steps) (`/Steps`)
    * [ExecutedSteps](#executedsteps) (`/ExecutedSteps`)
    * [Sessions](#sessions) (`/Sessions`)
    * [LiveAgentEscalations](#liveagentescalations) (`/LiveAgentEscalations`)

    The URL for accessing the OData endpoint 2.3 is as follows:
    `https://<odata-domain>/v2.3/<collection>?apikey=<your-api-key>`
  </Tab>
</Tabs>

Each collection can be filtered by its fields, which are often retrieved directly from the [Input](/ai/agents/develop/ai-agent-memory/input) object results.

### Analytics

The Analytics collection offers a log of comprehensive details such as organizational information, Project identifiers, Flow specifics, and essential performance metrics. Each chat or voice AI Agent session in Cognigy creates an Analytics record for an interaction. Each interaction is logged as a single record. The information is saved only after the end of the session. You can change this data using the [Overwrite Analytics Node](/ai/agents/develop/node-reference/analytics/overwrite-analytics).

**Requests**

Select an API request to extract data from the collection. The API version in the request must match [your Cognigy version](#supported-versions).

<Tabs>
  <Tab title="OData 2.4">
    `https://odata-trial.cognigy.ai/v2.4/Analytics?apikey=<your-api-key>`
  </Tab>

  <Tab title="OData 2.3">
    `https://odata-trial.cognigy.ai/v2.3/Analytics?apikey=<your-api-key>`
  </Tab>
</Tabs>

**Response**

When retrieving this collection, the endpoint returns the following fields:

<Accordion title="Response Fields">
  | Field Name         | <div>Description</div>                                                                                                                                                                                                                                                                                                                                                          | Type               | Example                                                          |
  | ------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------ | ---------------------------------------------------------------- |
  | id                 | The unique ID of the record. Applicable to the OData endpoint 2.4 or later.                                                                                                                                                                                                                                                                                                     | String             | 5a91d194fde28b0011ce2423                                         |
  | \_id               | The unique ID of the record. Applicable to the OData endpoint 2.3 or earlier.                                                                                                                                                                                                                                                                                                   | String             | 5a91d194fde28b0011ce2423                                         |
  | organisation       | The name of your organization.                                                                                                                                                                                                                                                                                                                                                  | String             | cognigy                                                          |
  | projectId          | The ID of the Project related to the record.                                                                                                                                                                                                                                                                                                                                    | String             | 5a91d194fde28b0011ce2422                                         |
  | flowReferenceId    | The ID of the Flow.                                                                                                                                                                                                                                                                                                                                                             | String             | 66ceb9cc-ac5b-4cdd-9f4b-8a48dfbb35d9                             |
  | entrypoint         | The ID of the Project or Snapshot used by the Endpoint.                                                                                                                                                                                                                                                                                                                         | String             | 621d3deb87016d3855ec0c92                                         |
  | ip                 | The IP address from which the request originated.                                                                                                                                                                                                                                                                                                                               | String             | 78.143.45.111                                                    |
  | contactId          | The hash of the user ID.                                                                                                                                                                                                                                                                                                                                                        | String             | 99ea95befe61812656bf963ba7d44ab7                                 |
  | sessionId          | The session ID.                                                                                                                                                                                                                                                                                                                                                                 | String             | 5a91d194fde28b0011ce2425                                         |
  | inputId            | The ID of the input message.                                                                                                                                                                                                                                                                                                                                                    | String             | 5a91d194fde28b0011ce2424                                         |
  | inputText          | The input text.                                                                                                                                                                                                                                                                                                                                                                 | String             | Hello World!                                                     |
  | inputData          | The input data object as a string.                                                                                                                                                                                                                                                                                                                                              | String             | `{"key":"value"}`                                                |
  | state              | The state of the Flow at input.                                                                                                                                                                                                                                                                                                                                                 | String             | default                                                          |
  | mode               | The [mode](/ai/agents/develop/ai-agent-memory/input#properties) of the input.                                                                                                                                                                                                                                                                                                   | String             | TextOnly                                                         |
  | userType           | The type of the connecting user. Either `external` for external user or `admin` for admin user.                                                                                                                                                                                                                                                                                 | String             | external                                                         |
  | channel            | The channel through which the record was received.                                                                                                                                                                                                                                                                                                                              | String             | facebook                                                         |
  | flowLanguage       | The language of the Flow.                                                                                                                                                                                                                                                                                                                                                       | String             | en-EN                                                            |
  | intent             | The found Intent. This field can be empty.                                                                                                                                                                                                                                                                                                                                      | String             | orderFood                                                        |
  | intentFlow         | The Parent ID of the Flow in which the Intent was found. This field can be empty.                                                                                                                                                                                                                                                                                               | String             | 5e33b160e6236da3aa54221461a53f04                                 |
  | intentScore        | The numerical score of the input against a recognized Intent. This field can be empty.                                                                                                                                                                                                                                                                                          | Number             | 0.836322430892043                                                |
  | completedGoalsList | A list of completed [tasks](/ai/agents/analyze/goals-and-tasks/tasks) in this session. If you have the [goals](/ai/agents/analyze/goals-and-tasks/tasks) feature enabled, use the [Goals collection](/ai/agents/analyze/odata#goals) through the OData 2.4 endpoint.                                                                                                            | String             | orderedFood                                                      |
  | foundSlots         | Found Slot tags.                                                                                                                                                                                                                                                                                                                                                                | String             | DATE                                                             |
  | foundSlotDetails   | Found Slot tags with details.                                                                                                                                                                                                                                                                                                                                                   | String             | DATE\[2018-2-25T12:32:32.000]                                    |
  | timestamp          | The date and time of the input.                                                                                                                                                                                                                                                                                                                                                 | DateTime           | 2018-2-25T12:32:32.000Z                                          |
  | executionTime      | The time it took to execute the Flow in milliseconds.                                                                                                                                                                                                                                                                                                                           | Number             | 32                                                               |
  | execution          | The Flow execution count.                                                                                                                                                                                                                                                                                                                                                       | Number             | 3                                                                |
  | custom1            | The custom value created by the Flow. The custom field has a maximum character limit of 1024.                                                                                                                                                                                                                                                                                   | String             | custom value                                                     |
  | custom2            | The custom value created by the Flow. The custom field has a maximum character limit of 1024.                                                                                                                                                                                                                                                                                   | String             | custom value                                                     |
  | custom3            | The custom value created by the Flow. The custom field has a maximum character limit of 1024.                                                                                                                                                                                                                                                                                   | String             | custom value                                                     |
  | custom4            | The custom value created by the Flow. The custom field has a maximum character limit of 1024.                                                                                                                                                                                                                                                                                   | String             | custom value                                                     |
  | custom5            | The custom value created by the Flow. The custom field has a maximum character limit of 1024.                                                                                                                                                                                                                                                                                   | String             | custom value                                                     |
  | custom6            | The custom value created by the Flow. The custom field has a maximum character limit of 1024.                                                                                                                                                                                                                                                                                   | String             | custom value                                                     |
  | custom7            | The custom value created by the Flow. The custom field has a maximum character limit of 1024.                                                                                                                                                                                                                                                                                   | String             | custom value                                                     |
  | custom8            | The custom value created by the Flow. The custom field has a maximum character limit of 1024.                                                                                                                                                                                                                                                                                   | String             | custom value                                                     |
  | custom9            | The custom value created by the Flow. The custom field has a maximum character limit of 1024.                                                                                                                                                                                                                                                                                   | String             | custom value                                                     |
  | custom10           | The custom value created by the Flow. The custom field has a maximum character limit of 1024.                                                                                                                                                                                                                                                                                   | String             | custom value                                                     |
  | localeReferenceId  | The reference ID for the locale.                                                                                                                                                                                                                                                                                                                                                | String             | 4bfdcf40-a605-4686-8e92-b17fa9f178af                             |
  | localeName         | The name of the locale.                                                                                                                                                                                                                                                                                                                                                         | String             | English                                                          |
  | endpointUrlToken   | The URL token for the Endpoint.                                                                                                                                                                                                                                                                                                                                                 | String             | ea30b8f20db52f9d86ea36fd55a7d66bd2c4c60eb24ac7ad52f1c9e173dd4cdb |
  | endpointName       | The name of the Endpoint.                                                                                                                                                                                                                                                                                                                                                       | String             | MyEndpoint                                                       |
  | rating             | The rating for the session: -1 for a negative rating or 1 for a positive rating.                                                                                                                                                                                                                                                                                                | Number             | 1                                                                |
  | ratingComment      | The comment left when given a rating.                                                                                                                                                                                                                                                                                                                                           | String             | Great service!                                                   |
  | snapshotName       | The name for the Snapshot used by the Endpoint.                                                                                                                                                                                                                                                                                                                                 | String             | Bot Release 2.2                                                  |
  | isMasked           | This field is available only for OData 2.4. A flag indicating whether the record has personally identifiable information (PII) has been masked using mask and keep analytics options from either the [Deactivate Profile](/ai/agents/develop/node-reference/analytics/deactivate-profile) or [Delete Profile](/ai/agents/develop/node-reference/analytics/delete-profile) Node. | Boolean (nullable) | `true`                                                           |

  <Note>
    The Analytics step label can be updated from the Cognigy UI at any time.
    Therefore, to ensure compatibility between past and future data records, the `entityReferenceId` field is used.
    This field corresponds to either the Node ID or Intent ID assigned as an Analytics step.
    The current Analytics step label can be retrieved
    by mapping the `entityReferenceID` field to the [`Steps`](#steps) record.
  </Note>
</Accordion>

### Conversations

The Conversations collection offers a log of all session messages, including messages from the end user, AI Agent, and human agent.
Each time one of these sources sends a message to a Flow, Cognigy creates a record to log the interaction.
Each interaction is logged as a single record.

**Requests**

Select an API request to extract data from the collection. The API version in the request must match [your Cognigy version](#supported-versions).

<Tabs>
  <Tab title="OData 2.4">
    `https://odata-trial.cognigy.ai/v2.4/Conversations?apikey=<your-api-key>`
  </Tab>

  <Tab title="OData 2.3">
    `https://odata-trial.cognigy.ai/v2.3/Conversations?apikey=<your-api-key>`
  </Tab>
</Tabs>

**Response**

When retrieving this collection, the endpoint returns the following fields:

<Accordion title="Response Fields">
  | Field Name             | <div>Description</div>                                                                                                                                                                                                                                                                                                                                                          | Type               | Example                                                                                                                |
  | ---------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------ | ---------------------------------------------------------------------------------------------------------------------- |
  | id                     | The unique ID of the record. Applicable to the OData endpoint 2.4 or later.                                                                                                                                                                                                                                                                                                     | String             | 5a91d194fde28b0011ce2423                                                                                               |
  | \_id                   | The unique ID of the record. Applicable to the OData endpoint 2.3 or earlier.                                                                                                                                                                                                                                                                                                   | String             | 5a91d194fde28b0011ce2423                                                                                               |
  | projectId              | The ID of the Project related to the record.                                                                                                                                                                                                                                                                                                                                    | String             | 5a91d194fde28b0011ce2422                                                                                               |
  | projectName            | The name of the Project.                                                                                                                                                                                                                                                                                                                                                        | String             | Project 1                                                                                                              |
  | inputId                | The ID of the input message.                                                                                                                                                                                                                                                                                                                                                    | String             | 5a91d194fde28b0011ce2424                                                                                               |
  | sessionId              | The session ID.                                                                                                                                                                                                                                                                                                                                                                 | String             | 5a91d194fde28b0011ce2425                                                                                               |
  | contactId              | The ID of the connecting user.                                                                                                                                                                                                                                                                                                                                                  | String             | myContactID                                                                                                            |
  | organisation           | The name of your organization.                                                                                                                                                                                                                                                                                                                                                  | String             | cognigy                                                                                                                |
  | inputText              | The input text.                                                                                                                                                                                                                                                                                                                                                                 | String             | Hello World!                                                                                                           |
  | inputData              | The input data object as a string.                                                                                                                                                                                                                                                                                                                                              | String             | `{"key":"value"}`                                                                                                      |
  | type                   | Whether the message is an input or output of the Flow.                                                                                                                                                                                                                                                                                                                          | String             | "input" or "output"                                                                                                    |
  | source                 | The source of the message.                                                                                                                                                                                                                                                                                                                                                      | String             | "user" or "bot" or "agent" or "suggestion"                                                                             |
  | timestamp              | The date and time of the input.                                                                                                                                                                                                                                                                                                                                                 | DateTime           | 2018-2-25T12:32:32.000Z                                                                                                |
  | flowName               | The name of the Flow.                                                                                                                                                                                                                                                                                                                                                           | String             | MainFlow                                                                                                               |
  | flowParentId           | The Parent ID of the Flow.                                                                                                                                                                                                                                                                                                                                                      | String             | 5e33b160e6236da3aa54221461a53f04                                                                                       |
  | channel                | The channel through which the record was received.                                                                                                                                                                                                                                                                                                                              | String             | facebook                                                                                                               |
  | inHandoverRequest      | A flag indicating whether the session is in a handover request.                                                                                                                                                                                                                                                                                                                 | Boolean            | false                                                                                                                  |
  | inHandoverConversation | A flag indicating whether the session is in a handover session.                                                                                                                                                                                                                                                                                                                 | Boolean            | true                                                                                                                   |
  | outputId               | The output ID.                                                                                                                                                                                                                                                                                                                                                                  | String             | f514b7b2-7dc0-4e75-be62-a53fed5b2bb7                                                                                   |
  | inputAttachments       | The list of attachments that the end user uploads to the chat. Each attachment will contain the URL of the file, file name, and file type. <br /><br /> When a human agent uploads a file, it will be recorded in the OData format as follows, without any specific information about the file: `[{"*****":"*****"}]`.                                                          | Object             | `{"url":"https://files-api-trial.cognigy.ai/v1.0/63babf6e92add771923e87b7/","name":"abandon-rate.png","type":"image"}` |
  | reference              | The Reference ID. For example, to reference a suggestion.                                                                                                                                                                                                                                                                                                                       | String             | null                                                                                                                   |
  | localeReferenceId      | The ID of the locale.                                                                                                                                                                                                                                                                                                                                                           | String             | 7eff993c-b801-4556-b111-1c319e8577cf                                                                                   |
  | endpointUrlToken       | The URL token for the Endpoint.                                                                                                                                                                                                                                                                                                                                                 | String             | ea30b8f20db52f9d86ea36fd55a7d66bd2c4c60eb24ac7ad52f1c9e173dd4cdb                                                       |
  | endpointName           | The name of the Endpoint.                                                                                                                                                                                                                                                                                                                                                       | String             | MyEndpoint                                                                                                             |
  | snapshotId             | The ID of the Snapshot.                                                                                                                                                                                                                                                                                                                                                         | String             | 63ff0cc47a466cab278fd19b                                                                                               |
  | snapshotName           | The name of the Snapshot.                                                                                                                                                                                                                                                                                                                                                       | String             | Bot Release 2.2                                                                                                        |
  | rating                 | The rating for the session: -1 for a negative rating or 1 for a positive rating.                                                                                                                                                                                                                                                                                                | Number             | 1                                                                                                                      |
  | ratingComment          | The comment left when given a rating.                                                                                                                                                                                                                                                                                                                                           | String             | Great service!                                                                                                         |
  | isMasked               | This field is available only for OData 2.4. A flag indicating whether the record has personally identifiable information (PII) has been masked using mask and keep analytics options from either the [Deactivate Profile](/ai/agents/develop/node-reference/analytics/deactivate-profile) or [Delete Profile](/ai/agents/develop/node-reference/analytics/delete-profile) Node. | Boolean (nullable) | `true`                                                                                                                 |
</Accordion>

### Steps

The Steps collection provides a list of all entities,
which are either Nodes or Intents, assigned as [Analytics Steps](/ai/agents/analyze/analytics-steps) within any Flow.
An Analytics step appear in this OData collection only after it has been triggered at least once in a session within the Flow.
Each step is logged as a single record.

**Requests**

Select an API request to extract data from the collection. The API version in the request must match [your Cognigy version](#supported-versions).

<Tabs>
  <Tab title="OData 2.4">
    `https://odata-trial.cognigy.ai/v2.4/Steps?apikey=<your-api-key>`
  </Tab>

  <Tab title="OData 2.3">
    `https://odata-trial.cognigy.ai/v2.3/Steps?apikey=<your-api-key>`
  </Tab>
</Tabs>

**Response**

When retrieving this collection, the endpoint returns the following fields:

<Accordion title="Response Fields">
  | Field Name        | <div>Description</div>                                                        | Type   | Example                          |
  | ----------------- | ----------------------------------------------------------------------------- | ------ | -------------------------------- |
  | id                | The unique ID of the record. Applicable to the OData endpoint 2.4 or later.   | String | 5a91d194fde28b0011ce2423         |
  | \_id              | The unique ID of the record. Applicable to the OData endpoint 2.3 or earlier. | String | 5a91d194fde28b0011ce2423         |
  | label             | The Analytics step label defined for the entity (Node or Intent) in Cognigy.  | String | Question (2)                     |
  | type              | The type of entity.                                                           | String | `node` or `intent`               |
  | entityReferenceId | The unique ID for the entity (Node ID or Flow ID).                            | String | 5a91d194fde28b0011ce2423         |
  | flowReferenceId   | The ID of the Flow.                                                           | String | 5a91d194fde28b0011ce2423         |
  | flowName          | The name of the Flow where the Analytics step is configured.                  | String | Main Flow                        |
  | projectName       | The name of the Project.                                                      | String | Project 1                        |
  | snapshotId        | The ID of the Snapshot.                                                       | String | 5e33b160e6236da3aa54221461a53f04 |
  | snapshotName      | The name of the Snapshot.                                                     | String | Bot Release 2.2                  |
</Accordion>

<Check>
  The [Step Monitoring with OData](https://support.cognigy.com/hc/en-us/articles/360021089180-Step-Monitoring-with-OData) article explains how these records can be manipulated in BI software to build insightful session path analytics.
</Check>

### ExecutedSteps

The ExecutedSteps collection logs all step events in sessions, including a reference to the prior (parent) step.
Whenever a Node or Intent with an assigned step is executed, the collection creates a record for that step.
Each step is logged as a single record.

<Tabs>
  <Tab title="OData 2.4">
    `https://odata-trial.cognigy.ai/v2.4/ExecutedSteps?apikey=<your-api-key>`
  </Tab>

  <Tab title="OData 2.3">
    `https://odata-trial.cognigy.ai/v2.3/ExecutedSteps?apikey=<your-api-key>`
  </Tab>
</Tabs>

**Response**

When retrieving this collection, the endpoint returns the following fields:

<Accordion title="Response Fields">
  | Field Name        | <div>Description</div>                                                        | Type     | Example                                                          |
  | ----------------- | ----------------------------------------------------------------------------- | -------- | ---------------------------------------------------------------- |
  | id                | The unique ID of the record. Applicable to the OData endpoint 2.4 or later.   | String   | 5a91d194fde28b0011ce2423                                         |
  | \_id              | The unique ID of the record. Applicable to the OData endpoint 2.3 or earlier. | String   | 5a91d194fde28b0011ce2423                                         |
  | userId            | The ID of the connecting user.                                                | String   | myContactID                                                      |
  | sessionId         | The session ID.                                                               | String   | 5a91d194fde28b0011ce2425                                         |
  | inputId           | The ID of the input message.                                                  | String   | 5a91d194fde28b0011ce2425                                         |
  | stepLabel         | The Analytics step label defined for the entity (Node or Intent)              | String   | Question (2)                                                     |
  | parentStep        | The ID of the step that occurred prior to this step.                          | String   | 5a91d194fde28b0011ce2425                                         |
  | type              | The type of entity.                                                           | String   | `node` or `intent`                                               |
  | entityReferenceId | The unique ID for the entity (Node ID or Intent ID).                          | String   | 5a91d194fde28b0011ce2423                                         |
  | flowReferenceId   | The ID of the Flow.                                                           | String   | 5a91d194fde28b0011ce2423                                         |
  | flowName          | The name of the Flow where the step exists.                                   | String   | Main Flow                                                        |
  | timestamp         | The date and time when the step was executed.                                 | DateTime | 2018-2-25T12:32:32.000Z                                          |
  | projectName       | The name of the Project.                                                      | String   | Project 1                                                        |
  | projectId         | The ID of the Project related to the record.                                  | String   | 6067352c18887e471da4e392                                         |
  | organisationId    | The ID of the organization associated with the record.                        | String   | 5f8833dae72b850ad2ed4d53                                         |
  | snapshotId        | The ID of the Snapshot.                                                       | String   | 5e33b160e6236da3aa54221461a53f04                                 |
  | snapshotName      | The name of the Snapshot.                                                     | String   | Bot Release 2.2                                                  |
  | localeReferenceId | The ID of the locale.                                                         | String   | 7eff993c-b801-4556-b111-1c319e8577cf                             |
  | localeName        | The name of the locale.                                                       | String   | English                                                          |
  | endpointUrlToken  | The URL token for the Endpoint.                                               | String   | ea30b8f20db52f9d86ea36fd55a7d66bd2c4c60eb24ac7ad52f1c9e173dd4cdb |
  | endpointName      | The name of the Endpoint.                                                     | String   | MyEndpoint                                                       |
</Accordion>

### Sessions

The Sessions collection contains a list of all sessions that have occurred.
The primary objective of this collection is
to provide a list of the Analytics steps that took place in any given session, along with the order of execution.
This information is included as a comma-separated list within a field called `stepPath`.
Each session is logged as a single record.

<Tabs>
  <Tab title="OData 2.4">
    `https://odata-trial.cognigy.ai/v2.4/Sessions?apikey=<your-api-key>`
  </Tab>

  <Tab title="OData 2.3">
    `https://odata-trial.cognigy.ai/v2.3/Sessions?apikey=<your-api-key>`
  </Tab>
</Tabs>

**Response**

When retrieving this collection, the endpoint returns the following fields:

<Accordion title="Response Fields">
  | Field Name          | <div>Description</div>                                                                                                                                                                                                                                                                                                                                                          | Type               | Example                                                                        |
  | ------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------ | ------------------------------------------------------------------------------ |
  | id                  | The unique ID of the record. Applicable to the OData endpoint 2.4 or later.                                                                                                                                                                                                                                                                                                     | String             | 5a91d194fde28b0011ce2423                                                       |
  | \_id                | The unique ID of the record. Applicable to the OData endpoint 2.3 or earlier.                                                                                                                                                                                                                                                                                                   | String             | 5a91d194fde28b0011ce2423                                                       |
  | goals               | All the goals that were achieved in the session.                                                                                                                                                                                                                                                                                                                                | String             | Goal1, Goal2                                                                   |
  | stepPath            | A comma-separated list of executed steps.                                                                                                                                                                                                                                                                                                                                       | String             | 9ac4f679-beae-4461-b9e3-43aece8b3430,f1e72fe3-f04b-48f5-b862-1e35ad253f18, ... |
  | stepsCount          | The count of the number of steps.                                                                                                                                                                                                                                                                                                                                               | Number             | 10                                                                             |
  | handoverEscalations | The number of times the session triggered a handover.                                                                                                                                                                                                                                                                                                                           | Number             | 3                                                                              |
  | startedAt           | The date and time when first message was received.                                                                                                                                                                                                                                                                                                                              | DateTime           | 2018-2-25T12:32:32.000Z                                                        |
  | userId              | The ID of the connecting user.                                                                                                                                                                                                                                                                                                                                                  | String             | myContactID                                                                    |
  | sessionId           | The session ID.                                                                                                                                                                                                                                                                                                                                                                 | String             | 5a91d194fde28b0011ce2425                                                       |
  | localeReferenceId   | The ID of the locale.                                                                                                                                                                                                                                                                                                                                                           | String             | 7eff993c-b801-4556-b111-1c319e8577cf                                           |
  | localeName          | The name of the locale.                                                                                                                                                                                                                                                                                                                                                         | String             | English                                                                        |
  | endpointReferenceId | The ID of the Endpoint.                                                                                                                                                                                                                                                                                                                                                         | String             | 5e33b160e6236da3aa54221461a53f04                                               |
  | endpointName        | The name of the Endpoint.                                                                                                                                                                                                                                                                                                                                                       | String             | Webchat                                                                        |
  | projectName         | The name of the Project.                                                                                                                                                                                                                                                                                                                                                        | String             | Project 1                                                                      |
  | projectId           | The ID of the Project related to the record.                                                                                                                                                                                                                                                                                                                                    | String             | 6067352c18887e471da4e392                                                       |
  | organisationId      | The ID of the organization associated with the record.                                                                                                                                                                                                                                                                                                                          | String             | 5f8833dae72b850ad2ed4d53                                                       |
  | snapshotId          | The ID of the Snapshot.                                                                                                                                                                                                                                                                                                                                                         | String             | 5e33b160e6236da3aa54221461a53f04                                               |
  | snapshotName        | The name of the Snapshot.                                                                                                                                                                                                                                                                                                                                                       | String             | Bot Release 2.2                                                                |
  | endpointUrlToken    | The URL token for the Endpoint.                                                                                                                                                                                                                                                                                                                                                 | String             | ea30b8f20db52f9d86ea36fd55a7d66bd2c4c60eb24ac7ad52f1c9e173dd4cdb               |
  | rating              | The rating for the session: -1 for a negative rating or 1 for a positive rating.                                                                                                                                                                                                                                                                                                | Number             | 1                                                                              |
  | ratingComment       | The comment left when given a rating.                                                                                                                                                                                                                                                                                                                                           | String             | Great service!                                                                 |
  | isMasked            | This field is available only for OData 2.4. A flag indicating whether the record has personally identifiable information (PII) has been masked using mask and keep analytics options from either the [Deactivate Profile](/ai/agents/develop/node-reference/analytics/deactivate-profile) or [Delete Profile](/ai/agents/develop/node-reference/analytics/delete-profile) Node. | Boolean (nullable) | `true`                                                                         |
</Accordion>

### LiveAgentEscalations

The LiveAgentEscalations collection contains records from [Cognigy Live Agent](/live-agent/overview) and includes the related data about handovers to Live Agent.
Each handover (escalation) is logged as a single record.

<Tabs>
  <Tab title="OData 2.4">
    `https://odata-trial.cognigy.ai/v2.4/LiveAgentEscalations?apikey=<your-api-key>`
  </Tab>

  <Tab title="OData 2.3">
    `https://odata-trial.cognigy.ai/v2.3/LiveAgentEscalations?apikey=<your-api-key>`
  </Tab>
</Tabs>

**Response**

When retrieving this collection, the endpoint returns the following fields:

<Accordion title="Response Fields">
  | Field Name        | <div>Description</div>                                                                                                                        | Type     | Example                                                          |
  | ----------------- | --------------------------------------------------------------------------------------------------------------------------------------------- | -------- | ---------------------------------------------------------------- |
  | id                | The unique ID of the record. Applicable to the OData endpoint 2.4 or later.                                                                   | String   | 5a91d194fde28b0011ce2423                                         |
  | \_id              | The unique ID of the record. Applicable to the OData endpoint 2.3 or earlier.                                                                 | String   | 5a91d194fde28b0011ce2423                                         |
  | organisationId    | The ID of the organization associated with the record.                                                                                        | String   | 5eb4a49d426cd3d05f2892a9                                         |
  | projectId         | The ID of the Project related to the record.                                                                                                  | String   | 6409b6c53c9687b5e78e6403                                         |
  | sessionId         | The session ID.                                                                                                                               | String   | session-2a38246e-0395-49a5-876d-dd5133fd5418                     |
  | timestamp         | The date and time when the record was created.                                                                                                | DateTime | 2023-06-28T07:59:04.628Z                                         |
  | localeName        | The name of the locale.                                                                                                                       | String   | en                                                               |
  | status            | The current status of the escalated conversation in Live Agent. The possible statures are: `opened`, `assigned`, `resolved`, and `abandoned`. | String   | opened                                                           |
  | inboxId           | The ID of the inbox associated with the record.                                                                                               | String   | 7983                                                             |
  | inboxName         | The name of the inbox.                                                                                                                        | String   | LA Chatbot-default                                               |
  | teamId            | The ID of the team assigned to the conversation.                                                                                              | String   | 8                                                                |
  | teamName          | The name of the team assigned to the conversation.                                                                                            | String   | IT Support                                                       |
  | labels            | The list of labels assigned to the conversation.                                                                                              | Array    | \[Hardware]                                                      |
  | agentId           | The ID of the human agent assigned to the conversation.                                                                                       | String   | 618cf7687614774ba028dcac                                         |
  | agentName         | The name of the human agent assigned to the conversation. This field is supported from Cognigy 4.80.                                          | String   | John Doe                                                         |
  | contactId         | The ID of the connecting user.                                                                                                                | String   | 649ba8450d5df5cadf8b75b8                                         |
  | endpointName      | The name of the Endpoint.                                                                                                                     | String   | LA Chatbot                                                       |
  | endpointType      | The type of the Endpoint.                                                                                                                     | String   | webchat2                                                         |
  | endpointUrlToken  | The URL token for the Endpoint.                                                                                                               | String   | ea30b8f20db52f9d86ea36fd55a7d66bd2c4c60eb24ac7ad52f1c9e173dd4cdb |
  | channel           | The channel through which the record was received.                                                                                            | String   | admin-webchat                                                    |
  | localeReferenceId | The ID of the locale.                                                                                                                         | String   | 7eff993c-b801-4556-b111-1c319e8577cf                             |
  | snapshotId        | The ID of the Snapshot.                                                                                                                       | String   | 63ff0cc47a466cab278fd19b                                         |
  | snapshotName      | The name of the Snapshot.                                                                                                                     | String   | Prod-Snapshot 26.7.23                                            |
</Accordion>

### Goals

The Goals collection contains records from [goals](/ai/agents/analyze/goals-and-tasks/overview).
Each goal is logged as a single record.

<Tabs>
  <Tab title="OData 2.4">
    `https://odata-trial.cognigy.ai/v2.4/Goals?apikey=<your-api-key>`
  </Tab>
</Tabs>

**Response**

When retrieving this collection, the endpoint returns the following fields:

<Accordion title="Response Fields">
  | Field Name     | <div>Description</div>                                                                                                                      | Field Type | Example                                                      |
  | -------------- | ------------------------------------------------------------------------------------------------------------------------------------------- | ---------- | ------------------------------------------------------------ |
  | goalId         | The goal ID.                                                                                                                                | String     | 66ebf71693dfb26c3ca4c827                                     |
  | name           | The name of the goal.                                                                                                                       | String     | Goal\_07                                                     |
  | version        | The UUID for tracking the current version of the goal configuration. Each update or change of the goal configuration creates a new version. | String     | 8ec4de84-71e2-4c98-a7c3-b18173f36223                         |
  | description    | The description of the goal.                                                                                                                | String     | The goal is to gather metrics for a flight booking use case. |
  | referenceId    | The Reference ID associated with the goal.                                                                                                  | String     | 04dd272f-6c8c-4957-9714-937e6f0847c2                         |
  | projectId      | The ID of the Project related to the record.                                                                                                | String     | 5a91d194fde28b0011ce2422                                     |
  | organisationId | The ID of the organization associated with the record.                                                                                      | String     | 5eb4a49d426cd3d05f2892a9                                     |
  | createdAt      | Timestamp indicating when the goal was created.                                                                                             | Timestamp  | 2024-09-19T10:04:06.000Z                                     |
  | lastChanged    | Timestamp of the last update to the goal.                                                                                                   | Timestamp  | 2024-09-19T10:04:06.000Z                                     |
  | createdBy      | The identifier for the user who created the goal.                                                                                           | String     | 63bff78cd74f7c7f7dd12944                                     |
  | lastChangedBy  | The identifier for the user who last modified the goal.                                                                                     | String     | 63bff78cd74f7c7f7dd12944                                     |
</Accordion>

### GoalSteps

The GoalSteps collection contains records about a specific Goal step from [goals](/ai/agents/analyze/goals-and-tasks/overview).
Each goal is logged as a single record.

<Tabs>
  <Tab title="OData 2.4">
    `https://odata-trial.cognigy.ai/v2.4/GoalSteps?apikey=<your-api-key>`
  </Tab>
</Tabs>

**Response**

When retrieving this collection, the endpoint returns the following fields:

<Accordion title="Response Fields">
  | Field Name     | <div>Description</div>                                                                                                                                                                                                    | Field Type | Example                                                      |
  | -------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------- | ------------------------------------------------------------ |
  | goalStepId     | The goal step ID.                                                                                                                                                                                                         | String     | 66eac7950dc3fd6e4b3bcd8c                                     |
  | name           | The name of the goal step.                                                                                                                                                                                                | String     | Start                                                        |
  | version        | The UUID for tracking step the version of the specific step configuration within the goal. Each update or change of the goal configuration creates a new version.                                                         | String     | 6cb467d0-9f79-4df0-ad31-991d1bb91910                         |
  | description    | The description of the goal step.                                                                                                                                                                                         | String     | The goal is to gather metrics for a flight booking use case. |
  | order          | The order of the step within the goal configuration.                                                                                                                                                                      | Integer    | null                                                         |
  | text           | Additional text information for the goal step                                                                                                                                                                             | String     | null                                                         |
  | type           | The type of the goal step: <br />`start` — the initial (first) step of the goal. It shows the beginning of the process.<br />`completion` — the completion (last) step of the goal. It shows that the objectives are met. | String     | start                                                        |
  | goalId         | The goal ID for the goal associated with the step.                                                                                                                                                                        | String     | 66eac7950dc3fdb39c3bcd8b                                     |
  | projectId      | The ID of the Project related to the record.                                                                                                                                                                              | String     | 5a91d194fde28b0011ce2422                                     |
  | organisationId | The ID of the organization associated with the record.                                                                                                                                                                    | String     | 5eb4a49d426cd3d05f2892a9                                     |
</Accordion>

### GoalStepMetrics

The GoalStepMetrics collection contains records about specific goal-related events from [goals](/ai/agents/analyze/goals-and-tasks/overview).
Each metric is logged as a single record.

<Tabs>
  <Tab title="OData 2.4">
    `https://odata-trial.cognigy.ai/v2.4/GoalStepMetrics?apikey=<your-api-key>`
  </Tab>
</Tabs>

**Response**

When retrieving this collection, the endpoint returns the following fields:

<Accordion title="Response Fields">
  | Field Name       | <div>Description</div>                                                                                                                                               | Field Type | Example                              |
  | ---------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------- | ------------------------------------ |
  | goalStepMetricId | The metric ID for the goal step.                                                                                                                                     | String     | 66ebf5a2cabacf6ecb0b5a81             |
  | name             | The name of the metric.                                                                                                                                              | String     | null                                 |
  | version          | The UUID for tracking step the version of the specific metric configuration within the goal. Each update or change of the goal configuration creates a new version.  | String     | 6cb467d0-9f79-4df0-ad31-991d1bb91910 |
  | description      | The description of the metric.                                                                                                                                       | String     | null                                 |
  | type             | The category of metrics that can be tracked for goals:<br />- `duration` — time in minutes or hours.<br />- `currency` — currency specified in the Project settings. | String     | currency                             |
  | value            | The value of the metric.                                                                                                                                             | Integer    | 20                                   |
  | goalId           | The goal ID.                                                                                                                                                         | String     | 66ebf5a2cabacf2ac70b5a7e             |
  | goalStepId       | The goal step ID.                                                                                                                                                    | String     | 66ebf5a2cabacf7c810b5a80             |
  | projectId        | The ID of the Project related to the record.                                                                                                                         | String     | 5a91d194fde28b0011ce2422             |
  | organisationId   | The ID of the organization associated with the record.                                                                                                               | String     | 5eb4a49d426cd3d05f2892a9             |
</Accordion>

### GoalEvents

The GoalEvents collection contains records about specific goal-related events from [goals](/ai/agents/analyze/goals-and-tasks/overview).
Each event is logged as a single record.

<Tabs>
  <Tab title="OData 2.4">
    `https://odata-trial.cognigy.ai/v2.4/GoalEvents?apikey=<your-api-key>`
  </Tab>
</Tabs>

**Response**

When retrieving this collection, the endpoint returns the following fields:

<Accordion title="Response Fields">
  | Field Name        | <div>Description</div>                                                                                                                                                                                                                                                                                                              | Field Type         | Example                                                          |
  | ----------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------ | ---------------------------------------------------------------- |
  | id                | The unique ID of the record. Applicable to the OData endpoint 2.4 or later.                                                                                                                                                                                                                                                         | String             | 5a91d194fde28b0011ce2423                                         |
  | version           | The UUID for tracking the current version of the goal configuration. Each update or change of the goal configuration creates a new version.                                                                                                                                                                                         | String             | 8ec4de84-71e2-4c98-a7c3-b18173f36223                             |
  | timestamp         | The date and time indicating when a specific event related to the goal occurred.                                                                                                                                                                                                                                                    | Timestamp          | 2024-09-19T10:00:02.322Z                                         |
  | goalCycleId       | The ID of the goal cycle to track each iteration of a goal. The cycle groups different goal tracking events into one attempt to complete a goal.                                                                                                                                                                                    | String             | a0da54a5-ee2e-43be-b743-23d4bccf513c                             |
  | stepId            | The ID of the specific goal step.                                                                                                                                                                                                                                                                                                   | String             | 66ebeca3cabacf41170b4a6d                                         |
  | goalId            | The goal ID.                                                                                                                                                                                                                                                                                                                        | String             | 66ebeca3cabacf6c820b4a6c                                         |
  | sessionId         | The session ID.                                                                                                                                                                                                                                                                                                                     | String             | session-2a38246e-0395-49a5-876d-dd5133fd5418                     |
  | projectId         | The ID of the Project related to the record.                                                                                                                                                                                                                                                                                        | String             | 5a91d194fde28b0011ce2422                                         |
  | organisationId    | The ID of the organization associated with the record.                                                                                                                                                                                                                                                                              | String             | 5eb4a49d426cd3d05f2892a9                                         |
  | expiresAt         | The [expiration date](/ai/administer/access/management-ui) for the goal record. After that time, the goal record will be removed from the Analytics database.                                                                                                                                                                       | Timestamp          | null                                                             |
  | localeName        | The name of the locale.                                                                                                                                                                                                                                                                                                             | String             | English                                                          |
  | endpointName      | The name of the Endpoint.                                                                                                                                                                                                                                                                                                           | String             | LA Chatbot                                                       |
  | endpointType      | The type of the Endpoint.                                                                                                                                                                                                                                                                                                           | String             | webchat2                                                         |
  | endpointUrlToken  | The URL token for the Endpoint.                                                                                                                                                                                                                                                                                                     | String             | ea30b8f20db52f9d86ea36fd55a7d66bd2c4c60eb24ac7ad52f1c9e173dd4cdb |
  | channel           | The channel through which the record was received.                                                                                                                                                                                                                                                                                  | String             | admin-webchat                                                    |
  | localeReferenceId | The ID of the locale.                                                                                                                                                                                                                                                                                                               | String             | 7eff993c-b801-4556-b111-1c319e8577cf                             |
  | snapshotId        | The ID of the Snapshot.                                                                                                                                                                                                                                                                                                             | String             | 63ff0cc47a466cab278fd19b                                         |
  | snapshotName      | The name for the Snapshot used by the Endpoint.                                                                                                                                                                                                                                                                                     | String             | Bot Release 2.2                                                  |
  | isMasked          | A flag indicating whether the record has personally identifiable information (PII) has been masked using mask and keep analytics options from either the [Deactivate Profile](/ai/agents/develop/node-reference/analytics/deactivate-profile) or [Delete Profile](/ai/agents/develop/node-reference/analytics/delete-profile) Node. | Boolean (nullable) |                                                                  |
</Accordion>

## Cognigy OData Collections: Entity Relationship Diagram (ERD)

The Entity Relationship Diagram (ERD) describes the relationships between the collection entities.

```mermaid theme={null}
erDiagram
    %% ERD diagram for Cognigy OData v2.4 collections 
    %% Layout: Top-down data flow — Sessions → ExecutedSteps → Goals

    %% ========== Session Layer ==========
    Sessions ||--o{ Conversations         : "logs messages"
    Sessions ||--o{ ExecutedSteps         : "executes"
    Sessions ||--o{ Analytics             : "summarized by"
    Sessions ||--o{ LiveAgentEscalations  : "escalates via"
    Sessions ||--o{ GoalEvents            : "tracks goals via"

    Sessions {
        string id PK
        string sessionId UK
        string userId
        string goals
        string stepPath
        number stepsCount
        datetime startedAt
        string projectId
        string organisationId
        string more "..."
    }

    Conversations {
        string id PK
        string sessionId FK
        string inputId
        string type
        string source
        string inputText
        datetime timestamp
        string projectId
        string organisation
        string more "..."
    }

    Analytics {
        string id PK
        string sessionId FK
        string flowReferenceId
        string organisation
        string projectId
        string intent
        string completedGoalsList
        datetime timestamp
        number executionTime
        string more "..."
    }

    LiveAgentEscalations {
        string id PK
        string sessionId FK
        string organisationId
        string projectId
        string status
        string teamName
        string agentName
        datetime timestamp
        string more "..."
    }

    %% ========== Execution Layer ==========
    Steps         ||--o{ ExecutedSteps : "executed as"
    ExecutedSteps ||--o{ ExecutedSteps : "parentStep (self)"

    ExecutedSteps {
        string id PK
        string sessionId FK
        string parentStep FK
        string entityReferenceId FK
        string flowReferenceId
        string stepLabel
        datetime timestamp
        string projectId
        string organisationId
        string more "..."
    }

    Steps {
        string id PK
        string entityReferenceId UK
        string flowReferenceId
        string label
        string type
        string flowName
        string projectName
        string more "..."
    }

    %% ========== Goal Layer ==========
    Goals     ||--o{ GoalSteps       : "has"
    Goals     ||--o{ GoalStepMetrics : "measured by"
    Goals     ||--o{ GoalEvents      : "tracked by"
    GoalSteps ||--o{ GoalStepMetrics : "measured by"

    Goals {
        string goalId PK
        string name
        string version
        string referenceId
        string projectId
        string organisationId
        datetime createdAt
        string more "..."
    }

    GoalSteps {
        string goalStepId PK
        string goalId FK
        string name
        string type
        int order
        string projectId
        string organisationId
        string more "..."
    }

    GoalStepMetrics {
        string goalStepMetricId PK
        string goalId FK
        string goalStepId FK
        string name
        string type
        int value
        string projectId
        string organisationId
        string more "..."
    }

    GoalEvents {
        string id PK
        string goalId FK
        string sessionId FK
        string stepId
        string goalCycleId
        string version
        datetime timestamp
        string projectId
        string organisationId
        string more "..."
    }

    %% ========== VISUAL LEGEND ==========
    Legend
    Legend {
        string info "One-to-Many: ||--o{ (1..*)"
        string info2 "One-to-One: O--O (1..1)"
        string info3 "Many-to-Many: }o--o{ (*..*)"
        string info4 "Optional-to-One: |o--|| (0..1)"
        string info5 "Strong One-to-One: ||--||"
        string abbr "PK=Primary Key, FK=Foreign Key, UK=Unique Key"
    }
    classDef legend fill:#fefce8,stroke:#facc15,color:#444
    class Legend legend
```

## Integrations

The Cognigy OData endpoint can integrate with all tools compatible with the OData protocol version 4.
The table outlines popular integration tools, including data visualization tools, recommended for integration.

| Tool                                                                | Description                                                                                                                                                                                                                                                                                            |
| ------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| Microsoft Excel                                                     | Microsoft Excel is a spreadsheet application widely used for data analysis and reporting. When connecting from Microsoft Excel 2016, use the PowerQuery feature, which can be found under `Data > Get & Transform > New Query > From Other Sources > From OData Feed`.                                 |
| Microsoft Power BI                                                  | Microsoft Power BI is a business analytics service that provides interactive visualizations and business intelligence capabilities. For integrating with the Cognigy OData endpoint, refer to the [Microsoft Power BI](https://docs.microsoft.com/en-us/power-bi/desktop-connect-odata) documentation. |
| OData Consumer Ecosystem                                            | The OData Consumer Ecosystem encompasses a variety of tools and applications that support the consumption of OData services. For a full list of available OData Consumer options, refer to the [OData](https://www.odata.org/ecosystem/#consumers) documentation.                                      |
| Client Libraries in .NET, Java, JavaScript, C++ and other platforms | Client libraries are available in multiple programming languages and platforms, providing developers with OData protocol implementations. For a full list of available libraries, refer to the [OData](https://www.odata.org/libraries/) documentation.                                                |

If you want to use another tool not listed in the table, verify its compatibility with the OData endpoint on the [OData website](https://www.odata.org/) and in the tool's documentation before integrating. For example, the data visualization tool Tableau can't be used for integration because it [doesn't support the OData protocol version 4](https://help.tableau.com/current/pro/desktop/en-us/examples_odata.htm) on which the OData endpoint is based.

## Keyset Pagination

Keyset pagination is a more efficient way to retrieve data from large datasets where you need to set `$skip` to high values, such as `10000`. To use keyset pagination, follow these steps:

<Accordion title="Keyset Pagination">
  1. Retrieve the first page with the `$top` and `$orderby` query options, for example, `https://odata-trial.cognigy.ai/v2.4/Conversations?apikey=<your-api-key>&$top=10000&$orderby=timestamp asc`.
  2. Use the last `timestamp` value from the previous page in the next page request, for example, `https://odata-trial.cognigy.ai/v2.4/Conversations?apikey=<your-api-key>&$filter=timestamp gt <LAST_TIMESTAMP_FROM_PREVIOUS_PAGE>&$top=10000&$orderby=timestamp asc`.
  3. Repeat step 2 until you have retrieved all the records.
</Accordion>

## More Information

* [Collecting Data](/ai/agents/analyze/collecting-data)
* [Insights](/ai/agents/analyze/insights)

***

<span><sup>1</sup>: Microsoft Excel and Power BI do not support for the `$count` query. Use Postman or other options. To perform this query, consider using tools such as Postman or other applications that support advanced API querying.</span>


# Analytics
Source: https://docs.cognigy.com/ai/agents/analyze/overview



When working on conversational AI projects, keeping track of how your AI and human agents are performing is essential. Cognigy offers a variety of tools that can help you analyze and evaluate your agents' effectiveness. You can use these tools to monitor important metrics, such as:

* Average speed of answer
* Resolution rate
* Customer satisfaction

With this data, you can find areas where your agents may need more training or improvement, and make informed decisions on how to optimize their performance.

## Cognigy Insights

[Cognigy Insights](/ai/agents/analyze/insights) is Cognigy's built-in analytics solution that provides advanced analytics and visualizations for the interactions between end users and both AI and human agents. With Insights, you can view data, track metrics, analyze performance, and export data to use in other business intelligence tools for each [Project](/ai/agents/develop/projects-and-flows/projects).

## OData

Cognigy provides an [OData Analytics Endpoint](/ai/agents/analyze/odata) that lets administrators import agent performance data into external data visualization tools, allowing for the creation of custom dashboards.

## Call Tracing

[Call tracing](/ai/agents/analyze/call-tracing) involves recording calls made by a voice AI Agent using Voice Gateway for debugging purposes. Call tracing helps you troubleshoot issues, identify latency sources, and ensure voice agents function seamlessly.

## Contact Profiles

[Contact Profiles](/ai/agents/analyze/contact-profiles) store user information, and you can access them through Flows and Endpoints. Contact Profiles let you store data persistently and personalize conversations with end users.

## Dashbot

Any Endpoint can be connected to the third-party analytics vendor [Dashbot](/ai/agents/deploy/endpoints/data-protection-and-analytics).

## More information

* [Cognigy Insights](/insights/overview)
* [OData](/ai/agents/analyze/odata)
* [Call tracing](/ai/agents/analyze/call-tracing)
* [Contact Profiles](/ai/agents/analyze/contact-profiles)
* [Dashbot](/ai/agents/deploy/endpoints/data-protection-and-analytics)


# User and Session Identification
Source: https://docs.cognigy.com/ai/agents/analyze/user-and-session-identification



Cognigy uses unique user IDs and session IDs to keep track of users' interactions with AI Agents.

## Restrictions

* You must use session IDs only within one organization in your Cognigy installation.
* Except for the [Voice Copilot Endpoint](/ai/agents/deploy/endpoint-reference/voice-copilot), reusing the combination of user and session IDs across Endpoints causes Endpoint collisions and unexpected behavior.
* You must keep user and session IDs unchanged during interactions with AI Agents. Changing user and session IDs with [transformers](/ai/for-developers/transformers/overview), [xApps](/xApps/overview), [Extensions](/ai/for-developers/extensions), or [Code Nodes](/ai/for-developers/code/overview) can cause unexpected issues and terminate interactions with AI Agents.

## User ID

*User ID* is a unique identifier assigned to each user that helps manage user data safely, personalize user experiences, and keep track of user history and preferences. The AI Agent uses this information to improve its responses and customize the conversation according to the user's needs.
The user ID can be hashed in [transformers](/ai/for-developers/transformers/overview) to enhance privacy and support GDPR compliance.

## Session ID

*Session ID* is a unique identifier given to each session initiated by a user. A session represents a period of interaction between the user and the AI Agent. The AI Agent uses session IDs to manage user sessions, track user activity, ensure security, and provide personalized experiences within the conversation.

## Get User and Session IDs

To get the user ID and session ID, you can:

* Retrieve them from the [Input object](/ai/agents/develop/ai-agent-memory/input) by using the `input.userId` and `input.sessionId` [CognigyScript](/ai/platform-features/cognigyscript) expressions, or the **User ID** and **Session ID** Tokens.
* Copy them from the [Logs](/ai/agents/test/logs) page.

## More Information

* [Input](/ai/agents/develop/ai-agent-memory/input)
* [Logs](/ai/agents/test/logs)
* [xApps](/xApps/overview)


# Deploying your AI Agents
Source: https://docs.cognigy.com/ai/agents/deploy/deploying-your-ai-agents

Deploy AI Agents via Cognigy Endpoints and move AI Agents across Projects with Snapshots.

Deploy AI Agents using Cognigy Endpoints such as Webchat, WhatsApp, Voice Gateway, Amazon, Twilio, and others. Learn how to use Snapshots to save and move AI Agents across different Projects and environments.

<CardGroup>
  <Card title="Endpoints" href="/ai/agents/deploy/endpoints/overview">
    <br />Learn how to connect your AI Agents to different third-party providers via Endpoints.
  </Card>

  <Card title="Endpoint Reference" href="/ai/agents/deploy/endpoint-reference/overview">
    <br />Explore detailed documentation for configuring Endpoints.
  </Card>

  <Card title="Snapshots" href="/ai/agents/deploy/snapshots">
    <br />See how to create and manage Snapshots to move AI Agents across Projects and environments.
  </Card>
</CardGroup>


# 8x8
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/8x8



<a href="/release-notes/2026.16"><Badge>Updated in 2026.16.0</Badge></a>

<Frame>
  <img alt="8x8 Endpoint logo" />
</Frame>

The 8x8 Endpoint enables you to connect your AI Agent to the [8x8 platform](https://www.8x8.com/).

## Prerequisites

* Access to [8x8 platform](https://www.8x8.com/).

## Generic Endpoint Settings

Learn about the generic Endpoint settings on the following pages:

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Session Management](/ai/agents/deploy/endpoints/session-management)
* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)
* [Transformer Functions](/ai/for-developers/transformers/overview)
* [File Storage](/ai/agents/deploy/endpoints/file-storage)
* [Copilot](/agent-copilot/configure/copilot-section)

## Specific Endpoint Settings

<Accordion title="8x8 Setup">
  | Parameter  | Description                                                                                   |
  | ---------- | --------------------------------------------------------------------------------------------- |
  | Base URL   | The URL of the 8x8 Server.                                                                    |
  | API Key    | The header value required for authenticating API calls to the 8x8 system.                     |
  | 8x8-Tenant | The Tenant ID, the field is mandatory if a customer has multiple tenants in the organization. |
</Accordion>

## How to Set Up

### Setup on the Third-Party Provider Side

<Accordion title="Configure 8x8">
  1. Log in to the 8x8 Contact Center for admins.
  2. On the **My Applications** page, select **Configuration Manager for Contact Center**.
  3. In the left-side menu, click **Integration**, then go to the **Webhooks** tab.
  4. On the **Webhooks List** tab, click **+ Add webhook**.
  5. Fill in the following fields:
     * **Name** — specify a unique name.
     * **URL** — paste the URL in the following pattern: `https://endpoint-your-environment/handover/eightByEight`. For instance, in a Cognigy trial environment, the URL would be `https://endpoint-trial.cognigy.ai/handover/eightByEight`.
     * **Chat API Version** — select **Chat Gateway v1.0**.
  6. To check if the connection works, click **Test Connection** and save changes.
  7. In the left-side menu, click **Queues/Skills**.
  8. On the **Queue List** tab, create a queue with a chat type by clicking **+ Chat**.
  9. Fill in the following fields:
     * **Queue name** — specify a unique name.
     * **Queue description** — specify a unique description for the queue to differentiate it from others.
       Save changes.
  10. On the **Members** tab, assign and enable human agents who will operate within your queue to assist end users. Save changes.
  11. In the left-side menu, click **Channels**.
  12. On the **Chat API** tab, click **+ New Channel**.
  13. Fill in the following fields:
      * **Name** — specify a unique name.
      * **Description** — specify a description of your channel.
      * **Webhook** — select the webhook that you recently created.
      * **Queue** — select the queue that you recently created.
        Save changes.
</Accordion>

### Setup on the Cognigy Side

<Accordion title="Configure 8x8 Endpoint">
  1. In the left-side menu of your Project, click **Deploy > Endpoints**.
  2. On the **Endpoints** page, click **+ New Endpoint**.
  3. In the **New Endpoint** section, do the following:
     1. Select the **8x8** Endpoint type.
     2. Add a unique name.
     3. Select a relevant Flow from the list.
  4. Click **Save**.
  5. In the Endpoint, go to the **8x8 Setup** settings.
  6. Specify a Base URL, API Key, and 8x8-Tenant that you received from the 8x8 server administrator.
  7. Copy the URL from the **Endpoint URL** field.
  8. *(Optional)* Activate the **Accept Conversation Active Event** setting. When activated, the setting allows the system to accept and process `CONVERSATION UPDATE` events with the state set as `ACTIVE`. These events will be sent to the Flow for further handling. This setting allows accessing the event payload via the [Input object](/ai/agents/test/interaction-panel/overview), which provides relevant information related to the ongoing conversation.
     * If the toggle is activated, the end user doesn't have to send the first message. The bot will automatically start replying.
     * If the toggle is not activated, the end user needs to engage in conversation with the bot first in order to receive a reply.
  9. Click **Save**.

  To check the AI Agent's performance, go to the installation that your server administrator has deployed.
</Accordion>

## More Information

* [8x8 handover provider](/ai/escalate/handover-reference/8x8)


# Amazon Alexa
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/amazon-alexa

The Alexa Endpoint enables you to connect your AI Agent to the Amazon Alexa Skills.

<a href="/release-notes/2026.17"><Badge>Updated in 2026.17</Badge></a>

<Frame>
  <img alt="Amazon Alexa Endpoint configuration in Cognigy" />
</Frame>

<Warning>
  As of Cognigy 2026.17.0, the Amazon Alexa Endpoint is deprecated. You can't create new Alexa Endpoints. The removal is planned for January 2028. Existing Alexa Endpoints continue to work until the removal date.
</Warning>

The Alexa Endpoint enables you to connect your AI Agent to the [Amazon Alexa Skills](https://alexa.amazon.com).

## Prerequisites

* Access to [Amazon Alexa](https://alexa.amazon.com).

## Restrictions

* This Endpoint uses the [Alexa NLU Connector](/ai/platform-features/nlu/external/nlu-connector-reference/alexa), meaning all Natural Language Understanding (NLU) is processed by Amazon's NLU engine. Consequently, the Cognigy NLU model can't be used in this configuration.
* Amazon Alexa passes only the intent name to the Cognigy Flow, not the actual user input. Therefore, any Nodes that rely on user inputs, such as If Nodes, or Intent rules won't function as intended in this setup.

## Generic Endpoint Settings

Learn about the generic Endpoint settings on the following pages:

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Transformer Functions](/ai/for-developers/transformers/overview)
* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)

## Specific Endpoint Settings

<Accordion title="Alexa Settings">
  | Parameter           | Description                                                                                                                                                                                                                                                                                                                                                                                   |
  | ------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
  | Reparse Alexa Slots | Allows slots detected by Alexa to be passed to Cognigy and processed by the Cognigy NLU.                                                                                                                                                                                                                                                                                                      |
  | Deployment          | Select the Alexa skill you configured on the Amazon Alexa side. You can also log in to Amazon from the [My Profile](/ai/administer/user-menu/my-profile) page. Once logged in, you will be able to select the **Alexa Skill** you want to connect with your Cognigy Endpoint. After selecting a skill, click the **Deploy** button to configure your skill to use this Endpoint when invoked. |
</Accordion>

## How to Set Up

<AccordionGroup>
  <Accordion title="1. Create an Alexa Endpoint">
    1. In the left-side menu of your Agent, click **Deploy > Endpoints**.
    2. On the **Endpoints** page, click **+ New Endpoint**.
    3. In the **New Endpoint** section, do the following:
       1. Select the **Alexa** Endpoint type.
       2. Specify a unique name.
       3. Select a relevant Flow from the list. Save changes.
    4. In the Endpoint, go to the **Alexa Settings**.
    5. Click **Login With Amazon** and log in to your Amazon Developer Account.
  </Accordion>

  <Accordion title="2. Create an Alexa Skill">
    1. In the Amazon Developer Portal, click the **Alexa** tab and select **Create Skill**.
    2. Name the skill and select the **Custom** model for flexibility in building the skill. Save and proceed to the Skill configuration.
    3. Add the following built-in intents to handle **Yes/No** questions in your Flow's Question Nodes.
       * `AMAZON.YesIntent`
       * `AMAZON.NoIntent`
         You can use additional Amazon-built-in intents by accessing them via `ci.intent` in Cognigy.
    4. Go to the **Endpoints** page. In the **Service Endpoint Type** section, select **HTTPS**.
    5. In the **Default Region** field, enter the copied Endpoint URL. From the **Select SSL certificate type** list, select **My development endpoint is a sub-domain of a domain that has a wildcard certificate from a certificate authority**. You can also configure another region if needed. Save the changes.
  </Accordion>

  <Accordion title="3. Finish Setup on the Cognigy side">
    1. Go to the Alexa Endpoint that you created previously.
    2. Select the target Alexa skill from the **Deployment** list.
    3. Click **Deploy** to automatically configure the skill to use the Endpoint.
  </Accordion>
</AccordionGroup>

## More Information

* [Cognigy Help Center: Amazon Alexa Intro ](https://support.cognigy.com/hc/en-us/articles/360016192979)
* [Cognigy Help Center: Amazon Alexa Deployment ](https://support.cognigy.com/hc/en-us/articles/360016193099-Amazon-Alexa-Deploy-an-Endpoint)
* [Cognigy Help Center: Amazon Alexa Account Linking ](https://support.cognigy.com/hc/en-us/articles/360016153600-Amazon-Alexa-Account-Linking)


# Amazon Lex
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/amazon-lex



<Frame>
  <img alt="Amazon Lex Endpoint logo" />
</Frame>

The Amazon Lex Endpoint enables you to connect your AI Agent with [Amazon Lex](https://aws.amazon.com/lex/).

## Prerequisites

* Access to [Amazon Lex](https://aws.amazon.com/lex/).
* Access to [AWS Lambda](https://aws.amazon.com/lambda/).

## Generic Endpoint Settings

Learn about the generic Endpoint settings on the following pages:

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Transformer Functions](/ai/for-developers/transformers/overview)
* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)

## Specific Endpoint Settings

<Accordion title="Amazon Lex Settings">
  | Parameter           | Description                                                                                                                                                                                                                         |
  | ------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
  | Reparse Alexa Slots | Allows Slots detected by Amazon Lex to be passed to Cognigy and be processed by the Cognigy NLU model. To make this parameter available, select **Amazon Lex (Built-in)** from the **NLU Connector** list in the Endpoint settings. |
</Accordion>

## How to Set Up

## Setup on the Cognigy Side

<Accordion title="Create an Amazon Lex Endpoint">
  1. In the left-side menu of your Agent, click **Deploy > Endpoints**.
  2. On the **Endpoints** page, click **+ New Endpoint**.
  3. In the **New Endpoint** section, do the following:
     1. Select the **Lex** Endpoint type.
     2. Specify a unique name.
     3. Select a relevant Flow from the list. Save changes.
  4. Go to the **Configuration Information** section and copy the Endpoint URL for later use in Amazon Lex.
  5. *(Optional)* In the **Amazon Lex Settings** section, activate the **Reparse Alexa Slots** option only if you want the Cognigy NLU model to process Amazon Lex slots.
</Accordion>

## Setup on the Amazon Side

Select the use case:

<AccordionGroup>
  <Accordion title="Use Amazon Lex with Amazon NLU">
    Use your existing Amazon Lex bots along with their native NLU capabilities to process user input within Cognigy.

    1. Open the [Lambda console](https://console.aws.amazon.com/lambda).

    2. Select **Create function**. Configure the following settings:
       * **Function name** – enter a name for the function.
       * **Runtime** – select **Node.js 22.x**.

    3. Select **Create function**. The console creates a Lambda function with a single source file named `index.mjs`. Edit this file in the built-in code editor. Copy the following code and paste it into your `index.mjs` file:

       ```js theme={null}
       const COGNIGY_ENDPOINT_HOSTNAME = "endpoint-trial.cognigy.ai";
       const COGNIGY_ENDPOINT_PATH = "/your-urltoken"; 

       // Change these attributes to the session attributes that contain your userId and sessionId 
       const USERID_ATTRIBUTE = "CustomerNumber";
       const SESSIONID_ATTRIBUTE = "ContactId"; 

       // Pagination doesn't work at this API so we need to specify the maximum number of slots that are used for a single intent as page size 
       const PAGESIZE = 20; 

       const AWS = require("aws-sdk");
       const lexmodelsv2 = new AWS.LexModelsV2({
           apiVersion: "2020-08-07"
       });

       exports.handler = async (event, context, callback) => {
           const slotMap = await retrieveSlots(event);
           const postObject = {
               userId: event["sessionState"]["sessionAttributes"][USERID_ATTRIBUTE] || AWS.util.uuid.v4(),
               sessionId: event["sessionState"]["sessionAttributes"][SESSIONID_ATTRIBUTE] || AWS.util.uuid.v4(),
               text: event["transcriptions"][0]["transcription"],
               data: {
                   request: event,
                   slotMap
               },
           };
           const resultMap = await sendMessageToCognigy(postObject);
           const defaultSessionAtrributes = {
               action: "NONE",
               action_data: "",
           };
           if (resultMap["data"] && resultMap["data"]["connect_action"]) {
               defaultSessionAtrributes["action"] = resultMap["data"]["connect_action"];
           }
           if (resultMap["data"] && resultMap["data"]["connect_action_data"]) {
               defaultSessionAtrributes["action_data"] = JSON.stringify(resultMap["data"]["connect_action_data"]);
           }
           const response = {
               sessionState: {
                   sessionAttributes: {
                       action: defaultSessionAtrributes.action,
                       action_data: JSON.stringify(defaultSessionAtrributes.action_data),
                   },
                   dialogAction: {
                       type: "Close",
                       fulfillmentState: "Fulfilled",
                       message: resultMap["text"],
                   },
                   intent: {
                       confirmationState: "Confirmed",
                       name: event["sessionState"]["intent"]["name"],
                       slots: {},
                       state: "Fulfilled",
                   },
               },
               messages: [{
                   contentType: "PlainText",
                   content: resultMap["text"],
               }, ],
           };
           callback(null, response);
       };
       const https = require("https");
       async function sendMessageToCognigy(postObject) {
           return new Promise((resolve, reject) => {
               const postData = JSON.stringify(postObject);
               const options = {
                   hostname: COGNIGY_ENDPOINT_HOSTNAME,
                   path: COGNIGY_ENDPOINT_PATH,
                   method: "POST",
                   headers: {
                       "Content-Type": "application/json",
                   },
               };
               const request = https.request(options, handleResponse);
               request.on("error", error => {
                   reject(error);
               });
               request.write(postData);
               request.end();

               function handleResponse(response) {
                   let data = "";
                   response.on("data", d => {
                       data += d.toString();
                   });
                   response.on("end", () => {
                       resolve(JSON.parse(data));
                   });
               }
           });
       }
       async function retrieveSlots(event) {
           const params = {
               botId: event.bot.id,
               botVersion: event.bot.version,
               localeId: event.bot.localeId,
               filters: [{
                   name: "IntentName",
                   operator: "EQ",
                   values: [event.sessionState.intent.name],
               }, ],
           };
           const intents = await lexmodelsv2.listIntents(params).promise();
           const slotParams = {
               botId: event.bot.id,
               botVersion: event.bot.version,
               localeId: event.bot.localeId,
               intentId: intents.intentSummaries[0].intentId,
               maxResults: PAGESIZE,
           };
           const slotList = await lexmodelsv2.listSlots(slotParams).promise();
           const slots = slotList.slotSummaries.reduce((obj, item) => Object.assign(obj, {
               [item.slotName]: item.slotTypeId
           }), {}, );
           return slots;
       }
       ```

       In the script, replace the `COGNIGY_ENDPOINT_HOSTNAME` and `COGNIGY_ENDPOINT_PATH` variables with your Endpoint hostname and the URL token of your Endpoint.

       For example, if your Cognigy trial Endpoint URL is `https://endpoint-trial.cognigy.ai/175525264283054b52edb3b9ece270755e46e651ac00015fad9f1175bb47c110`, specify the following values:

       ```js theme={null}
       const COGNIGY_ENDPOINT_HOSTNAME = "endpoint-trial.cognigy.ai";
       const COGNIGY_ENDPOINT_PATH = "/175525264283054b52edb3b9ece270755e46e651ac00015fad9f1175bb47c110";
       ```

       In the **Deploy** section, select **Deploy** to update your function's code. Then, to run your code, select **Create test event** in the **Test events** section.
       Modify the service that was created for the Lambda function. Add the following permissions: `lex:ListIntents` and `lex:ListSlots`. Alternatively, you can also use a predefined role such as `AmazonLexReadOnly`. Save the Lambda function.

    4. Open the [Amazon Lex console](https://console.aws.amazon.com/lex/).

    5. Select **Create bot**. For the **Creation** method, select **Start with an example**. In the **Example bots** section, select **OrderFlowers** from the list.

    6. In the **Bot configuration** section, give the bot a name and an optional description. The name must be unique in your account.

    7. Open your first intent. Go to the **Fulfillment** section and enable **Fulfillment Lambda code hook** in the **Advanced options**.

    8. Click **Update Options** and **Save intent**. Open each intent in the current language, including the **FallbackIntent**, and repeat these steps. Build the language, then repeat these steps for additional languages if required.

    9. Go to **Bot versions** and create a new version. Then, go to **Aliases** and create a new alias and link it to the latest bot's version. Reopen the alias and click the name of each language that is supported.

    10. Select the previously created Lambda function with its latest version and save changes.
  </Accordion>

  <Accordion title="Use Amazon Lex with Cognigy NLU">
    Use the Amazon Lex Endpoint to transcribe user input, while Cognigy's NLU model handles intent recognition and slot parsing.

    1. Open the [Lambda console](https://console.aws.amazon.com/lambda).

    2. Select **Create function**. Configure the following settings:
       * **Function name** – enter a name for the function.
       * **Runtime** – select Node.js 22.x.

    3. Select **Create function**. The console creates a Lambda function with a single source file named `index.mjs`. Edit this file in the built-in code editor. Copy the following code and paste it into your `index.mjs` file:

       ```js theme={null}

       const COGNIGY_ENDPOINT_HOSTNAME = "endpoint-trial.cognigy.ai";
       const COGNIGY_ENDPOINT_PATH = "/your-url-token";

       // Change these attributes to the session attributes that contain your userId and sessionId
       const USERID_ATTRIBUTE = "CustomerNumber";
       const SESSIONID_ATTRIBUTE = "ContactId";

       const AWS = require("aws-sdk");

       exports.handler = async (event, context, callback) => {
           try {
               const postObject = {
                   userId: event["sessionState"]["sessionAttributes"][USERID_ATTRIBUTE] || AWS.util.uuid.v4(),
                   sessionId: event["sessionState"]["sessionAttributes"][SESSIONID_ATTRIBUTE] ||
                       AWS.util.uuid.v4(),
                   text: event["transcriptions"][0]["transcription"],
                   data: {},
               };

               const defaultSessionAtrributes = {
                   action: "NONE",
                   action_data: "",
               };

               const resultMap = await sendMessageToCognigy(postObject);

               if (resultMap["data"] && resultMap["data"]["connect_action"]) {
                   defaultSessionAtrributes["action"] = resultMap["data"]["connect_action"];
               }

               if (resultMap["data"] && resultMap["data"]["connect_action_data"]) {
                   defaultSessionAtrributes["action_data"] = JSON.stringify(resultMap["data"]["connect_action_data"]);
               }

               const response = {
                   sessionState: {
                       sessionAttributes: {
                           action: defaultSessionAtrributes.action,
                           action_data: defaultSessionAtrributes.action_data,
                       },
                       dialogAction: {
                           type: "Close",
                           fulfillmentState: "Fulfilled",
                           message: resultMap["text"],
                       },
                       intent: {
                           confirmationState: "Confirmed",
                           name: event["sessionState"]["intent"]["name"],
                           slots: {},
                           state: "Fulfilled",
                       },
                   },
                   messages: [{
                       contentType: "PlainText",
                       content: resultMap["text"],
                   }, ],
               };

               callback(null, response);
           } catch (e) {
               return e;
           }
       };

       const https = require("https");

       async function sendMessageToCognigy(postObject) {
           return new Promise((resolve, reject) => {
               const postData = JSON.stringify(postObject);
               const options = {
                   hostname: COGNIGY_ENDPOINT_HOSTNAME,
                   path: COGNIGY_ENDPOINT_PATH,
                   method: "POST",
                   headers: {
                       "Content-Type": "application/json",
                   },
               };

               const request = https.request(options, handleResponse);

               request.on("error", error => {
                   reject(error);
               });

               request.write(postData);
               request.end();

               function handleResponse(response) {
                   let data = "";

                   response.on("data", d => {
                       data += d.toString();
                   });

                   response.on("end", () => {
                       resolve(JSON.parse(data));
                   });
               }
           });
       }
       ```

       In the script, replace the `COGNIGY_ENDPOINT_HOSTNAME` and `COGNIGY_ENDPOINT_PATH` variables with your Endpoint hostname and the URL token of your Endpoint.

       For example, if your Cognigy trial Endpoint URL is `https://endpoint-trial.cognigy.ai/175525264283054b52edb3b9ece270755e46e651ac00015fad9f1175bb47c110`, specify the following values:

       ```js theme={null}
       const COGNIGY_ENDPOINT_HOSTNAME = "endpoint-trial.cognigy.ai";
       const COGNIGY_ENDPOINT_PATH = "/175525264283054b52edb3b9ece270755e46e651ac00015fad9f1175bb47c110";
       ```

       In the **Deploy** section, select **Deploy** to update your function's code. Then, to run your code, select **Create test event** in the **Test events** section.

    4. Open the [Amazon Lex console](https://console.aws.amazon.com/lex/).

    5. Select **Create bot**. For the **Creation** method, select **Start with an example**. In the **Example bots** section, select **OrderFlowers** from the list.

    6. In the **Bot configuration** section give, the bot a name and an optional description. The name must be unique in your account.

    7. Create a single intent with one sample utterance that will never get called, for example, `THIS_WILL_NEVER_BE_CALLED`. Save the intent and open **FallbackIntent**.

    8. Go to the **Fulfillment** section and enable **Fulfillment Lambda code hook** in **Advanced options**.

    9. Click **Update Options** and **Save intent**. Build the language.

    10. Go to **Bot versions** and create a new version. Then, go to **Aliases** and create a new alias and link it to the latest bot's version. Reopen the alias and click the name of the primary language.

    11. Select the previously created Lambda function with its latest version and save changes.
  </Accordion>
</AccordionGroup>

## More Information

* [Cognigy Help Center: Amazon Connect: Intro](https://support.cognigy.com/hc/en-us/articles/6229662732444)
* [Cognigy Help Center: Amazon Connect: Setup Amazon Connect](https://support.cognigy.com/hc/en-us/articles/6229664156060)
* [Cognigy Help Center: Amazon Connect: Control Amazon Connect from Cognigy](https://support.cognigy.com/hc/en-us/articles/6229683335196)


# AudioCodes
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/audiocodes



<Frame>
  <img alt="AudioCodes Endpoint logo" />
</Frame>

<Warning>
  As of April 2025, the AudioCodes Endpoint is deprecated. The full removal of the AudioCodes Endpoint is planned for Q3 2026. Migrate to the [Voice Gateway Endpoint](/ai/agents/deploy/endpoint-reference/voice-gateway) before the removal date.
</Warning>

The AudioCodes Endpoint enables you to connect your AI Agent to your contact center or other phone numbers using the [AudioCodes](https://www.audiocodes.com/) platform. This Endpoint allows your customers to speak with your AI Agents instead of using a chat.

## Prerequisites

* A Flow configured with [AudioCodes](/ai/agents/develop/node-reference/voice/audiocodes/overview) or [Voice Generic](/ai/agents/develop/node-reference/voice/generic/overview) Nodes.
* To link the AudioCodes Endpoint to a phone number, contact your Cognigy representative to provision and configure a Voice Gateway instance.
* The Cognigy AudioCodes Endpoint is 100% compatible with [AudioCodes VoiceAI Connect](https://voiceaiconnect.audiocodes.com), where you can sign up for free and get a phone number within minutes.

## Generic Endpoint Settings

Learn about the generic Endpoint settings on the following pages:

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [NLU Connectors](/ai/platform-features/nlu/external/nlu-connectors/overview)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)
* [Transformer Functions](/ai/for-developers/transformers/overview)

## Specific Endpoint Settings

<Accordion title="Voice Gateway Settings">
  | **Parameter** | **Description**                                                                                                                                                                                                                                                                                                                                                                                                                         |
  | ------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
  | Async Mode    | When async mode is enabled for the AudioCodes Endpoint, outputs are sent directly as they are processed in a Flow. This mode is useful when the Flow needs to send an HTTP request. <br /><br />When async mode is off, outputs are sent only after the entire Flow has finished processing the input.<br /><br />The [Inject and Notify](/ai/agents/deploy/endpoints/inject-and-notify) feature only works when async mode is enabled. |
</Accordion>

## How to Set Up

### Configuration on the Cognigy Side

<Accordion title="Create an AudioCodes Endpoint">
  1. In the left-side menu of the Cognigy Project, go to **Deploy > Endpoints**.
  2. On the **Endpoints** page, click **+ New Endpoint**.
  3. In the **New Endpoint** section, do the following:
     1. Select the **AudioCodes** Endpoint type.
     2. Add a unique name.
     3. Select a relevant Flow from the list.
  4. *(Optional)* Activate the **Async Mode** setting.
</Accordion>

## Additional Information

<Accordion title="Call Metadata">
  AudioCodes identifies information about the caller and adds it to the [Input](/ai/agents/develop/ai-agent-memory/input) object as `input.data.numberMetaData`.

  | Parameter            | Type    | Description                                                                                                                                                                                    | Example              |
  | -------------------- | ------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------- |
  | `number`             | string  | The complete phone number of the caller, including the country code.                                                                                                                           | `+4921154591991`     |
  | `country`            | string  | The 2-letter ISO country code.                                                                                                                                                                 | `DE`                 |
  | `countryCallingCode` | string  | The country calling code (without the `+` sign).                                                                                                                                               | `49`                 |
  | `nationalNumber`     | string  | The national number, excluding the country code and leading zero.                                                                                                                              | `21154591991`        |
  | `valid`              | boolean | Indicates whether the phone number is valid.                                                                                                                                                   | `true`               |
  | `type`               | string  | The type of phone number. Possible values: `PREMIUM_RATE`, `TOLL_FREE`, `SHARED_COST`, `VOIP`, `PERSONAL_NUMBER`, `PAGER`, `UAN`, `VOICEMAIL`, `FIXED_LINE_OR_MOBILE`, `FIXED_LINE`, `MOBILE`. | `FIXED_LINE`         |
  | `uri`                | string  | The URI representation of the phone number.                                                                                                                                                    | `tel:+4921154591991` |
</Accordion>

## More Information

* [All Endpoints](/ai/agents/deploy/endpoint-reference/overview)
* [Migration to Voice Gateway](/voice-gateway/migrate-from-ac-to-vg)


# Avaya CPaaS
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/avaya-cpaas



<Frame>
  <img alt="Avaya CPaaS Endpoint logo" />
</Frame>

The Avaya CPaaS Endpoint enables you to connect your AI Agent to [Avaya CPaaS](https://www.avaya.com/).

## Prerequisites

* Access to the [Avaya CPaaS](https://accounts.avayacloud.com/).

## Generic Endpoint Settings

Learn about the generic Endpoint settings available on the following pages:

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [NLU Connectors](/ai/platform-features/nlu/external/nlu-connectors/overview)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Transformer Functions](/ai/for-developers/transformers/overview)

## Specific Endpoint Settings

In the Avaya CPaaS Endpoint, you can customize the setup of conversations with Avaya CPaaS.

<AccordionGroup>
  <Accordion title="Avaya Settings">
    | Setting            | Description                                                                                               |
    | ------------------ | --------------------------------------------------------------------------------------------------------- |
    | STT + TTS Language | The language that Avaya CPaaS should use to read out the outputs from the bot.                            |
    | Voice              | The voice to use with Avaya CPaaS.                                                                        |
    | Avaya CPaaS Token  | An optional security token from Avaya CPaaS to validate that the request doesn't come from a third party. |
  </Accordion>

  <Accordion title="Gather Settings">
    | Setting       | Description                                                                                                                                                   |
    | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Action URL    | The URL Avaya CPaaS will use to send the user inputs. Leave empty to send requests to the Avaya CPaaS Endpoint.                                               |
    | Input Type    | The types of input Avaya CPaaS should accept. Default is `digits + voice`.                                                                                    |
    | Hints         | A list of words or phrases that Avaya CPaaS will listen for. Separate each item with a comma.                                                                 |
    | Method        | The HTTP request method used to contact the Action URL. The default value is `POST`.                                                                          |
    | Timeout       | Time in seconds the user has to answer before the call disconnects. The default value is `30`. Note that the timer starts as soon as the bot begins speaking. |
    | Finish On Key | The key the user should press to submit input via DTMF.                                                                                                       |
    | Num digits    | The maximum number of digits the user can enter.                                                                                                              |
  </Accordion>
</AccordionGroup>

## How to Set Up

### Setup on the Cognigy Side

<Accordion title="Create an Avaya CPaaS Endpoint">
  1. In the left-side menu of the Cognigy Project, go to **Deploy > Endpoints**.
  2. On the **Endpoints** page, click **+ New Endpoint**.
  3. In the **New Endpoint** section, do the following:
     1. Select the **Avaya CPaaS** Endpoint type.
     2. Add a unique name.
     3. Select a relevant Flow from the list.
  4. Copy the Endpoint URL for later use in the Avaya CPaaS phone number configuration.
</Accordion>

### Setup on the Avaya CPaaS Side

<Accordion title="Configure Avaya CPaaS">
  To connect the Avaya CPaaS Endpoint in Cognigy to a phone number in Avaya CPaaS, enter the Endpoint URL in the Voice Request URL field within the Avaya CPaaS phone number configuration.
</Accordion>

## More Information

* [All Endpoints](/ai/agents/deploy/endpoint-reference/overview)


# Azure Bot Services
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/azure-bot-services



<a href="/release-notes/2026.17"><Badge>Updated in 2026.17</Badge></a>

<Frame>
  <img alt="Azure Bot Services Endpoint logo" />
</Frame>

The Azure Bot Services Endpoint enables you to connect your AI Agent to the [Azure Bot Services](https://azure.microsoft.com/en-us/services/bot-services/) platform. This integration allows you to deploy your AI Agent to Microsoft Teams and other channels supported by Azure Bot Services, and supports [multiple SSO connections](#set-up-multiple-sso-connections-optional-accordion-title).

## Prerequisites

* Access to the [Azure Bot Services](https://azure.microsoft.com/en-us/services/bot-services/) platform.
* An account on the [Microsoft Entra Admin Center](https://entra.microsoft.com/).
* An account on the [Microsoft Teams Developer Portal](https://dev.teams.microsoft.com/).

## Restrictions

* In Microsoft Teams, postback buttons in [quick reply](/ai/agents/develop/node-reference/basic/say), [gallery](/ai/agents/develop/node-reference/basic/say), [list](/ai/agents/develop/node-reference/basic/say), or [button](/ai/agents/develop/node-reference/basic/say) configurations will always display the button title in the user's chat when clicked. This behavior can't be disabled but doesn't affect other channels supported by the Azure Bot Services Endpoint.

## Generic Endpoint Settings

Learn about the generic Endpoint settings available on the following pages:

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [NLU Connectors](/ai/platform-features/nlu/external/nlu-connectors/overview)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Transformer Functions](/ai/for-developers/transformers/overview)
* [NLU Connectors](/ai/platform-features/nlu/external/nlu-connectors/overview)
* [Session Management](/ai/agents/deploy/endpoints/session-management)
* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)
* [Transformer Functions](/ai/for-developers/transformers/overview)
* [Inject and Notify](/ai/agents/deploy/endpoints/inject-and-notify)
* [Copilot](/agent-copilot/configure/copilot-section)

## Specific Endpoint Settings

<Accordion title="Azure Bot Services Settings">
  | Setting                       | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
  | ----------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
  | Access Scope                  | Defines who can access this app:<br />- **Multi Tenant** — allows users from any Azure Active Directory (AAD) tenant to access the app, for example, users from different organizations.<br />- **Single Tenant** — restricts access to one specific AAD tenant. Only accounts within your organization's directory can log in.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
  | Restrict to Azure AD Accounts | When **Multi Tenant** is selected, this toggle restricts access to Azure AD accounts only. It prevents public access and allows access only for accounts managed in your organization's Azure Active Directory.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
  | Microsoft App ID              | The app ID for your bot. You can find it on the **Settings** page of your bot in the Azure Bot Services Dashboard.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
  | Microsoft App Password        | The Client Secret for your bot. You can find it on the **App Registrations** page in the Azure Bot Services Dashboard.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
  | SSO Connection Name           | The name of the OAuth connection you create in Azure Bot Services.<br />The [Get Token](/ai/agents/develop/node-reference/service/get-token), [Token Status](/ai/agents/develop/node-reference/service/token-status), and [Invalidate Token](/ai/agents/develop/node-reference/service/invalidate-token) Nodes use this connection unless you turn on the **Overwrite Connection** toggle.<br />Depending on your Cognigy version, this connection's token and approval status are accessible from the following fields:<ul><li>Cognigy 2026.16.0 and earlier — `input.data.microsoftSsoToken` and `input.data.microsoftSsoPermissionRequest`.</li><li>Cognigy 2026.17.0 and later — from the `ssoToken` and `permissionRequest` properties in `input.data.microsoftSsoConnections['<connection-name>']`, which allow [multiple SSO connections](#set-up-multiple-sso-connections-optional-accordion-title), and from legacy `input.data.microsoftSsoToken` and `input.data.microsoftSsoPermissionRequest`, which allow Flows configured previous to Cognigy 2026.16.0 to keep working.</li></ul> |
</Accordion>

## How to Set Up

### Setup on the Cognigy Side

<AccordionGroup>
  <Accordion title="Create an Azure Bot Services Endpoint">
    1. In the left-side menu of your Project, click **Deploy > Endpoints**.
    2. On the **Endpoints** page, click **+ New Endpoint**.
    3. In the **New Endpoint** section, do the following:
       1. Select the **Azure Bot Services** Endpoint type.
       2. Specify a unique name.
       3. Select a Flow from the list. Save changes.
    4. In the **Configuration Information** section, copy the Endpoint URL and save it for later use in Azure Bot Services.
  </Accordion>

  <Accordion title="Set up Multiple SSO Connections (Optional)">
    The Azure Bot Services Endpoint supports holding Microsoft SSO tokens for multiple OAuth connections simultaneously within a single conversation. Multiple OAuth connections allow your Flow to authenticate users with different Azure Bot OAuth connections, for example, Microsoft Graph and a custom API, in a single session.

    To target a different connection's token, use the **Overwrite Connection** toggle in the Get Token, Invalidate Token, and Token Status Nodes. Activating this toggle creates further keys in `input.data.microsoftSsoConnections`, in addition to the connection [configured in the Endpoint settings](#specific-endpoint-settings).

    Each connection's data is accessible from the Flow under `input.data.microsoftSsoConnections`, keyed by connection name:

    ```json theme={null}
    input.data.microsoftSsoConnections['<connection-name>'] = {
      ssoToken: "<jwt-access-token>",
      permissionRequest: "completed",
      audience: "<token-audience-url>"
    }
    ```

    The `audience` field holds the token's decoded `aud` claim. This claim identifies the API or resource that the token is valid for. The field is informational and may be `null` for opaque tokens, such as Microsoft Graph tokens.
  </Accordion>
</AccordionGroup>

### Setup on the Azure Bot Services Side

<AccordionGroup>
  <Accordion title="1. Create an App in Azure AD B2C">
    1. Go to the Microsoft Entra Admin Center and select **Azure AD B2C**.

    2. In the **Manage** section, select **App registrations**, then click **New registration**.

    3. In the **Supported account types** section, select one of the following options:

       <Tabs>
         <Tab title="Single Tenant">
           1. Under **Redirect URI**, pick **Single-page application (SPA)** and enter the URL `https://login.microsoftonline.com/<tenantID>/oauth2/v2.0/token`.
           2. Click **Register**, then copy **Tenant ID**.
           3. Go to **Entra ID > App registrations**, find your application that you created. Then copy and save the **Application (client) ID**.
           4. Open the app by clicking on its name. Go to **Manage > Authentication**. Enter the Cognigy Endpoint URL in the **Front-channel logout URL** field. In the **Supported account types** section, ensure that **Accounts in this organizational directory only** for a single tenant is enabled.
         </Tab>

         <Tab title="Multitenant">
           1. Under **Redirect URI**, select **Single-page application (SPA)** and enter the URL `https://token.botframework.com/.auth/web/redirect`.
           2. Go to **Entra ID > App registrations**, find the app that you created. Then copy and save the **Application (client) ID**.
           3. Open the app by clicking its name. Go to **Manage > Authentication**. Enter the Cognigy Endpoint URL in the **Front-channel logout URL** field. In the **Supported account types** section, ensure that **Accounts in any organizational directory** for a multitenant is enabled.
         </Tab>
       </Tabs>

    4. Go to **Manage > Certificates & secrets**. Click **+ New client secret**. Name your secret, set expiry, click **Save**. Copy the value and save it for later use.

    5. Go to **Manage > API permissions**, click **+ Add a permission**, then **Microsoft Graph > Delegated permissions**. Select `openid`, which is required for SSO. Select other permissions like `email` if needed.

    6. Click **Add permissions**. Click **Grant admin consent if required** — permissions get a green check.

    7. Go to **Manage > Expose an API**. In the **Application ID URI** field, enter `api://botid-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx`. Replace `x` with your saved Client ID. Save changes.

    8. Click **+ Add a scope**. Name the scope. Allow both admins and users to consent. Fill in other required fields. Save changes.

    9. Click **+ Add a client application** and add:
       * `5e3ce6c0-2b1f-4285-8d4b-75ee78787346` (MS Teams Web)
       * `1fec8e78-bce4-4aaf-ab1b-5451cc387264` (MS Teams Mobile & Desktop)

    10. Go to **Manage > Manifest**, update the following parameters to the values:
        * `AccessTokenAcceptedVersion` to `2`.
        * `Oauth2RequirePostResponse` to `true` (use only for Single Tenant).
        * `SignInUrl` – enter your Cognigy Endpoint URL. Save changes.
  </Accordion>

  <Accordion title="2. Create a Resource Group in Azure Bot Services">
    If you already have a resource group, you can skip this section.

    1. Click the **Create a resource** card. In the Marketplace search, enter **resource group**.
    2. In the search results, find the **Resource group** card. At the bottom of this card, click **Create > Resource group**.
    3. On the **Create a Resource group** page, select your subscription from the list. A trial or usage-based one will work.
    4. Specify a unique name for the group.
    5. Select a region to store the resources.
    6. Save changes by clicking **Review + create**, then **Create**.

    Your resource group will be created. You can check it by clicking the bell icon, then **Go to resource group**.
  </Accordion>

  <Accordion title="3. Create a Bot in Azure Bot Services">
    1. Click the **Create a resource** card.

    2. In the Marketplace search, enter **azure bot**. In the search results, find the **Azure bot** card. At the bottom of this card, click **Create > Azure bot**.

    3. On the **Create an Azure bot** page, fill in the following fields:
       * **Bot handle** - enter a display name for the Azure bot.
       * **Subscription and Resource group** - select the same values that you previously configured in the **Resource group** section.
       * **Data residency** — select **Global**.
       * **Pricing tier** — select **Free plan** by clicking **Change plan**.

    4. In the **Microsoft App ID** section, fill in the following fields according to the scope of your app:

       <Tabs>
         <Tab title="Single Tenant">
           1. Select **Single Tenant**.
           2. In the **Creation type** list, select **Use an existing app registration**.
           3. In the **App ID** field, enter the Client ID you previously copied and saved.
           4. In the **App tenant ID** field, enter the Client Secret you previously copied and saved.
         </Tab>

         <Tab title="Multitenant">
           1. Select **Multi Tenant**.
           2. In the **Creation type** list, select **Use an existing app registration**.
           3. In the **App ID** field, enter the Client ID you previously copied and saved.
         </Tab>
       </Tabs>

    5. Save changes by clicking **Review + create**, then **Create**. Wait for the deployment, which might take 1-2 minutes, then go to the resource.

    6. In the **Settings** section, select **Configuration**.

    7. The **Messaging endpoint** field must contain the previously copied Endpoint URL from the Azure Bot Services Endpoint. Activate **Enable Streaming Endpoint**. Click **Add OAuth Connection Settings**.

    8. On the **New Connection Setting** window, fill in the following fields:
       * **Name** — add a unique name. Copy and save this name for later use in Cognigy.
       * **Service provider** — select **Azure Active Directory v2**.
       * **Client ID** — enter the Client ID you previously copied and saved.
       * **Client Secret** — enter the Client Secret you previously copied and saved.
       * **Token Exchange URL** — enter the Cognigy Endpoint URL.
       * **Tenant ID** — enter the word `common` for a multitenant or the tenant ID for a single tenant.
       * **Scopes** — add the permissions your bot should have. Must be in a space-delimited list. For example, `email offline_access openid profile User.Read`.

    9. Save changes. At the bottom of the **Configuration** page, click **Apply**.
  </Accordion>

  <Accordion title="4. Add the Teams Channel to your Azure Bot">
    1. Add the **Teams** channel to your channel list.
    2. Go to the Azure Bot that you have just configured.
    3. In the **Settings** section, select **Channels**.
    4. In the **Available channels** section, select **Microsoft Teams**.
    5. Accept **Terms of Service** by clicking **Agree**. Click **Save**. The Teams channel appears in your channel list.

    <Note>The Teams channel might display errors. You can ignore these errors because each message is properly delivered to and processed by Cognigy.</Note>
  </Accordion>

  <Accordion title="5. Update the Azure Bot Services Endpoint">
    To update the Azure Bot Services Endpoint:

    1. In the Cognigy interface, go to **Deploy > Endpoints**. Select the Azure Bot Services Endpoint you previously created.
    2. In the **Bot Framework Settings** section, fill in the fields as follows:
       * **Microsoft App ID** — enter the Client ID.
       * **Microsoft App Password** — enter the Client Secret.
       * **SSO Connection Name** — enter the Connection Name.
  </Accordion>

  <Accordion title="6. Create a Microsoft Teams App">
    1. Go to the [Microsoft Teams Developer Portal](https://dev.teams.microsoft.com/) and select **Apps** from the left pane.
    2. Select **+ New app**, enter your app name and click **Add**. Go to **Configure > Basic Information**.
    3. Fill in the following fields:
       * **Short description** — a brief summary about your app.
       * **Long description** — more details about your app, such as its functions or restrictions.
       * **Developer or Company name** — for example, `Cognigy`.
       * **Website** — for example, [https://www.cognigy.com/](https://www.cognigy.com/).
       * **Privacy policy** — for example, [https://www.cognigy.com/privacy-policy](https://www.cognigy.com/privacy-policy).
       * **Terms of use** — for example, [https://www.cognigy.com/legal-notice](https://www.cognigy.com/legal-notice).
       * **Application (client) ID** — the Client ID you saved earlier.
       * *(Optional)* **Microsoft Partner Network ID** — only if you have admin consent permissions.
    4. Save changes and specify which features your bot will use:
       1. Go to **Configure > App features**.
       2. Select the **Bot** card.
       3. On the **Bot** page, configure the following settings:
          3.1 Select **Enter a bot ID** and enter your saved **Client ID**.
          3.2 Enable **Upload and download files** — required to prevent a bug that blocks attachments.
          3.3 Select **Personal scope**.
       4. Save changes.
  </Accordion>

  <Accordion title="(Optional) 6.1 Configure Single Sign-On">
    If you want to use Single Sign-On (SSO) with your bot, follow these steps:

    1. Go to **Configure > Single sign-on**.
    2. In the **Application ID URI** field, enter `api://botid-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx`. Replace `x` with your saved Client ID. Save changes.
    3. Still under **Single sign-on**, add these domains:
       * `token.botframework.com` — for token authentication.
       * `mydomain.com` — your Endpoint's hosting domain.
    4. To complete the SSO configuration, go to the [Cognigy Help Center: SSO and Graph API Nodes](https://support.cognigy.com/hc/en-us/articles/6422300559388-Azure-Bot-Services-Advanced-Deployment-with-SSO-Microsoft-Teams-SSO-and-Graph-API-Nodes#creating-an-sso-flow-0-0) guide.
  </Accordion>

  <Accordion title="7. Configure the App Package">
    1. Go to **Configure > App package**.
    2. At the top, click **Download App package**.
    3. Unzip the downloaded package, you'll get two images and a `manifest.json` file. Open the `manifest.json` in a code editor. Make these adjustments:
       * Around line 6, replace `packageName` with your unique package name — do not use `com.package.name`. Otherwise, SSO won't work.
       * Add this key at the top level:
         ```json theme={null}
         "permissions": ["identity", "messageTeamMembers"]
         ```
    4. Save the changes and zip the modified `manifest.json` file and the two images. The new zip file must contain only these files.
    5. Open your Teams client or go to [https://teams.microsoft.com](https://teams.microsoft.com).
    6. In the bottom-left corner, click **Apps**.
    7. Click **Upload a custom app** and select your new `.zip` package. Click **Add** — the bot should now appear as a contact you can chat with.
  </Accordion>
</AccordionGroup>

## Access Connection Data from Flows

To access OAuth token connections from your Flows, make sure you do the following:

* Add one **Get Token** Node per OAuth connection to your Flow. Each connection requires its own consent process.
* Each connection must be registered in Azure Bot Services under **Configuration > OAuth Connection Settings**.

You can use Code Nodes and CognigyScript in text fields to retrieve OAuth connection data. For example, if you configured a `microsoft-graph` connection and a `contoso-crm` connection in Azure Bot Services:

<Tabs>
  <Tab title="Code Node">
    ```javascript theme={null}
    // Access the token for the microsoft-graph connection
    const graphToken = input.data.microsoftSsoConnections['microsoft-graph'].ssoToken;

    // Access the approval status for the contoso-crm connection
    const crmStatus = input.data.microsoftSsoConnections['contoso-crm'].permissionRequest;

    // Access the token for the connection configured in the Endpoint settings (legacy field)
    const legacyToken = input.data.microsoftSsoToken;
    ```
  </Tab>

  <Tab title="Text Field">
    Use [CognigyScript](/ai/platform-features/cognigyscript) to access the same data in a text field:

    ```javascript theme={null}
    // Access the token for the microsoft-graph connection
    {{input.data.microsoftSsoConnections['microsoft-graph'].ssoToken}}

    // Access the approval status for the contoso-crm connection
    {{input.data.microsoftSsoConnections['contoso-crm'].permissionRequest}}

    // Access the token for the connection configured in the Endpoint settings (legacy field)
    {{input.data.microsoftSsoToken}}
    ```
  </Tab>
</Tabs>

## More Information

* [Cognigy Help Center: Microsoft Bot Framework Deployment](https://support.cognigy.com/hc/en-us/articles/360016183720-Microsoft-Bot-Framework-Deploy-an-Endpoint)
* [Cognigy Help Center: Microsoft Bot Framework Adaptive Cards](https://support.cognigy.com/hc/en-us/articles/360016184000-Microsoft-Bot-Framework-Using-Adaptive-Cards)


# Bandwidth
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/bandwidth



<a href="/release-notes/4.60"><Badge>Updated in 4.60</Badge></a>

<Frame>
  <img alt="Bandwidth Endpoint logo" />
</Frame>

The Bandwidth Endpoint enables you to connect your AI Agent to your contact center or other phone numbers using the [Bandwidth Voice API](https://www.bandwidth.com/voice/voice-api/). This Endpoint allows your customers to speak with your AI Agents instead of using a chat.

## Prerequisites

* Access to the [Bandwidth Voice API](https://www.bandwidth.com/).

## Generic Endpoint Settings

Learn about the generic Endpoint settings available on the following pages:

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [NLU Connectors](/ai/platform-features/nlu/external/nlu-connectors/overview)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Transformer Functions](/ai/for-developers/transformers/overview)
* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)
* [Inject and Notify](/ai/agents/deploy/endpoints/inject-and-notify)

## Specific Endpoint Settings

<Accordion title="Bandwidth Settings">
  | **Parameter** | **Description**                                                                                                                                                                                                                                                                                                                                                                                                                        |
  | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
  | Async Mode    | When async mode is enabled for the Bandwidth Endpoint, outputs are sent directly as they are processed in a Flow. This mode is useful when the Flow needs to send an HTTP request. <br /><br />When async mode is off, outputs are sent only after the entire Flow has finished processing the input.<br /><br />The [Inject and Notify](/ai/agents/deploy/endpoints/inject-and-notify) feature only works when async mode is enabled. |
</Accordion>

## How to Set Up

### Configuration on the Cognigy Side

<Accordion title="Create a Bandwidth Endpoint">
  1. In the left-side menu of the Cognigy Project, go to **Deploy > Endpoints**.
  2. On the **Endpoints** page, click **+ New Endpoint**.
  3. In the **New Endpoint** section, do the following:
     1. Select the **Bandwidth** Endpoint type.
     2. Add a unique name.
     3. Select a relevant Flow from the list.
  4. *(Optional)* Activate the **Async Mode** setting.
</Accordion>

## More Information

* [All Endpoints](/ai/agents/deploy/endpoint-reference/overview)


# Dialogflow
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/dialogflow

The Dialogflow Endpoint enables you to connect your AI Agent to the Dialogflow platform.

<Frame>
  <img alt="Dialogflow Endpoint logo" />
</Frame>

<Warning>
  As of Cognigy 2026.17.0, the Dialogflow Endpoint is deprecated. You can't create new Dialogflow Endpoints. The removal is planned for January 2028. Existing Dialogflow Endpoints continue to work until the removal date.
</Warning>

The Dialogflow Endpoint enables you to connect your AI Agent to the [Dialogflow platform](https://dialogflow.cloud.google.com/).

## Prerequisites

* Access to [the Dialogflow platform](https://dialogflow.cloud.google.com/).
* A fallback intent named **Default Fallback Intent** in Dialogflow with webhooks enabled.
* For Phone Gateway, a configured Dialogflow Phone Gateway and a connected phone number.

## Generic Endpoint Settings

Learn about the generic Endpoint settings on the following pages:

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [NLU Connectors](/ai/platform-features/nlu/external/nlu-connectors/overview)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)
* [Transformer Functions](/ai/for-developers/transformers/overview)

## Specific Endpoint Settings

<Accordion title="Dialogflow Settings">
  | **Parameter** | **Description**                                                                                                                                                    |
  | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
  | Reparse Slots | A toggle that lets you combine Dialogflow and Cognigy NLU. This feature allows slots found by Dialogflow to be passed to Cognigy and further processed by its NLU. |
</Accordion>

## How to Set Up

### On the Dialogflow Side

<AccordionGroup>
  <Accordion title="1. Access an Agent">
    1. Log in to Dialogflow, then select your agent.
    2. Ensure intents are configured.
    3. Create an intent named **Default Fallback Intent**.
    4. In the **Fulfillment** section of your Dialogflow agent, enter the Endpoint URL to enable webhooks.
  </Accordion>

  <Accordion title="2. Set up Integrations">
    1. Choose from supported integrations:
       * **Built-in Telephony**: AudioCodes, Avaya, Genesys, SignalWire, Voximplant.
       * **Dialogflow Built-in**: Messenger, Phone Gateway, Web Demo, Facebook, Google Assistant (legacy), and more.
       * **Google-contributed Open Source**: Kik, Skype, Twilio, Viber, and more.
    2. Then, follow Dialogflow's instructions to enable Google Actions Account linking.
  </Accordion>
</AccordionGroup>

### On the Cognigy Side

<AccordionGroup>
  <Accordion title="1. Create a Dialogflow Endpoint">
    1. In the left-side menu of your Agent, click **Deploy > Endpoints**.
    2. On the **Endpoints** page, click **+ New Endpoint**.
    3. In the **New Endpoint** section, do the following:
       1. Select the **Dialogflow** Endpoint type.
       2. Specify a unique name.
       3. Select a relevant Flow from the list. Save changes.
    4. Copy the URL from the **Endpoint URL** field.
  </Accordion>

  <Accordion title="2. Handle the `TELEPHONY_WELCOME` Message">
    1. In your Flow, add a Once Node or an Intent rule that triggers on `TELEPHONY_WELCOME` and greet the user, for example, `Welcome to our service! How can I help you today?`.
  </Accordion>
</AccordionGroup>

## More Information

* [Cognigy Help Center: Dialogflow Intro ](https://support.cognigy.com/hc/en-us/articles/360016153980)
* [Cognigy Help Center: Dialogflow Deployment](https://support.cognigy.com/hc/en-us/articles/360016154040)


# Facebook (Meta) Messenger
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/facebook-messenger



<a href="/release-notes/2026.16"><Badge>Updated in 2026.16.0</Badge></a>

<Frame>
  <img alt="Facebook Messenger Endpoint logo" />
</Frame>

The Facebook Messenger Endpoint enables you to connect AI Agents to the [Facebook Messenger](https://www.messenger.com/) platform.

## Prerequisites

* A Facebook account.
* Access to [Facebook (Meta) Messenger](https://www.messenger.com/).
* Access to the [Facebook (Meta) Developer Center](https://developers.facebook.com/).

## Generic Endpoint Settings

Learn about the generic Endpoint settings on the following pages:

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Transformer Functions](/ai/for-developers/transformers/overview)
* [NLU Connectors](/ai/platform-features/nlu/external/nlu-connectors/overview)
* [Session Management](/ai/agents/deploy/endpoints/session-management)
* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)
* [Inject and Notify](/ai/agents/deploy/endpoints/inject-and-notify)
* [Copilot](/agent-copilot/configure/copilot-section)

## Specific Endpoint Settings

<AccordionGroup>
  <Accordion title="Facebook Page Setup">
    | Parameter                  | Description                                                                                                      |
    | -------------------------- | ---------------------------------------------------------------------------------------------------------------- |
    | Facebook Page Access Token | A token generated by your Facebook App to authenticate the connection with Cognigy.                              |
    | App ID                     | The App ID of your Facebook App.                                                                                 |
    | App secret                 | The App Secret of your Facebook App. For additional security, it verifies that messages are sent from Messenger. |
  </Accordion>

  <Accordion title="Facebook Profiles">
    | Parameter               | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |
    | ----------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Request profile data    | Fetches data from the user's Facebook profile and stores in the `ci.data.profile` object. The `ci.data.profile` object will contain the following parameters: <ul><li>**Id** – the Facebook user's PSID, which is a unique user ID for each Facebook Page the user interact with.</li><li>**first\_name** – the user's first name.</li><li>**last\_name** – the user's last name.</li><li>**profile\_pic** – a URL linking to the user's profile picture.</li><li>**locale** – information about the user's locale.</li><li>**timezone** – the user's time zone.</li><li>**gender** – the user's gender.</li></ul><br />Note that the same data types are exposed in the profile object for every Facebook user. However, the actual values depend on the information available in their Facebook profile and on their privacy settings. |
    | Update contact profiles | Updates the Contact Profile in Cognigy with data from the user's Facebook profile. If the setting is enabled, the information a Facebook user has on their profile will be copied to their Contact Profile.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |
    | Merge contact profiles  | Each Facebook user has a user's PSID for each Facebook Page they interact with. This means that if you own two or more Facebook Pages, you can't immediately identify users across these pages because they will have a different ID for each page. However, if you enable this setting, Cognigy will fetch all of a user's PSIDs for your pages and merge these profiles into one. This setting allows you to easily identify a single user across all your pages and personalize the chat experience more effectively. To perform the necessary API calls to get all the PSIDs for a user, the same Business must own all the Facebook Pages.                                                                                                                                                                                          |
  </Accordion>

  <Accordion title="Messenger Settings">
    | Parameter                | Description                                                                                                                                                                               |
    | ------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Enable typing indicators | Enables a typing indicator while the bot is replying. You can configure whether typing indicators should be used and how many milliseconds of delay there should be between each message. |
    | Message Delay            | The amount of time to wait between sending bot replies, in milliseconds.                                                                                                                  |
  </Accordion>
</AccordionGroup>

## How to Set Up

<AccordionGroup>
  <Accordion title="1. Create a Facebook Page and a Facebook App">
    1. Log in to your Facebook account. Go to the **Pages** section.
    2. Click **Create New Page**, select a type (Business or Community), and provide a name and category.
    3. Go to the [Facebook (Meta) Developer Center](https://developers.facebook.com/).
    4. Click **Manage Apps**, then **Add a New App**.
    5. Fill in the required details to complete the app creation.
    6. In your app's product overview, locate **Messenger** and click **Set Up**.
  </Accordion>

  <Accordion title="2. Generate a Page Access Token">
    1. Go to **Messenger > Access Tokens**. Add the Facebook Page created earlier.
    2. Click **Generate Token**, acknowledge the warning, and **copy the token**.
  </Accordion>

  <Accordion title="3. Create a Messenger Endpoint">
    1. In the left-side menu of the Cognigy Project, go to **Deploy > Endpoints**.
    2. On the **Endpoints** page, click **+ New Endpoint**.
    3. In the **New Endpoint** section, do the following:
       1. Select the **Messenger** Endpoint type.
       2. Add a unique name.
       3. Select a relevant Flow from the list.
    4. Open the **Facebook Page Setup** section. Paste the **Page Access Token** from the Facebook App.
    5. *(Optional)* Enter your **App ID** and **App Secret** for additional security.
    6. Copy the Endpoint URL for later use.
  </Accordion>

  <Accordion title="4. Configure Webhooks">
    1. Go to the Facebook app. Under **Messenger > Webhooks**, click **Add Callback URL**.
    2. Paste the Endpoint URL as the callback URL.
    3. Enter the **Verify Token** from Cognigy:
       * For `https://trial.cognigy.ai`, use the token `6a1ed94ef26462d1a529ab80afabe135e5ac7cc8b414ef762e1208f3339f887196d3433c996c0f30ea15295fbd86293cd7516abaa9b9e78d641ebb514eea32b4`.
       * For other environments, contact Cognigy technical support for more information.
    4. Click **Verify and Save**.
    5. Add **Subscription Fields**: select `messages`, `messaging_postbacks`, and optionally `messaging_referrals`. Save the settings.
  </Accordion>
</AccordionGroup>

## More Information

* [Help Center: Facebook (Meta) Messenger Intro](https://support.cognigy.com/hc/en-us/articles/360015980560-Facebook-Messenger-Intro)
* [Help Center: Facebook (Meta) Messenger Deploy an Endpoint](https://support.cognigy.com/hc/en-us/articles/360015980660-Facebook-Messenger-Deploy-an-Endpoint)
* [Help Center: Use Facebook (Meta) Handover](https://support.cognigy.com/hc/en-us/articles/360016021919-Facebook-Messenger-Use-Facebook-Handover)
* [Help Center: Facebook (Meta) Messenger Facebook Tools](https://support.cognigy.com/hc/en-us/articles/360015980780-Facebook-Messenger-Facebook-Tools)


# Genesys
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/genesys



<a href="/release-notes/4.64"><Badge>Updated in 4.64</Badge></a>

<Frame>
  <img alt="Genesys Endpoint logo" />

  <img alt="Genesys Endpoint logo" />
</Frame>

The Genesys Endpoint enables you to connect your AI Agent to the [Genesys Cloud CX platform](https://apps.mypurecloud.de/) using a [Genesys Bot Connector](https://help.mypurecloud.com/articles/about-genesys-bot-connector/).

## Prerequisites

* Access to [Genesys Cloud](https://apps.mypurecloud.de/).
* Activate the Genesys Endpoint on the Cognigy side:
  * If you have a SaaS installation, contact Cognigy technical support.
  * If you have an on-premises installation, specify the following feature flags:
    * `FEATURE_USE_GENESYS_BOT_CONNECTOR_ENDPOINT_WHITELIST="organization1,organization2"`. To enable the Genesys Endpoint for all organizations in your installation, use `*` instead of a list of organizations.
    * `FEATURE_ENABLE_AGENT_ASSIST_WORKSPACE_GENESYS_CREDENTIALS_WHITELIST="organization1,organization2"`. Add this feature flag to use Genesys [Notifications APIs](https://developer.genesys.cloud/notificationsalerts/notifications/notifications-apis) with Agent Copilot. To enable this feature for all organizations within the Genesys Endpoint in your installation, use `*` instead of a list of organizations.

## Generic Endpoint Settings

Learn about the generic Endpoint settings available for this Endpoint on the following pages:

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)
* [Transformer Functions](/ai/for-developers/transformers/overview)
* [Copilot](/agent-copilot/configure/copilot-section)

## Specific Endpoint Settings

<Accordion title="Genesys Bot Connector Setup">
  | Parameter    | Description                                                                                                                                                                                                     |
  | ------------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
  | Verify Token | A verification token for connecting the Cognigy Genesys Endpoint to the Genesys Bot Connector. This token needs to be generated on the Cognigy side and can consist of letters, numbers, and punctuation marks. |
</Accordion>

## How to Set Up

### Setup on the Cognigy Side

<Accordion title="1. Configure Genesys Endpoint">
  1. In the left-side menu of the Project, go to **Deploy > Endpoints**.
  2. On the **Endpoints** page, click **+ New Endpoint**.
  3. In the **New Endpoint** section, do the following:
     1. Select the **Genesys** Endpoint type.
     2. Add a unique name.
     3. Select a relevant Flow from the list.
  4. In the **Configuration Information** section, copy the Endpoint URL and save it for later use in Genesys.
  5. Activate the **Enable Endpoint** setting.
  6. In the **Genesys Bot Connector Setup** section, in the **Verify Token** field, create a verification token for connecting the Cognigy Genesys Endpoint to the Genesys Bot Connector. The token can consist of letters, numbers, and punctuation marks. Copy this token for later use in Genesys. Save changes.
</Accordion>

### Setup on the Genesys Cloud CX Side

<AccordionGroup>
  <Accordion title="1. Create Genesys Bot Connector Integration">
    1. Open the Genesys Cloud interface, click **Menu** in the upper-left corner and go to **IT and Integrations > Integrations**.
    2. Click **Add Integration** in the upper-right corner.
    3. On the **Genesys Bot Connector** card, click **Install**. Assign a unique name to your integration.
    4. On the **Configuration** tab, navigate to the **Properties** section.
    5. In the **Bot Connector Handle Utterance URI** field, enter the Endpoint URL value you saved earlier from the Cognigy setup.
    6. In the **Credentials** section, click **Configure** in the **Genesys Bot Connector connection secret** card.
    7. Click **+ Add Credential Field** and enter the following credentials:
       * **newKey Key** — use `verify-token` as the name.
       * **newKey** — enter the token created earlier in the **Verify Token** field on the Cognigy side.
    8. Click **Confirm**, then **Save**.
    9. On the **Details** tab, click **Copy Integration ID** and save this ID for later use.

    Your integration will be listed among integrations.
  </Accordion>

  <Accordion title="2. Add Bot for Bot Connector Integration">
    1. Open the [API Explorer interface](https://developer.genesys.cloud/devapps/api-explorer) in the Genesys Developer Tools. Make sure you are logged in by using the **Account Selection** list in the upper-right corner.
    2. In the **API Operation Filter**, enter `botconnector`, then select **PUT** in the **HTTP Verb Filter**.
    3. Click the `PUT /api/v2/integrations/botconnector/{integrationId}` operation, then click the **Reading Mode** toggle to deactivate it.
    4. Fill in the **integrationId** field under **Path Parameters** with the integration ID that you copied and saved earlier.
    5. In the **Request Body** editor under **Query Parameters**, enter the following JSON:

       ```json theme={null}
       {
          "chatBots": [
             {
                "id": "<your UUID>",
                "name": "Cognigy Bot",
                "versions": [
                   {
                      "version": "1",
                      "supportedLanguages": ["en-us"],
                      "intents": [
                         {
                            "name": "Success",
                            "slots": {
                               "processedTimestamp": {
                                  "name": "processedTimestamp",
                                  "type": "integer"
                               }
                            }
                         }
                      ]
                   }
                ]
             }
          ]
       }   
       ```

    The `id` field must contain a UUID, which you can create using an online [UUID generator](https://www.uuidgenerator.net/).<br />

    6. Click **Execute Request**.
  </Accordion>

  <Accordion title="3. Build Inbound Message Flow">
    To create a digital bot flow in Architect, configure the inbound message flow:

    1. Open the Genesys Cloud interface, click **Menu** in the upper-left corner and go to **Orchestration > Architect**.

    2. Hover over the <img alt="menu" /> icon on the **Flows** tab and select **Inbound Message**.

    3. Click **+ Add** in the upper-left corner. The **Create 'Inbound Message Flow'** dialog box opens. In the **Name** field, enter a unique name for the inbound message flow.

    4. From the **Division** list, select the division in which to place the flow.

    5. Click **Create Flow**. The flow's configuration page opens. To configure a flow, click **Edit**.

    6. Go to the **Toolbox** section, enter `Call Bot Connector` in the search field and drag the action to the messaging flow editor.

    7. In the right-side **Call Bot Connector** panel, fill in the following fields:
       * **Bot Integration** — select the integration you created earlier.
       * **Bot Name** — select the bot you created earlier.
       * **Bot Version** — select the bot version you created earlier.

    8. In the **Toolbox** section, enter `Send Response` in the search field. Then, drag the action to the messaging flow editor and drop it below the **Success** action. Drag another **Send Response** action below the **Failure** action.

    9. In the **Message Body** of the **Send Response** action below **Success**, enter `success message` and select **Literal** from the list next to the field. In the **Send Response** action below **Failure**, enter `fail message` and select **Literal** from the list next to the field.

    10. Below the **Send Response** actions, place the [Transfer to ACD](https://help.mypurecloud.com/articles/transfer-acd-action/) action to transfer an interaction to a queueing system.

    11. In the **Queue** field of the **Transfer to ACD** action, select the queue to which you want to transfer the interaction.

        <Frame>
          <img alt="Genesys inbound message flow configuration example" />
        </Frame>

    12. In the upper-left corner, click **Save**, then **Publish**.

    After creating your inbound message flow, you will see this flow in the architect list.

    To learn more about designing the flow, see [Configure Inbound Message Flow](https://help.mypurecloud.com/articles/inbound-message-flows/) settings.
  </Accordion>

  <Accordion title="4. Configure Messenger">
    To create a new version of a messenger configuration and a messenger deployment, follow these steps:

    <Steps>
      <Step title="Set up Messenger">
        1. Go to **Digital and Telephony > Message > Messenger Configurations** and click **+ New Setup**.
        2. In the **Channel Setup** step, configure the following:
           * **Name** — enter a unique configuration name reflecting your bot integration.
           * **Choose a Digital Channel you want to set up** — select **Messaging**. Click **Next**.
        3. *(Optional)* In the **Appearance** step, configure the appearance of the chat widget. Click **Next**.
        4. In the **Behavior** step, configure the following:
           * *(Optional)* In the **Messaging Configuration** section, set the behavior of the chat widget.
           * In the **Language Support** section, set how the chat widget handles different languages:
             * **Select your Supported Languages** — select UI languages and customize text labels. Browser language determines the user's language.
             * **Select Default Language** — select the default language for cases when the browser language isn't detected or isn't in the list. Click **Next**.
        5. In the **Review Step**, review your messenger configuration, then click **Publish Configuration**.
      </Step>

      <Step title="Deploy Messenger">
        1. In the left-side menu, click **Messenger Deployments** and click **+ New Deployment**.
        2. In the **Name** field, enter a unique deployment name that corresponds to your messenger configuration.
        3. Click the **Select your Configuration** button, select the message configuration you created, the configuration version, and click **Change Assignment**.
        4. In the **Restrict domain access** section, select **Allow all domains** or specify trusted domains where the Messenger can run.
        5. From the **Select your Architect Flow** list, select a published inbound message flow to trigger the bot behavior or connect users with agents.
        6. Click **Publish Deployment**. The **Deployment Options** section is displayed.
        7. In the **Deploy your snippet** section, click **Copy to Clipboard** to copy the embedding code snippet. You will use this snippet to embed the Messenger in your website. After deploying the snippet, any future changes to the configuration and deployment will be applied automatically.
        8. To test the AI Agent, visit [CodeSandbox](https://codesandbox.io/) and select a default **HTML** project. Paste the copied code snippet into the `<body>` element of the HTML document.
      </Step>
    </Steps>

    To check the AI Agent's performance, access the installation that your server administrator has deployed.
  </Accordion>
</AccordionGroup>

### (Optional) Activate Agent Copilot

If you want to use [Agent Copilot](/agent-copilot/overview) and [Notifications APIs](https://developer.genesys.cloud/notificationsalerts/notifications/notifications-apis) within the Genesys integration, create credentials on the Genesys Cloud CX side and configure the **Copilot** settings in the Endpoint.

<AccordionGroup>
  <Accordion title="1. Create Credentials">
    To configure credentials, follow these steps:

    <Steps>
      <Step title="Set up Client">
        1. In the left-side menu, select **IT and Integrations > OAuth**, then click **+ Add Client**.
        2. On the **Add New Client** page, configure the following:
           * **App Name** — enter a unique name for the client.
           * **Grant Types** — select **Client Credentials**. Click **Next**.
        3. In the **Roles** list, activate the corresponding role for the client. Click **Next**.
        4. In the **Token Duration in seconds** field, enter the token expiration time. Click **Next**.
        5. Click **Generate New Client Secret**, then **Confirm**.
        6. Copy the Client ID and Client Secret, save them for future use, and click **Finish**. Confirm that you copied the Client ID and Client Secret in the dialog box.
      </Step>

      <Step title="Authorize Client">
        1. In the left-side menu, select **Authorized Applications**.
        2. In the upper-right corner, click **+ Authorize a Client**.
        3. In the **Authorize Client** window, enter the Client ID that you copied previously and click **Authorize Client**.
        4. *(Optional)* In the **Users that can use this application** section, select which user roles can access the application.
        5. In the **Scope** section, select the scopes of the application. Click **Authorize**.
      </Step>
    </Steps>
  </Accordion>

  <Accordion title="2. Configure Settings for Agent Copilot">
    To configure the Agent Copilot workspace settings, follow these steps:

    1. In the **Genesys Endpoint settings**, navigate to the **Copilot** section.
    2. Configure Agent Copilot by creating an [Agent Copilot config](https://api-trial.cognigy.ai/openapi#post-/v2.0/agentassistconfigs) and adding a Flow. For more information, refer to the [Agent Copilot documentation](/agent-copilot/overview).
    3. Set up Genesys Cloud Credentials:
       * **Connection Name** — create a unique name for your connection.
       * **oAuth2Url** — enter the login URL for Genesys Cloud customers, which varies by region, such as `mypurecloud.de` for Germany. You can use the base domain like `mypurecloud.com` or `cac1.pure.cloud`, omitting the `apps.` or `login.` part. For more information on Genesys Cloud regions, refer to [AWS regions for deployment](https://help.mypurecloud.com/articles/aws-regions-for-genesys-cloud-deployment/).
       * **oAuth2ClientId** — enter the Client ID that you generated on the Genesys Cloud CX side.
       * **oAuth2ClientSecret** — enter the Client Secret that you generated on the Genesys Cloud CX side.
       * **oAuth2Scope** — enter `*`. This system field is only necessary for authorizing the application.
    4. Click **Create**, then **Save**.
  </Accordion>
</AccordionGroup>

## What's Next?

Now you can configure Say, Question, and Optional Question Nodes by selecting the [Genesys](/ai/agents/develop/node-reference/basic/say) channel in the Node Editor settings. It's important to create messages that are compatible with the Genesys Endpoint, as the Genesys channel supports a [limited set of output types](/ai/agents/deploy/endpoints/content-conversion).


# Intercom
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/intercom



<Frame>
  <img alt="Intercom Endpoint logo" />
</Frame>

The Intercom Endpoint enables you to connect your AI Agent to the [Intercom platform](https://app.intercom.com/).

## Prerequisites

* Access to the [Intercom platform](https://app.intercom.com/).

## Generic Endpoint Settings

Learn about the generic Endpoint settings on the following pages:

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [NLU Connectors](/ai/platform-features/nlu/external/nlu-connectors/overview)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)
* [Transformer Functions](/ai/for-developers/transformers/overview)

## Specific Endpoint Settings

<Accordion title="Intercom Settings">
  | **Parameter** | **Description**                                                                                                                  |
  | ------------- | -------------------------------------------------------------------------------------------------------------------------------- |
  | Access Token  | The token generated by your Intercom app that has the necessary permissions to communicate with Intercom.                        |
  | Bot User ID   | The unique identifier of the Intercom teammate that will be used as your Intercom bot.                                           |
  | Hub Secret    | The client secret provided by Intercom. The secret allows Cognigy to authenticate and communicate with your Intercom bot.        |
  | Message Delay | A delay in milliseconds for each letter in a message. This parameter allows you to control the typing speed of the Intercom bot. |
</Accordion>

## How to Set Up

<AccordionGroup>
  <Accordion title="1. Create an Intercom Endpoint">
    1. In the left-side menu of your Project, click **Deploy > Endpoints**.
    2. On the **Endpoints** page, click **+ New Endpoint**.
    3. In the **New Endpoint** section, do the following:
       1. Select the **Intercom** Endpoint type.
       2. Specify a unique name.
       3. Select a Flow from the list. Save changes.
    4. Go to the **Configuration Information** section and copy the URL from the **Endpoint URL** field.
  </Accordion>

  <Accordion title="2. Create an App">
    1. Log in to your [Intercom account](https://app.intercom.com/).
    2. Go to **Settings > Home > Integrations**, then click **Developer Hub**.
    3. Click the app you want to use with Cognigy or create a new one.
    4. In the app, go to **Configure > Webhooks**. Paste the copied Endpoint URL into the **Endpoint URL** field. Click **Send a test request**.
    5. Go to **Configure > Basic information**. Copy the client secret from the **Client secret** field.
    6. Go to **Configure > Authentication**. Make sure `Read Conversations` and `Write Conversations` are enabled. Copy the access token from the **Access token** field.
  </Accordion>

  <Accordion title="3. Create a Teammate">
    1. In the Intercom interface, go to **Settings > Home > Workspace**, click **+ Add teammate**, then **+ Add new teammates**.
    2. Enter an email address for your bot, for example, `bot@yourcompany.com`. Click **Continue and set permissions**.
    3. Set the role and permissions. Grant limited permissions to your bot if you want to restrict access to inboxes or admin settings. Click **Send Invite**.
    4. Accept the invitation via email. Open the teammate's profile and copy the Admin ID from the URL. The Admin ID is the 7-digit number in the URL: `https://app.intercom.io/a/apps/<app-id>/admins/<admin-id>`.
  </Accordion>

  <Accordion title="4. Complete the Endpoint Configuration">
    1. In the Cognigy interface, go to the **Intercom Endpoint** you created.
    2. In the **Access Token** field, paste the access token you copied from Intercom.
    3. In the **Bot User ID** field, paste the Admin ID of the Intercom teammate you created.
    4. In the **Hub Secret** field, paste the client secret you copied from Intercom.
    5. *(Optional)* Set a **Message Delay** to control the typing speed of your bot. Save changes.
  </Accordion>
</AccordionGroup>

## More Information

* [Cognigy Help Center: Intercom Deployment](https://support.cognigy.com/hc/en-us/articles/360016183040-Intercom-Deploy-an-Endpoint)
* [Cognigy Help Center: Intercom Managing a Bot](https://support.cognigy.com/hc/en-us/articles/360016183200-Intercom-Managing-the-Intercom-Bot)


# LINE
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/line-messenger



<a href="/release-notes/2026.16"><Badge>Updated in 2026.16.0</Badge></a>

<Frame>
  <img alt="LINE Messenger Endpoint logo" />
</Frame>

The LINE Endpoint enables you to connect your AI Agent to the [LINE platform](https://www.line.me/).

## Prerequisites

* Access to [LINE Developer Console](https://developers.line.biz/console/).

## Restrictions

* The LINE Messaging API requires you to follow rules that might block some messages. Check your Node settings with the [Line Messaging API requirements](https://support.cognigy.com/hc/en-us/articles/5721825020444) to ensure all messages are delivered.

## Generic Endpoint Settings

Learn about the generic Endpoint settings on the following pages:

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [NLU Connectors](/ai/platform-features/nlu/external/nlu-connectors/overview)
* [Session Management](/ai/agents/deploy/endpoints/session-management)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)
* [Transformer Functions](/ai/for-developers/transformers/overview)
* [Inject and Notify](/ai/agents/deploy/endpoints/inject-and-notify)
* [Copilot](/agent-copilot/configure/copilot-section)

## Specific Endpoint Settings

<Accordion title="LINE Setup">
  | **Parameter**        | **Description**                                                                                                                   |
  | -------------------- | --------------------------------------------------------------------------------------------------------------------------------- |
  | Channel Access Token | The token obtained from your Line Developer account for each deployed channel. It is linked to the Channel Secret.                |
  | Channel Secret       | The secret associated with the Channel Access Token and also obtained from your Line Developer account for each deployed channel. |
</Accordion>

## How to Set Up

<AccordionGroup>
  <Accordion title="1. Create a LINE Messaging API Channel">
    1. Log in to the [LINE Developers Console](https://developers.line.biz/console/).
    2. Create a new provider (if needed).
    3. Create a new **Messaging API** channel under that provider.
    4. Copy the values from the **Channel Secret** and **Channel Access Token** fields for later use.
  </Accordion>

  <Accordion title="2. Create a LINE Endpoint">
    1. In the left-side menu of your Project, click **Deploy > Endpoints**.
    2. On the **Endpoints** page, click **+ New Endpoint**.
    3. In the **New Endpoint** section, do the following:
       1. Select the **Line** Endpoint type.
       2. Specify a unique name.
       3. Select a Flow from the list. Save changes.
    4. Go to the **Configuration Information** section and copy the URL from the **Endpoint URL** field.
  </Accordion>

  <Accordion title="3. Configure the Webhook in LINE">
    1. Go back to the LINE Developers Console.
    2. Open your Messaging API channel settings.
    3. Find the **Webhook URL** field and paste the Cognigy Endpoint URL you copied.
    4. Enable the webhook by toggling the **Use webhook** switch. Test the webhook.
  </Accordion>

  <Accordion title="4. Complete the Endpoint Configuration">
    1. In the Cognigy interface, open the LINE Endpoint configuration.
    2. Fill in the **Channel Secret** and **Channel Access Token** fields. Save changes.
  </Accordion>
</AccordionGroup>

## More Information

* [Cognigy Help Center: LINE Messenger Deployment](https://support.cognigy.com/hc/en-us/articles/360016222819)


# MCP Server
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/mcp-server



<a href="/release-notes/2026.17"><Badge>Updated in 2026.17 (experimental)</Badge></a>

<Frame>
  <img alt="MCP Server Endpoint logo" />
</Frame>

<Warning>
  The MCP Server Endpoint is experimental and isn't recommended for production use. This Endpoint may change or be removed in future releases. Use it only in development or staging environments.
</Warning>

The MCP Server Endpoint allows external AI applications to use tools from an AI Agent in Cognigy via the Model Context Protocol (MCP). Compatible clients, such as desktop AI assistants or custom MCP applications, can discover and execute these tools.

Unlike standard Endpoints, the MCP Server Endpoint isn't designed for conversational use. It doesn't use NLU, process user messages, or perform real-time translation. Instead, it works as a tool server: clients discover the available tools and invoke them using structured requests. Cognigy then executes the corresponding Flow logic and returns the results through the MCP protocol.

## Key Benefits

* **Tool Access through MCP**. Tools configured under an [AI Agent](/ai/agents/overview) or [LLM Prompt](/ai/agents/develop/node-reference/service/llm-prompt) Node are available to MCP-compatible clients.
* **Direct Flow Execution**. When a tool is called, Cognigy executes the related Flow and returns the result through MCP.
* **Standard Integration**. MCP provides a consistent interface for connecting external applications without custom APIs.

## Restrictions

* The following tools are available via MCP:
  * Regular Tool
  * Knowledge Search
  * Send Email
  * Execute Workflow
  * Handover to AI Agent
* Tool calls must finish within the configured timeout. The default timeout is 30 seconds. If you have an on-premises installation, you can change the default value via the `MCP_SERVER_TOOL_CALL_TIMEOUT` variable.
* This Endpoint supports OAuth 2.0 and no-authentication modes. User-level authentication isn't supported.

## Prerequisites

* A Flow that contains either an AI Agent or LLM Prompt Node.
* The AI Agent Node or LLM Prompt Node in the Flow must include at least one Tool Node.
  * Each Tool Node has a tool ID and a description.

## Generic Endpoint Settings

* [Advanced: Mocking](/ai/agents/test/mocking)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Transformer Functions](/ai/for-developers/transformers/overview)

## Specific Endpoint Settings

<AccordionGroup>
  <Accordion title="MCP Explorer">
    | **Parameter** | **Description**                                                                                                                                                                                                                                                                                                                                    |
    | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | MCP Explorer  | A link to test your tools and view connection details. It displays all exposed tools, including their names, descriptions, and parameters, and allows validation and testing of the configuration. The MCP Explorer also provides ready-to-use configuration snippets for Claude Desktop and Cursor, with a server key based on the Endpoint name. |
  </Accordion>

  <Accordion title="AI Agent Selection">
    This section allows you to select the Flow and AI Agent Node that will expose its tools via the Model Context Protocol (MCP). External AI clients can discover and call these tools through this endpoint.

    | **Parameter** | **Description**                                                                                                                                             |
    | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Flow          | The Flow that contains the selected AI Agent Node.                                                                                                          |
    | Node          | The AI Agent or LLM Prompt Node whose child Tool Nodes are exposed via MCP. Only tools defined as children of the selected Node are exposed to MCP clients. |
    | Exposed Tools | Displays all exposed tools.                                                                                                                                 |
  </Accordion>

  <Accordion title="Configuration Information">
    This section provides the MCP Server URL and a list of all exposed tools, including their names, descriptions, and parameters. This information is required to configure MCP clients.

    | **Parameter**     | **Description**                                                                                                                                                         |
    | ----------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | MCP Server URL    | The dedicated MCP URL in the format `https://endpoint.cognigy.ai/mcp/v1/{token}/mcp`. This is the single entry point for all MCP traffic and must be kept confidential. |
    | Enable MCP Server | Activates the MCP Server functionality for the Endpoint.                                                                                                                |
  </Accordion>
</AccordionGroup>

## How to Set Up

The diagram shows the complete MCP workflow. First, make sure your Flow is set up correctly with an AI Agent Node and a Tool Node, according to the prerequisites. Then, the MCP client connects to the MCP Server Endpoint and calls a tool. This triggers the Flow execution. The result is returned through a session-based connection.

```mermaid theme={null}
sequenceDiagram
    participant User
    participant Cognigy as Cognigy Platform
    participant MCP as MCP Endpoint
    participant Client as MCP Client

    %% Prerequisites
    rect rgb(240, 240, 240)
        note right of User: Prerequisites
        User->>Cognigy: Create a Flow
        User->>Cognigy: Add AI Agent or LLM Prompt Node
        User->>Cognigy: Add child Tool Node
    end

    %% Configuration
    User->>Cognigy: Configure Endpoint as MCP Server
    User->>Client: Provide MCP Server URL
    Client->>MCP: Connect using URL

    %% Runtime
    Client->>MCP: tools/call (with mcp-session-id)
    MCP->>Cognigy: Identify AI Agent Node
    Cognigy->>Cognigy: Execute Flow (Tool Node)
    Cognigy-->>MCP: Return Tool Result
    MCP-->>Client: Structured MCP Response
    Client-->>User: Return Final Response
```

### Setup on the Cognigy Side

<Accordion title="Configure an MCP Server Endpoint">
  1. In the left-side menu of your Project, click **Deploy > Endpoints**.
  2. On the **Endpoints** page, click **+ New Endpoint**.
  3. In the **New Endpoint** section, do the following:
     1. Select the **MCP Server** Endpoint type.
     2. Specify a unique name.
     3. Save changes.
  4. In **AI Agent Selection**, select the Flow and the AI Agent Node or LLM Prompt whose tools you want to expose via MCP. Save changes.
  5. In the **Configuration Information** section, copy the URL in the **MCP Server URL** field.
  6. Activate the **Enable MCP Server** toggle to start the server and allow clients to connect.
  7. Go to the **Authentication** section. From the **Authentication Type** list, select the authentication method for the MCP server:
     * **None** — no authentication is required to access the MCP server. Use this option for internal or trusted environments only. The option is selected by default.
     * **OAuth 2.0** — activate token-based authentication using OAuth 2.0. Use this option to verify that only authorized clients can access the MCP server. Configure the following fields:
       * **JWKS URI** — the URL of the JSON Web Key Set (JWKS) endpoint used to verify token signatures. You can find this URL in your identity provider's OpenID Connect discovery document, typically at `/.well-known/openid-configuration`. For example, `https://your-idp.example.com/.well-known/jwks.json`.
       * **Issuer** — the expected token issuer. This value must match the `iss` claim in the JWT. You can find this value in your identity provider's app settings, often labeled **Issuer URL** or **Authority**. For example, `https://your-idp.example.com`.
       * **Audience** — the intended recipient of the token. This value must match the `aud` claim in the JWT. You can find this value when registering your application with the identity provider such as the client ID or a custom API identifier. For example, `cognigy-mcp-client`.
       * **Required Scopes** — the OAuth scopes that the token must include to be granted access. Define these scopes in your identity provider when configuring the application or API permissions. For example, `cognigy:mcp`.
       * **Allowed Subject IDs** — this field is optional. A comma-separated list of subject identifiers (`sub` claim) permitted to access the server. You can find the subject ID in the JWT payload or in your identity provider's user or service account settings. For example, `user-123, service-account-456`. If the field is empty, all subjects are allowed.
       * **Key Refresh Interval (minutes)** — this field is optional. The frequency at which the JWKS public keys are refreshed. Increase this value to reduce the number of requests to your identity provider. Must be a value between `1` and `1440`, for example, `120`. If you leave this parameter empty, the server uses the default interval of 60 minutes.
</Accordion>

### Setup on the Third-Party Provider Side

<Accordion title="Configure an MCP Client">
  The configuration depends on the MCP client you use.

  <Tabs>
    <Tab title="Direct HTTP Integration">
      The MCP Server Endpoint supports the MCP Streamable HTTP transport. To integrate directly with the Endpoint, follow these steps:

      1. Send an `initialize` request.

         ```bash theme={null}
         curl -i -X POST https://endpoint-dev.cognigy.ai/mcp/v1/{token}/mcp \
           -H "Content-Type: application/json" \
           -d '{
             "jsonrpc": "2.0",
             "id": 1,
             "method": "initialize",
             "params": {
               "protocolVersion": "2024-11-05",
               "capabilities": {},
               "clientInfo": {
                 "name": "test",
                 "version": "1.0"
               }
             }
           }'
         ```

         Replace `{token}` with the token from the MCP Server URL.

      2. The response will include `mcp-session-id: <session-id>`. Store the session ID for use in subsequent requests.

      3. Request a list of available tools via `tools/list`:

         ```bash theme={null}
         curl -X POST https://endpoint-dev.cognigy.ai/mcp/v1/{token}/mcp \
           -H "Content-Type: application/json" \
           -H "mcp-session-id: <session-id>" \
           -d '{
             "jsonrpc": "2.0",
             "id": 2,
             "method": "tools/list"
           }'
         ```

         The request returns the list of available tools with their parameters and descriptions. This allows the client to understand how to call each tool correctly.

         For example, the response may include tools like this:

         ```json theme={null}
         {
           "jsonrpc": "2.0",
           "id": 2,
           "result": {
             "tools": [
               {
                 "name": "mcp",
                 "description": "",
                 "inputSchema": {
                   "type": "object",
                   "properties": {},
                   "required": []
                 }
               },
               {
                 "name": "SearchCognigyDocumentation",
                 "description": "",
                 "inputSchema": {
                   "type": "object",
                   "properties": {},
                   "required": []
                 }
               },
               {
                 "name": "unlock_account",
                 "description": "This tool unlocks a locked user account.",
                 "inputSchema": {
                   "type": "object",
                   "properties": {
                     "email": {
                       "type": "string",
                       "description": "User's login email for their account."
                     }
                   },
                   "required": ["email"]
                 }
               }
             ]
           }
         }
         ```

      4. Call tools using the `tools/call` method, providing the tool name and required arguments.
         For example, to call the `unlock_account` tool with the required `email` argument, send the following request:

         ```bash theme={null}
         curl -X POST https://endpoint-dev.cognigy.ai/mcp/v1/{token}/mcp \
           -H "Content-Type: application/json" \
           -H "mcp-session-id: <session_id>" \
           -d '{
             "jsonrpc": "2.0",
             "id": 2,
             "method": "tools/call",
             "params": {
               "name": "unlock_account",
               "arguments": {
                 "email": "user@example.com"
               }
             }
           }'
         ```

         After a successful tool call, the MCP server returns the result in a structured MCP response. The response contains the output of the executed tool inside the result field. For example, if the tool completes successfully, it may return a text message such as `Resolved` in the content array.

         ```json theme={null}
         {
           "jsonrpc": "2.0",
           "id": 2,
           "result": {
             "content": [
               {
                 "type": "text",
                 "text": "Resolved"
               }
             ]
           }
         }
         ```
    </Tab>

    <Tab title="Claude Desktop">
      To configure the MCP client in Claude Desktop, follow these steps:

      1. In the Claude Desktop application, go to **Settings > Developer > Local MCP servers**. Click **Edit Config**.
      2. Add the following configuration to `claude_desktop_config.json`:

      ```json theme={null}
      {
        "mcpServers": {
          "my-endpoint-name": {
            "command": "npx",
            "args": [
              "mcp-remote",
              "https://endpoint.cognigy.ai/mcp/v1/{token}/mcp"
            ]
          }
        }
      }
      ```

      Replace `{token}` with the token from the MCP Server URL and `my-endpoint-name` with a name for your Endpoint. The MCP Explorer automatically generates a key based on the Endpoint name.

      3. Close and reopen Claude Desktop. It automatically launches `mcp-remote`, connects to the MCP server, initializes the session, and loads the available tools.
      4. In Claude Desktop, open the chat and ask the AI to perform a task that requires a tool. Claude Desktop will automatically call the appropriate MCP tool when needed.
    </Tab>

    <Tab title="Cursor">
      To configure the MCP client in Cursor, follow these steps:

      1. In the root directory of your Cursor project, create the `.cursor/mcp.json` file. If the `.cursor` folder doesn't exist, create it manually.

      2. Add the following configuration to the `mcp.json` file:

         ```json theme={null}
         {
           "mcpServers": {
             "my-endpoint-name": {
               "command": "npx",
               "args": [
                 "-y",
                 "mcp-remote",
                 "https://endpoint.cognigy.ai/mcp/v1/{token}/mcp"
               ]
             }
           }
         }
         ```

         The `-y` flag lets `npx` run the package without asking for confirmation. Replace `{token}` with the token from the MCP Server URL and `my-endpoint-name` with a name for your Endpoint. The MCP Explorer automatically generates a key based on the Endpoint name.

      3. Close and reopen Cursor. It automatically launches `mcp-remote`, connects to the MCP server, initializes the session, and loads the available tools.

      4. Open the Cursor Chat in Agent mode. Ask the AI to perform a task that requires a tool. Cursor will automatically call the appropriate MCP tool when needed.
    </Tab>
  </Tabs>
</Accordion>

## Use Cases

<AccordionGroup>
  <Accordion title="Access Knowledge Bases">
    External AI clients can use MCP to query Cognigy knowledge bases through the Knowledge Search Tool.
    This approach enables real-time retrieval of relevant content, such as product information, policies, or support articles,
    directly from connected applications.
  </Accordion>

  <Accordion title="Automate Business Tasks">
    MCP clients can trigger operational workflows, such as CRM lookups, ticket creation, or email notifications.
    When a tool is called, the corresponding Flow is executed, allowing existing business logic to run
    without modification.
  </Accordion>

  <Accordion title="Expose Internal Logic">
    Existing Flows can be published as callable tools. This approach allows organizations to reuse and expose
    internal automation, decision logic, and AI Agent configurations as standardized tools
    available to external systems.
  </Accordion>

  <Accordion title="Connect Custom Applications">
    Custom AI applications, development environments, or third-party systems can integrate with Cognigy
    through the MCP protocol. This approach enables structured tool calls and consistent result handling
    without building and maintaining custom APIs.
  </Accordion>
</AccordionGroup>

## Troubleshooting

| **Issue**                                                | **Cause**                                                                                                                            | **Solution**                                                                                                                                                                                                                       |
| -------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Tools Not Visible                                        | The AI Agent Node doesn't contain tool Nodes, or tools are missing a tool ID or description. The wrong Flow or Node may be selected. | Ensure the AI Agent Node includes Tool Nodes. Verify that each tool has a tool ID and description. Confirm the correct Flow and Node are selected in the Endpoint settings.                                                        |
| Tool Call Times Out                                      | The tool execution exceeds the default 30-second timeout.                                                                            | If you have an on-premises installation, adjust the `MCP_SERVER_TOOL_CALL_TIMEOUT` configuration if a longer execution time is required.                                                                                           |
| Connection Fails                                         | The full MCP URL isn't specified, or an incorrect path is configured.                                                                | Ensure you are using the complete URL in the format `/mcp/v1/{token}/mcp`. The Endpoint doesn't respond on other paths.                                                                                                            |
| Session Expired                                          | The MCP session has timed out due to inactivity.                                                                                     | Reinitialize the session by sending a new `initialize` request without a session ID.                                                                                                                                               |
| Authentication Failed                                    | The Bearer token is missing, expired, or doesn't include the required scopes.                                                        | Verify that the token is valid and includes the scopes configured in the **Required Scopes** field. Re-authenticate if the token has expired.                                                                                      |
| The **Run Test** Button in the MCP Explorer isn't Active | A reverse proxy or security layer is changing the Content Security Policy (CSP) headers set by the MCP Explorer page.                | The MCP Explorer sets its own CSP headers using nonce-based script allowlisting. Make sure upstream proxies or security tools don't remove or change these headers. Contact your platform administrator if CSP changes are needed. |


# Microsoft Teams
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/microsoft-teams



<Frame>
  <img alt="Microsoft Teams Endpoint logo" />
</Frame>

The Microsoft Teams Endpoint allows you to connect your AI Agent to [Microsoft Teams](https://www.microsoft.com/en-us/microsoft-teams/), enabling responses when the Microsoft Teams bot is mentioned in a channel. In this case, users must mention (via `@mentions`) the bot to initiate interaction in the channel. For direct messaging scenarios, where users can interact with the bot in a one-on-one chat without the need for `@mentions`, use the Azure Bot Services Endpoint.

## Prerequisites

* A Microsoft Teams account with permissions to manage Microsoft Teams and create outgoing webhooks.

## Restrictions

* The postback functionality, for example `messageBack`, isn't supported by the Microsoft Teams Endpoint. For postback support, use the Azure Bot Services Endpoint instead.

## Generic Endpoint Settings

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [NLU Connectors](/ai/platform-features/nlu/external/nlu-connectors/overview)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)
* [Transformer Functions](/ai/for-developers/transformers/overview)

## Specific Endpoint Settings

<Accordion title="Microsoft Teams Settings">
  | **Parameter**               | **Description**                                                                           |
  | --------------------------- | ----------------------------------------------------------------------------------------- |
  | Security Token              | A token from the Microsoft Teams webhook setup.                                           |
  | Empty Message Fallback Text | A message sent if no output is provided, as Microsoft Teams requires a non-empty message. |
</Accordion>

## How to Set Up

<AccordionGroup>
  <Accordion title="1. Configure a Microsoft Teams Endpoint">
    1. In the left-side menu of your Project, click **Deploy > Endpoints**.
    2. On the **Endpoints** page, click **+ New Endpoint**.
    3. In the **New Endpoint** section, do the following:
       1. Select the **Microsoft Teams** Endpoint type.
       2. Add a unique name.
       3. Select a relevant Flow from the list. Save changes.
    4. Copy the URL from the **Endpoint URL** field and save it for later use.
  </Accordion>

  <Accordion title="2. Create a Webhook in Microsoft Teams">
    1. Open Microsoft Teams and find the team where you want to add the bot. To do this, in the **Chat** panel, go to **Microsoft Teams and channels > More Options > Your Microsoft Teams and channels**.
    2. Hover over the team, click the menu next to the team name and select **Manage team**.
    3. Access the **Apps** tab, go to **Upload an app > Create an outgoing webhook**. If the **Create an outgoing webhook** option isn't available, follow the [Microsoft Outgoing Webhooks](https://learn.microsoft.com/en-us/microsoftteams/platform/webhooks-and-connectors/how-to/add-outgoing-webhook?tabs=urljsonpayload%2Cdotnet) guide to configure your Office 365 tenant.
    4. In the **Create an outgoing webhook** window, fill in the following fields:
       * **Name** – enter the name for the webhook. This name will be used as the bot's name in Microsoft Teams.
       * **Callback URL** – enter the Endpoint URL from your Microsoft Teams Endpoint.
       * **Description** - enter a clear and concise description that explains the bot's purpose and helps users distinguish it from other bots in Microsoft Teams.
    5. Click **Create** to generate the webhook. A pop-up will display a security token. Copy and save it for later use. If the token is lost, you'll need to create a new webhook.
  </Accordion>

  <Accordion title="3. Finish the Integration">
    1. In the Cognigy interface, go to the **Microsoft Teams** Endpoint that you configured previously.
    2. In the **Security Token** field, paste the security token that you copied and saved previously. Save changes.
    3. To test the bot, go to the Microsoft Teams channel related to the team for which the bot was created, mention the bot, for example, `@Bot`, to send a message. The message will be routed to your Cognigy Flow. Always mention the bot to ensure that messages are sent to the Flow.
  </Accordion>
</AccordionGroup>

## Example

### Send a Response in the Adaptive Card Format

1. In the Cognigy interface, create a Flow and add a Say Node.
2. In the Node editor, select the **Microsoft Teams** channel, then **Adaptive Card** from the **Output Type** list.
3. In the code editor, enter the following script:

   <Accordion title="Adaptive Card">
     ```json theme={null}
     {
       "$schema": "http://adaptivecards.io/schemas/adaptive-card.json",
       "type": "AdaptiveCard",
       "version": "1.4",
       "body": [
         {
           "type": "TextBlock",
           "text": "Seattle Center Monorail",
           "weight": "Bolder",
           "size": "Large"
         },
         {
           "type": "TextBlock",
           "text": "Seattle Center Monorail",
           "isSubtle": true,
           "spacing": "None"
         },
         {
           "type": "Image",
           "url": "https://upload.wikimedia.org/wikipedia/commons/thumb/4/49/Seattle_monorail01_2008-02-25.jpg/1024px-Seattle_monorail01_2008-02-25.jpg",
           "size": "Stretch"
         },
         {
           "type": "TextBlock",
           "text": "The Seattle Center Monorail is an elevated train line between Seattle Center (near the Space Needle) and downtown Seattle. It was built for the 1962 World's Fair. Its original two trains, completed in 1961, are still in service.",
           "wrap": true
         }
       ],
       "actions": [
         {
           "type": "Action.OpenUrl",
           "title": "Official website",
           "url": "https://www.seattlemonorail.com"
         },
         {
           "type": "Action.Submit",
           "title": "My MessageBack button",
           "data": {
             "property": "propertyValue",
             "text": "User just clicked the MessageBack button"
           }
         }
       ]
     }
     ```
   </Accordion>
4. Save changes and go to the channel in Microsoft Teams to call your bot.

## More Information

* [Cognigy Help Center: Microsoft Teams Deployment](https://support.cognigy.com/hc/en-us/articles/360016224039-Microsoft-Teams-Deploy-an-Endpoint)


# NiCE CXone
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/nice



<a href="/release-notes/4.82"><Badge>Updated in 4.82</Badge></a>

<Frame>
  <img alt="NiCE CXone Endpoint logo" />

  <img alt="NiCE CXone Endpoint logo" />
</Frame>

The NiCE CXone Endpoint enables you to connect your AI Agent to the [NiCE CXone](https://www.nice.com) platform.

## Prerequisites

* Registered an account on the [NiCE CXone](https://cxone.niceincontact.com/) platform.
* Gained access to the [Virtual Agent Hub](https://help.nice-incontact.com/content/aiassistantsandbots/virtualagenthub/virtualagenthub.htm) in NiCE CXone.
* Created a Cognigy Flow.

## Generic Endpoint Settings

Learn about the generic Endpoint settings available for this Endpoint on the following pages:

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Transformer Functions](/ai/for-developers/transformers/overview)

## Specific Endpoint Settings

<Accordion title="NiCE CXone Setup">
  | Parameter    | Description                                                                                                                                                           |
  | ------------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
  | Verify Token | A verification token for connecting the NiCE CXone Endpoint to the CXone Virtual Agent Hub. This token is generated on the Cognigy side and unique for each Endpoint. |
</Accordion>

## How to Set Up

### Setup on the Cognigy Side

<Accordion title="Configure a NiCE CXone Endpoint">
  1. Go to **Deploy > Endpoints**.
  2. On the **Endpoints** page, click **+ New Endpoint**.
  3. In the **New Endpoint** section, do the following:
     1. Select the **NiCE CXone** Endpoint type.
     2. Add a unique name.
     3. Select a relevant Flow from the list. Save changes.
  4. In the **Configuration Information** section, copy the Endpoint URL and save it for later use in NiCE CXone.
  5. Activate the **Enable Endpoint** setting.
  6. In the **NiCE CXone Setup** section, copy the token from the **Verify Token** field and save it for later use in NiCE CXone.
</Accordion>

### Setup on the NiCE CXone Side

<Accordion title="Create an App in NiCE CXone">
  1. Go to the [NiCE CXone](https://cxone.niceincontact.com/) platform.
  2. In the upper-left corner, click the **Applications** app selector.
  3. Go to **Others > Automation & AI**. A new window opens.
  4. In the left-side **CXone Mpower Hubs** menu, select **Virtual Agent Hub**. The **Virtual Agents** panel appears. Click **Add Bot**.
  5. At **Step 1: General Information**, do the following:
     1. Enter a unique name in the **Virtual Agent (Bot) Name** field.
     2. In the **Select Bot Provider** field, choose the **Cognigy Bot** card. If this card is not present, contact NICE support to activate this feature for your account.
  6. At **Step 2: Configuration**, fill in the following fields:
     * **URL** — enter the Endpoint URL that you saved and copied previously.
     * **Verify Token** — enter the Verify Token that you saved and copied previously.
  7. At **Step 3: Voice**, in the **Voice Integration Options** section, activate the **Text Only** option. With this option, you can only use chat AI Agents.
  8. At **Step 4: Transcript**, leave the **Transcript** field with the **Nothing** value.
  9. At **Step 5: Test + Add**, test the bot by sending a welcome message to the chat. If you receive a response from the Cognigy Agent, the connection is established correctly.
  10. Click **Add** to complete the virtual agent configuration.

  For further integration,
  you will need [NiCE CXone Desktop Studio](https://help.nice-incontact.com/content/studio/fundamentals/fundamentals.htm) and the [channel](https://help.nice-incontact.com/content/acd/channels/acdchannels.htm?tocpath=ACD%7CACD%7CACD%20Channels%7C_____0)
  through which end users will communicate with the Cognigy Agent.
  This logic is configured on the NiCE CXone side.

  For more information, read the following articles in the NiCE CXone documentation:

  * [Cognigy in NiCE CXone](https://help.nicecxone.com/content/aiassistantsandbots/agentassisthub/cognigyagentassist/cognigyagentassist.htm?)
  * [Integrate Cognigy with NiCE CXone](https://help.nice-incontact.com/content/aiassistantsandbots/virtualagenthub/integratecognigy.htm?tocpath=Virtual%20Agent%20Hub%7CVirtual%20Agent%20Hub%7CVirtual%20Agent%20Integrations%7CCognigy%7C_____1)
</Accordion>


# NiCE CXone AAH
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/nice-aah



<a href="/release-notes/2025.25"><Badge>Added in 2025.25 (beta)</Badge></a>

<Frame>
  <img alt="NiCE CXone Agent Assist Hub Endpoint logo" />
</Frame>

The NiCE CXone AAH Endpoint enables you to connect [Agent Copilot](/agent-copilot/overview) directly to [NiCE CXone Agent Assist Hub (AAH)](https://help.nicecxone.com/content/aiassistantsandbots/agentassisthub/cognigyagentassist/cognigyagentassist.htm) in the NiCE CXone Mpower contact center. For more information, read the [Agent Copilot documentation on this Endpoint](/agent-copilot/deploy/nice-aah).


# Endpoint Reference
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/overview

Explore pre-built Cognigy Endpoints including Webchat, Voice Gateway, REST, and third-party integrations.

Cognigy includes many pre-built Endpoints that you can configure.

## Cognigy Endpoints

The following Endpoints are specific to the Cognigy platform:

<CardGroup>
  <Card title="Cognigy Webchat" href="/ai/agents/deploy/endpoint-reference/webchat">
    An open-source web-based chat widget that can be embedded on any website. You can use either the Webchat v3 Endpoint or the Webchat v2 Endpoint.
  </Card>

  <Card title="Cognigy Voice Gateway" href="/ai/agents/deploy/endpoint-reference/voice-gateway">
    Lets you connect to SIP interfaces, so your AI Agent can be connected to the phone network.
  </Card>

  <Card title="Cognigy Voice Copilot" href="/ai/agents/deploy/endpoint-reference/voice-copilot">
    Used for integration between call centers and Agent Copilot for voice use cases.
  </Card>

  <Card title="MCP Server" href="/ai/agents/deploy/endpoint-reference/mcp-server">
    Exposes AI Agent tools via MCP. Acts as a non-conversational tool server for structured tool discovery and execution.
  </Card>
</CardGroup>

## Generic Endpoints

The following Endpoints are provided for custom integrations:

<CardGroup>
  <Card title="REST" href="/ai/agents/deploy/endpoint-reference/rest">
    Connect to a Cognigy Flow directly through a REST interface.
  </Card>

  <Card title="Socket" href="/ai/agents/deploy/endpoint-reference/socketio">
    Connect to a Cognigy Flow directly through a socket client.
  </Card>

  <Card title="Webhook" href="/ai/agents/deploy/endpoint-reference/webhook">
    Expose Cognigy Flows via Webhook.
  </Card>
</CardGroup>

## Third-Party Endpoints

The following Endpoints provide integrations with third-party products:

<CardGroup>
  <Card title="8x8" href="/ai/agents/deploy/endpoint-reference/8x8">
    Connect your AI Agent to the 8x8 platform.
  </Card>

  <Card title="Amazon Alexa (deprecated)" href="/ai/agents/deploy/endpoint-reference/amazon-alexa">
    Connect Cognigy to Amazon's voice AI devices like Echo, Echo Dot, and Echo Show.
  </Card>

  <Card title="Amazon Lex" href="/ai/agents/deploy/endpoint-reference/amazon-lex">
    Connect your Amazon Lex bots to Cognigy via a Lambda function.
  </Card>

  <Card title="Avaya CPaaS" href="/ai/agents/deploy/endpoint-reference/avaya-cpaas">
    Connect your Cognigy resources to an Avaya CPaaS phone number.
  </Card>

  <Card title="Azure Bot Services" href="/ai/agents/deploy/endpoint-reference/azure-bot-services">
    Connect your AI Agent to all Azure Bot Service channels.
  </Card>

  <Card title="Bandwidth" href="/ai/agents/deploy/endpoint-reference/bandwidth">
    Connect your AI Agent to the Bandwidth communications platform.
  </Card>

  <Card title="Dialogflow (deprecated)" href="/ai/agents/deploy/endpoint-reference/dialogflow">
    Connect Flows to all channels supported by Dialogflow.
  </Card>

  <Card title="Facebook (Meta) Messenger" href="/ai/agents/deploy/endpoint-reference/facebook-messenger">
    Integrate with Facebook's messaging app and platform.
  </Card>

  <Card title="Genesys" href="/ai/agents/deploy/endpoint-reference/genesys">
    Connect to Genesys Cloud CX using the Genesys Bot Connector.
  </Card>

  <Card title="Intercom" href="/ai/agents/deploy/endpoint-reference/intercom">
    Connect with the Intercom live chat platform.
  </Card>

  <Card title="LINE Messenger" href="/ai/agents/deploy/endpoint-reference/line-messenger">
    Integrate with the popular LINE messaging and voice app.
  </Card>

  <Card title="Microsoft Teams" href="/ai/agents/deploy/endpoint-reference/microsoft-teams">
    Integrate your AI Agent with Microsoft Teams.
  </Card>

  <Card title="NiCE CXone" href="/ai/agents/deploy/endpoint-reference/nice">
    Connect AI Agents to the NiCE CXone platform.
  </Card>

  <Card title="RingCentral Engage" href="/ai/agents/deploy/endpoint-reference/ringcentral-engage">
    Customer engagement with built-in live chat and multichannel support.
  </Card>

  <Card title="Slack" href="/ai/agents/deploy/endpoint-reference/slack">
    Integrate with Slack for communication and collaboration.
  </Card>

  <Card title="Sunshine Conversations" href="/ai/agents/deploy/endpoint-reference/sunshine-conversations">
    Connect to multiple channels via Zendesk Sunshine Conversations.
  </Card>

  <Card title="Twilio" href="/ai/agents/deploy/endpoint-reference/twilio">
    Connect to Twilio to handle voice interactions.
  </Card>

  <Card title="Twilio SMS" href="/ai/agents/deploy/endpoint-reference/twilio-sms">
    Connect to Twilio to handle chat interactions via SMS and WhatsApp.
  </Card>

  <Card title="Userlike" href="/ai/agents/deploy/endpoint-reference/userlike">
    Integrate with the Userlike live-chat platform.
  </Card>

  <Card title="WhatsApp" href="/ai/agents/deploy/endpoint-reference/whatsapp">
    Connect Cognigy resources to WhatsApp via the Meta Developer Dashboard.
  </Card>

  <Card title="Workplace by Facebook (Meta)" href="/ai/agents/deploy/endpoint-reference/workplace-by-fb">
    Integrate with Workplace by Facebook.
  </Card>

  <Card title="Zoom" href="/ai/agents/deploy/endpoint-reference/zoom">
    Connect your AI Agent to the Zoom Contact Center platform.
  </Card>
</CardGroup>


# REST
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/rest



<a href="/release-notes/2026.16"><Badge>Updated in 2026.16</Badge></a>

<Frame>
  <img alt="REST Endpoint logo" />
</Frame>

The REST Endpoint enables you to connect your AI Agent to a REST client. The REST client can be any application, tool, or service capable of making HTTP requests to RESTful APIs.
This includes custom-built applications, third-party tools like Postman or Insomnia, or backend services that interact with external APIs.

## Prerequisites

* Set up a REST client, for example, Postman, Insomnia, a custom application, or any tool capable of sending HTTP requests.

## Generic Endpoint Settings

Learn about the generic Endpoint settings on the following pages:

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [NLU Connectors](/ai/platform-features/nlu/external/nlu-connectors/overview)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Transformer Functions](/ai/for-developers/transformers/overview)
* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)

## Specific Endpoint Settings

<Accordion title="API Key Authentication">
  Use this section to configure how incoming requests to this Endpoint are authenticated.

  | Parameter             | Type   | Description                                                                                                                                                                                                                                                                                                                                                                                                                                           |
  | --------------------- | ------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
  | Authentication Method | List   | Sets an authentication step for requests sent to this Endpoint. Select one of the options:<ul><li>**None** — no authentication step. This option is selected by default.</li><li>**API Key** — sets API key-based authentication. If you select this option, you need to enter an Endpoint-specific API key in the header of your requests to this Endpoint. To generate an Endpoint-specific API key, use the **Generate API Key** button.</li></ul> |
  | Generate API Key      | Button | This button appears in the **Endpoint API Keys** section when you select **API Key** in **Authentication Method**. Opens a dialog to generate an Endpoint-specific API key. You can generate multiple Endpoint-specific API keys.                                                                                                                                                                                                                     |
</Accordion>

## How to Set Up

### Setup on the Cognigy Side

<Accordion title="Create a REST Endpoint">
  1. In the left-side menu of your Project, go to **Deploy > Endpoints** and click **+ New Endpoint**.
  2. In the **New Endpoint** section, do the following:
     1. Select the **REST** Endpoint type.
     2. Specify a unique name.
     3. Select a Flow from the list.
  3. *(Optional)* To set up Endpoint-specific API authentication, select **API Key** from the **Authentication Method** list in the **API Key Authentication** section. The **Endpoint API Keys** section appears. Follow these steps:
     1. Click **+ Generate API Key**.
     2. In the dialog box, enter a unique name for your API key in the **Key name** field and click **Generate API Key**.
     3. Click the **Key** field to copy the generated API key for later use in the header of requests to the Endpoint URL. Save the API key because you can't retrieve it after closing the dialog. Alternatively, you can create and manage API keys for this Endpoint using the `/v2.0/endpoints/{endpointId}/apikeys` routes via [Cognigy API](https://api-trial.cognigy.ai/openapi#tag--Endpoints).
  4. Save changes and go to the **Configuration Information** section. Copy the URL from the **Endpoint URL** field.
</Accordion>

### Setup on the Third-Party Provider Side

<AccordionGroup>
  <Accordion title="1. Send a Request">
    Send a `POST` request to the Endpoint URL.

    <AccordionGroup>
      <Accordion title="No Authentication">
        <Tabs>
          <Tab title="cURL">
            Replace `https://<your-endpoint-url>` with the Endpoint URL from the Endpoint settings.

            ```bash theme={null}
            curl -X POST https://<your-endpoint-url> \
              -H "Content-Type: application/json" \
              -d '{
                "userId": "user123",
                "sessionId": "session123",
                "text": "Hello, I need help with my order",
                "data": {
                  "exampleKey": "exampleValue"
                }
              }'
            ```
          </Tab>

          <Tab title="Postman">
            1. Open the Postman collection, select **Add a request**, then set the request type to `POST`.
            2. Enter the Endpoint URL as the request URL.
            3. Go to the **Headers** tab and add `Content-Type: application/json`.
            4. Go to the **Body** tab, select **raw**, then select **JSON** as the format.
            5. Paste the request body:

            ```json theme={null}
            {
              "userId": "user123",
              "sessionId": "session123",
              "text": "Hello, I need help with my order",
              "data": {
                "exampleKey": "exampleValue"
              }
            }
            ```
          </Tab>

          <Tab title="Next.js">
            Replace `https://<your-endpoint-url>` with the Endpoint URL from the Endpoint settings.

            ```js theme={null}
            const axios = require('axios');

            async function sendRequest() {
              try {
                const response = await axios.post('https://<your-endpoint-url>', {
                  userId: "user123",
                  sessionId: "session123",
                  text: "Hello, I need help with my order",
                  data: {
                    exampleKey: "exampleValue"
                  }
                }, {
                  headers: {
                    "Content-Type": "application/json"
                  }
                });

                console.log(response.data);
              } catch (error) {
                console.error('Error:', error.message);
              }
            }

            sendRequest();
            ```
          </Tab>

          <Tab title="Python">
            Replace `https://<your-endpoint-url>` with the Endpoint URL from the Endpoint settings.

            ```python theme={null}
            import requests

            url = "https://<your-endpoint-url>"
            payload = {
                "userId": "user123",
                "sessionId": "session123",
                "text": "Hello, I need help with my order",
                "data": {
                    "exampleKey": "exampleValue"
                }
            }
            headers = {
                "Content-Type": "application/json"
            }

            response = requests.post(url, json=payload, headers=headers)
            print(response.json())
            ```

            | Parameter | Type   | Description                                                                                                                                                                                                 | Required                                             |
            | --------- | ------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------- |
            | userId    | String | The ID of the end user.                                                                                                                                                                                     | Yes                                                  |
            | sessionId | String | The ID used to track the current session and maintain its state. Generate a new unique ID for each new session. For testing, you can use any string and change it whenever you want to start a new session. | Yes                                                  |
            | text      | String | The text that the assigned Flow processes.                                                                                                                                                                  | No, if `data` is specified<sup>[1](#footnote1)</sup> |
            | data      | Object | The data that the assigned Flow processes.                                                                                                                                                                  | No, if `text` is specified<sup>[1](#footnote1)</sup> |
          </Tab>
        </Tabs>
      </Accordion>

      <Accordion title="With API Key Authentication">
        The API key authentication header name depends on your Cognigy version.

        <AccordionGroup>
          <Accordion title="For Cognigy 2026.16.0 and later">
            Use the `x-cognigy-endpoint-key` header. The Endpoint also accepts the legacy `x-rest-endpoint-key` header for backward compatibility, but `x-cognigy-endpoint-key` takes precedence if both are provided.

            <Tabs>
              <Tab title="cURL">
                Replace `https://<your-endpoint-url>` with the Endpoint URL and `<your-rest-endpoint-api-key>` with an active API key from your Endpoint settings.

                ```bash theme={null}
                curl -X POST https://<your-endpoint-url> \
                  -H "Content-Type: application/json" \
                  -H "x-cognigy-endpoint-key: <your-rest-endpoint-api-key>" \
                  -d '{
                    "userId": "user123",
                    "sessionId": "session123",
                    "text": "Hello, I need help with my order",
                    "data": {
                      "exampleKey": "exampleValue"
                    }
                  }'
                ```
              </Tab>

              <Tab title="Postman">
                1. Open the Postman collection, select **Add a request**, then set the request type to `POST`.
                2. Enter the Endpoint URL as the request URL.
                3. Go to the **Headers** tab and configure:
                   * `Content-Type` — select `application/json`
                   * `x-cognigy-endpoint-key` — the API key you created using the [Cognigy API](#api-key-authentication).
                4. Go to the **Body** tab, select **raw**, then select **JSON** as the format.
                5. Paste the request body:

                ```json theme={null}
                {
                  "userId": "user123",
                  "sessionId": "session123",
                  "text": "Hello, I need help with my order",
                  "data": {
                    "exampleKey": "exampleValue"
                  }
                }
                ```
              </Tab>

              <Tab title="Next.js">
                Replace `https://<your-endpoint-url>` with the Endpoint URL and `<your-rest-endpoint-api-key>` with an active API key from your Endpoint settings.

                ```js theme={null}
                const axios = require('axios');

                async function sendRequest() {
                  try {
                    const response = await axios.post('https://<your-endpoint-url>', {
                      userId: "user123",
                      sessionId: "session123",
                      text: "Hello, I need help with my order",
                      data: {
                        exampleKey: "exampleValue"
                      }
                    }, {
                      headers: {
                        "Content-Type": "application/json",
                        "x-cognigy-endpoint-key": "<your-rest-endpoint-api-key>"
                      }
                    });

                    console.log(response.data);
                  } catch (error) {
                    console.error('Error:', error.message);
                  }
                }

                sendRequest();
                ```
              </Tab>

              <Tab title="Python">
                Replace `https://<your-endpoint-url>` with the Endpoint URL and `<your-rest-endpoint-api-key>` with an active API key from your Endpoint settings.

                ```python theme={null}
                import requests

                url = "https://<your-endpoint-url>"
                payload = {
                    "userId": "user123",
                    "sessionId": "session123",
                    "text": "Hello, I need help with my order",
                    "data": {
                        "exampleKey": "exampleValue"
                    }
                }
                headers = {
                    "Content-Type": "application/json",
                    "x-cognigy-endpoint-key": "<your-rest-endpoint-api-key>"
                }

                response = requests.post(url, json=payload, headers=headers)
                print(response.json())
                ```
              </Tab>
            </Tabs>
          </Accordion>

          <Accordion title="For Cognigy 2026.14 and 2026.15">
            <Tabs>
              <Tab title="cURL">
                Replace `https://<your-endpoint-url>` with the Endpoint URL and `<your-rest-endpoint-api-key>` with an active API key from your Endpoint settings.

                ```bash theme={null}
                curl -X POST https://<your-endpoint-url> \
                  -H "Content-Type: application/json" \
                  -H "x-rest-endpoint-key: <your-rest-endpoint-api-key>" \
                  -d '{
                    "userId": "user123",
                    "sessionId": "session123",
                    "text": "Hello, I need help with my order",
                    "data": {
                      "exampleKey": "exampleValue"
                    }
                  }'
                ```
              </Tab>

              <Tab title="Postman">
                1. Open the Postman collection, select **Add a request**, then set the request type to `POST`.
                2. Enter the Endpoint URL as the request URL.
                3. Go to the **Headers** tab and configure:
                   * `Content-Type` — select `application/json`
                   * `x-rest-endpoint-key` — the API key you created using the [Cognigy API](#api-key-authentication).
                4. Go to the **Body** tab, select **raw**, then select **JSON** as the format.
                5. Paste the request body:

                ```json theme={null}
                {
                  "userId": "user123",
                  "sessionId": "session123",
                  "text": "Hello, I need help with my order",
                  "data": {
                    "exampleKey": "exampleValue"
                  }
                }
                ```
              </Tab>

              <Tab title="Next.js">
                Replace `https://<your-endpoint-url>` with the Endpoint URL and `<your-rest-endpoint-api-key>` with an active API key from your Endpoint settings.

                ```js theme={null}
                const axios = require('axios');

                async function sendRequest() {
                  try {
                    const response = await axios.post('https://<your-endpoint-url>', {
                      userId: "user123",
                      sessionId: "session123",
                      text: "Hello, I need help with my order",
                      data: {
                        exampleKey: "exampleValue"
                      }
                    }, {
                      headers: {
                        "Content-Type": "application/json",
                        "x-rest-endpoint-key": "<your-rest-endpoint-api-key>"
                      }
                    });

                    console.log(response.data);
                  } catch (error) {
                    console.error('Error:', error.message);
                  }
                }

                sendRequest();
                ```
              </Tab>

              <Tab title="Python">
                Replace `https://<your-endpoint-url>` with the Endpoint URL and `<your-rest-endpoint-api-key>` with an active API key from your Endpoint settings.

                ```python theme={null}
                import requests

                url = "https://<your-endpoint-url>"
                payload = {
                    "userId": "user123",
                    "sessionId": "session123",
                    "text": "Hello, I need help with my order",
                    "data": {
                        "exampleKey": "exampleValue"
                    }
                }
                headers = {
                    "Content-Type": "application/json",
                    "x-rest-endpoint-key": "<your-rest-endpoint-api-key>"
                }

                response = requests.post(url, json=payload, headers=headers)
                print(response.json())
                ```
              </Tab>
            </Tabs>
          </Accordion>
        </AccordionGroup>
      </Accordion>
    </AccordionGroup>
  </Accordion>

  <Accordion title="2. Get a Response">
    The response contains the output text, output data, and the outputStack, which is an array of all Flow outputs. The REST Endpoint can concatenate AI Agent outputs into a single `text` or `data` output, for example, when you place Say Nodes sequentially in the Flow. You can use the outputStack for debugging purposes.

    ```json theme={null}
    {
      "text": "Hello, this is the Cognigy Agent. How can I help you?",
      "data": {},
      "outputStack": [
        {
          "text": "Hello, this is the Cognigy Agent.",
          "data": {},
          "traceId": "endpoint-httpIncomingMessage-068f1748-dadb-4792-bd2e-971335e9c88c",
          "disableSensitiveLogging": false,
          "source": "bot"
        },
        {
          "text": "How can I help you?",
          "data": {},
          "traceId": "endpoint-httpIncomingMessage-068f1748-dadb-4792-bd2e-971335e9c88c",
          "disableSensitiveLogging": false,
          "source": "bot"
        }
      ],
      "userId": "user123",
      "sessionId": "session123"
    }
    ```

    | Parameter                           | Type    | Description                                                                                                                                                                                                    |
    | ----------------------------------- | ------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | userId                              | String  | The ID of the end user.                                                                                                                                                                                        |
    | sessionId                           | String  | The ID used to track the current session and to maintain its state. Generate a new unique ID for each new session. For testing, you can use any string and change it whenever you want to start a new session. |
    | text                                | String  | The text the assigned Flow processes.                                                                                                                                                                          |
    | data                                | Object  | The data the assigned Flow processes.                                                                                                                                                                          |
    | outputStack                         | Array   | An array of outputs generated by the Flow, each with detailed metadata.                                                                                                                                        |
    | outputStack.text                    | String  | The text of the output.                                                                                                                                                                                        |
    | outputStack.data                    | Object  | The data of the output.                                                                                                                                                                                        |
    | outputStack.traceId                 | String  | The ID used for tracing and debugging purposes.                                                                                                                                                                |
    | outputStack.disableSensitiveLogging | Boolean | The flag indicating if logging is disabled. If the value is `true`, this interaction won't be logged for privacy or compliance reasons.                                                                        |
    | outputStack.source                  | String  | The message source. Always `"bot"` for AI Agent replies.                                                                                                                                                       |
  </Accordion>
</AccordionGroup>

***

<sup>1</sup>: You must provide at least one of `text` or `data`. You can send either, or both. If both are missing or invalid, the REST Endpoint throws an error.


# RingCentral Engage
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/ringcentral-engage



<Frame>
  <img alt="RingCentral Engage Endpoint logo" />
</Frame>

<Warning>
  As of Cognigy 4.97.0, the Handover Providers section in the Endpoint settings is deprecated. From Cognigy 2026.4 on, the Handover Settings section in the Endpoint settings is in read-only mode. You can't use this section to configure new handover providers. You can still view configured handover provider settings, but you can't change them.
  To configure handover providers, use the [Handover Providers](/ai/escalate/migration) interface, which offers greater flexibility for managing your integration with contact centers.
  For a smooth migration, refer to the [migration guide](/ai/escalate/migration).
</Warning>

The RingCentral Engage Endpoint enables you to connect your AI Agent to the [RingCentral Engage platform](https://www.ringcentral.com/engage/).

## Prerequisites

* Access to [RingCentral Engage](https://www.ringcentral.com/engage/).
* `Admin` role in RingCentral Engage.

## Generic Endpoint Settings

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [NLU Connectors](/ai/platform-features/nlu/external/nlu-connectors/overview)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)
* [Transformer Functions](/ai/for-developers/transformers/overview)
* [Inject and Notify](/ai/agents/deploy/endpoints/inject-and-notify)

## Specific Endpoint Settings

<Accordion title="RingCentral Engage Setup">
  | **Parameter**      | **Description**                                                                                                                                |
  | ------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------- |
  | API URL            | The URL determined by your application name and API version. The URL format is `https://api.ringcentral.com/restapi/v1.0/accounts/{accountId}` |
  | Access Token       | The token for API requests.                                                                                                                    |
  | Verification Token | The token that adds an extra layer of security to verify the authenticity of API requests to your RingCentral API domain.                      |
  | Secret             | The secret key matching the RingCentral Engage webhook configuration, used for the handshake process.                                          |
  | Bot Category ID    | The category ID used to identify the bot, obtained from the RingCentral Engage platform.                                                       |
  | Agent Category ID  | The category ID used to identify the agent, obtained from the RingCentral Engage platform.                                                     |
</Accordion>

## How to Set Up

<AccordionGroup>
  <Accordion title="1. Create a RingCentral Engage Endpoint">
    1. In the left-side menu of your Project, click **Deploy > Endpoints**.
    2. On the **Endpoints** page, click **+ New Endpoint**.
    3. In the **New Endpoint** section, do the following:
       1. Select the **RingCentral Engage** Endpoint type.
       2. Specify a unique name.
       3. Select a Flow from the list. Save changes.
    4. Go to the **Configuration Information** section and copy the URL from the **Endpoint URL** field.
    5. In the **RingCentral Engage Setup** section, copy the **Secret** value for later use.
    6. Create a verification token and paste it into the **Verification Token** field. The token can be any string, but must be kept secret and secure. Copy the value for later use. This token will be used to verify the authenticity of API requests to your RingCentral API domain.
  </Accordion>

  <Accordion title="2. Create API Access Token">
    1. Log in to your [RingCentral Engage account](https://www.ringcentral.com/engage/).
    2. Go to **ADMIN > API Access Tokens**. Click **+** to create a new token.
    3. Assign an Agent with permissions to read and reply to your source.
    4. Copy the value from the **Access Token** field for later use.
  </Accordion>

  <Accordion title="3. Create Categories">
    1. Go to **ADMIN > Categories**.
    2. Create a new parent category. For example, `Cognigy`. Then, create two subcategories: `Bot` and `Agent`.
    3. Copy the category IDs from the URL for later use.
  </Accordion>

  <Accordion title="4. Configure Sources">
    1. Go to **ADMIN > Sources**.
    2. Select your source and under **MESSAGE SETTINGS**, set the **Default Category** to the **Bot** category.
    3. Save the configuration.
  </Accordion>

  <Accordion title="5. Create a Webhook">
    1. Go to **ADMIN > Webhooks**.
    2. Click **+** to create a new webhook. Configure the following settings:
       * **API Access Token** – use the token created earlier.
       * **Active** – enable the webhook.
       * **URL** – paste the Cognigy Endpoint URL.
       * **Secret** – paste the secret from Cognigy Endpoint.
       * **Sources Filtering Strategy** – select **No sources except**.
       * **Sources** – select your source.
       * **Registered Events** – enter `content.imported`.
       * **Verification Token** – paste the verification token from Cognigy Endpoint.
    3. Save the webhook.
  </Accordion>

  <Accordion title="6. Complete the Endpoint Configuration">
    1. In the **RingCentral Engage Setup** section of your Endpoint, fill in the following fields:
       * **API URL** – enter the URL in the following format: `https://api.ringcentral.com/restapi/v1.0/accounts/{accountId}`.
       * **Access Token** – paste the Access Token copied earlier from RingCentral Engage.
       * **Bot Category ID** – paste the Bot Category ID copied earlier from RingCentral Engage.
       * **Agent Category ID** – paste the Agent Category ID copied earlier.
    2. Save changes. Test the connection by sending a message via RingCentral Engage chat.
  </Accordion>
</AccordionGroup>

## More Information

* [Cognigy Help Center: RingCentral Engage Deployment](https://support.cognigy.com/hc/en-us/articles/360016306779-RingCentral-Engage-Deploy-an-Endpoint)
* [Cognigy Help Center: RingCentral Engage Send Structured Content](https://support.cognigy.com/hc/en-us/articles/360016306859-RingCentral-Engage-Send-Structured-Content)
* [Cognigy Help Center: RingCentral Engage Handover Integration](https://support.cognigy.com/hc/en-us/articles/360016275260-RingCentral-Engage-Handover-Integration)


# Slack
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/slack



<a href="/release-notes/2026.16"><Badge>Updated in 2026.16.0</Badge></a>

<Frame>
  <img alt="Slack Endpoint logo" />
</Frame>

The Slack Endpoint enables you to connect your AI Agent to the [Slack platform](https://api.slack.com/apps/).

## Prerequisites

* Access to the [Slack API](https://api.slack.com/apps/).

## Generic Endpoint Settings

Learn about the generic Endpoint settings on the following pages:

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [NLU Connectors](/ai/platform-features/nlu/external/nlu-connectors/overview)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Session Management](/ai/agents/deploy/endpoints/session-management)
* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)
* [Transformer Functions](/ai/for-developers/transformers/overview)
* [Inject and Notify](/ai/agents/deploy/endpoints/inject-and-notify)
* [Copilot](/agent-copilot/configure/copilot-section)

## Specific Endpoint Settings

<Accordion title="Slack Setup">
  | Parameter          | Description                                                                                                                        |
  | ------------------ | ---------------------------------------------------------------------------------------------------------------------------------- |
  | OAuth Access Token | The token for your Slack app's workspace, obtained from the **OAuth & Permissions** page of your Slack app.                        |
  | Verification Token | The token of your Slack app, obtained from the **Basic Information** page of your Slack app under the **App Credentials** section. |
</Accordion>

## How to Set Up

<AccordionGroup>
  <Accordion title="1. Create a Slack Endpoint">
    1. In the left-side menu of your Project, click **Deploy > Endpoints**.
    2. On the **Endpoints** page, click **+ New Endpoint**.
    3. In the **New Endpoint** section, do the following:
       1. Select the **Slack** Endpoint type.
       2. Specify a unique name.
       3. Select a Flow from the list. Save changes.
    4. Go to the **Configuration Information** section and copy the URL from the **Endpoint URL** field.
  </Accordion>

  <Accordion title="2. Create a Slack app">
    1. Go to the [Slack API](https://api.slack.com/apps) and click **Create New App**.
    2. Select **From scratch**. Enter a name for your app and select the workspace where you want to install it. Click **Create App**. Make sure your app is installed in the workspace.
    3. After creating the app, you will be redirected to the **Basic Information** page of your Slack app.
    4. In the **App Credentials** section, you will find the **Client ID** and **Client Secret**. You will need these later for OAuth configuration.
    5. In the left sidebar, go to **OAuth & Permissions**. Configure the necessary permissions for your app. Under **Scopes**, add the following app token scopes:
       * `app_mentions:read`
       * `chat:write`
       * `im:write`
       * `incoming-webhook`
    6. Go to **Settings > Basic Information** in your Slack app. In the **App Credentials** section, copy the value from the **Verification Token** field and save it for later use.
    7. Go to the **OAuth & Permissions** page of your Slack app and scroll to the **Tokens for Your Workspace** section. Copy the **Bot User OAuth Access Token** (it starts with `xoxb-`) and save it. You can also use the OAuth Access Token; in this case, the app will have the same permissions as your user account.
    8. In the **Event Subscriptions** section, enable events and set the **Request URL** to the Cognigy Endpoint URL. In the **Subscribe to Bot Events** section, add the following events:
       * `app_mention`
       * `message.im`
    9. To enable Quick Reply messages, follow Slack's official guide on [Interactivity & Shortcuts](https://api.slack.com/interactivity/handling#setup).
  </Accordion>

  <Accordion title="3. Complete the Slack Endpoint Configuration">
    1. In the Cognigy interface, go to the Slack Endpoint you created. Go to the **Slack Setup** section.
    2. In the **OAuth Access Token** field, paste the access token you copied from Slack.
    3. In the **Verification Token** field, paste the verification Token you copied from Slack. Save changes.
  </Accordion>
</AccordionGroup>

## More Information

* [Help Center: Slack Deployment](https://support.cognigy.com/hc/en-us/articles/360016307039-Slack-Deploy-an-Endpoint)
* [Help Center: Slack Renaming a Bot](https://support.cognigy.com/hc/en-us/articles/360016271340-Slack-Rename-your-bot)


# Socket.IO
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/socketio



<a href="/release-notes/2026.16"><Badge>Updated in 2026.16.0</Badge></a>

<Frame>
  <img alt="Socket.IO Endpoint logo" />
</Frame>

The Socket.IO Endpoint enables you to connect your AI Agent to the Socket.IO client.

## Prerequisites

* Familiarity with the [Socket.IO library](https://socket.io/).
* Setup of the [Cognigy Socket.IO client](https://github.com/Cognigy/SocketClient).

## Generic Endpoint Settings

Learn about the generic Endpoint settings on the following pages:

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [NLU Connectors](/ai/platform-features/nlu/external/nlu-connectors/overview)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Transformer Functions](/ai/for-developers/transformers/overview)
* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)
* [Inject and Notify](/ai/agents/deploy/endpoints/inject-and-notify)
* [Copilot](/agent-copilot/configure/copilot-section)

## How to Set Up

### Setup on the Cognigy Side

<Accordion title="1. Configure a Socket.IO Endpoint">
  1. In the left-side menu of your Project, click **Deploy > Endpoints**.
  2. On the **Endpoints** page, click **+ New Endpoint**.
  3. In the **New Endpoint** section, do the following:
     1. Select the **Socket.IO** Endpoint type.
     2. Specify a unique name.
     3. Select a Flow from the list. Save changes.
  4. In the Endpoint editor, go to the **Configuration Information** section and copy the URL from the **Endpoint URL** field.
     The Endpoint URL consists of two parts separated by a slash:

     * The server address (`<Socket-Endpoint-URL>`), for example, `https://endpoint-trial.cognigy.ai`.
     * The token (`<URLToken>`) that comes after the slash, for example, `0e77b9a19f33cb68b3f528b3b28d4b32386162559160ceb9ec85036d83dd8f8b`.

     For further configuration, you need to use these two parts separately.
</Accordion>

### Setup on the Third-Party Provider Side

<AccordionGroup>
  <Accordion title="1. Send a Request">
    <Tabs>
      <Tab title="Cognigy Socket.IO client">
        ```js theme={null}
        const { SocketClient } = require("@cognigy/socket-client");

        (async () => {
          // Create the SocketClient with the server address + URL Token + optional options
          const client = new SocketClient(
            "https://endpoint-trial.cognigy.ai",             // the Socket.IO Endpoint server address
            "0e77b9a19f33cb68b3f528b3b28d4b32386162559160ceb9ec85036d83dd8xxx", // the Socket.IO URL token
            {
              userId: "user@cognigy.com",  
              sessionId: "session123", 
              forceWebsockets: true 
            }
          );

          // Listen for messages from the AI Agent
          client.on("output", (output) => {
            console.log("AI Agent:", output.text);
            console.log("Additional data:", output.data);
          });

          // Listen for errors
          client.on("error", (error) => {
            console.error("AI Agent error:", error.message);
          });

          // Listen for finalPing (bot done processing)
          client.on("finalPing", () => {
            console.log("AI Agent finished processing a message");
          });

          // Connect to the Socket.IO Endpoint
          await client.connect();

          // Send a message — BASIC
          client.sendMessage("Hello from test!");

          // Send a message — WITH EXTRA DATA
          client.sendMessage(
            "Example text with custom data",
            {
              passthroughIP: "127.0.0.1", // custom IP if needed
              resetFlow: false,           // false means: continue Flow
              key: "value"                // any custom data for your Flow
            }
          );

          // Reset the Flow if needed
          // client.sendMessage(
          //   "Starting fresh!",
          //   {
          //     resetFlow: true
          //   }
          // );
        })();
        ```
      </Tab>

      <Tab title="Other clients">
        Messages are sent using the processInput event with a payload in the following format:

        ```json theme={null}
        {
            "URLToken": "urlToken",
            "sessionId": "sessionId",
            "userId": "user@cognigy.com",
            "passthroughIP": "127.0.0.1",
            "resetFlow": "false",  // Resets the Flow and starts a new one from the beginning
            "text": "Example text",  // Can be skipped with "resetFlow": true
            "data": {
                "key": "value"
            }
        }
        ```

        | Parameter     | Type    | Description                                                                                                                                                                                                                                                                     | Required                      |
        | ------------- | ------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------- |
        | URLToken      | String  | An authentication token for connecting to the Endpoint.                                                                                                                                                                                                                         | Yes                           |
        | userId        | String  | The ID of the end user.                                                                                                                                                                                                                                                         | Yes                           |
        | sessionId     | String  | The ID used to track the current session and maintain its state. Generate a new unique ID for each new session. For testing, you can use any string and change it whenever you want to start a new session.                                                                     | Yes                           |
        | passthroughIP | String  | The IP address of the user, used for logging or tracking purposes.                                                                                                                                                                                                              | No                            |
        | resetFlow     | Boolean | When set to `true`, resets the conversation Flow. This payload may or may not contain a message. If it contains a message, it will be the first in the new Flow. Make sure to set `resetFlow` to `false` for subsequent messages to continue the Flow instead of restarting it. | Yes                           |
        | text          | String  | The text that the assigned Flow should process.                                                                                                                                                                                                                                 | No, if `resetFlow` is `true`. |
        | data          | Object  | The data that the assigned Flow should process.                                                                                                                                                                                                                                 | No                            |
      </Tab>
    </Tabs>
  </Accordion>

  <Accordion title="2. Get a Response">
    Messages can be received by listening to the `output` event. The responses have the following format:

    <Tabs>
      <Tab title="Cognigy Socket.IO client">
        ```text theme={null}
        AI Agent: I’d be happy to help. Could you please provide your order number?
        Additional data: {}
        AI Agent finished processing a message
        ```
      </Tab>

      <Tab title="Other clients">
        ```json theme={null}
        {
          "type":"output",
          "data":
          {
            "text":"responseText",
            "data":
            {
              "key":"value"
            }
          }
        }
        ```
      </Tab>
    </Tabs>
  </Accordion>
</AccordionGroup>


# Sunshine Conversations
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/sunshine-conversations



<a href="/release-notes/2026.16"><Badge>Updated in 2026.16.0</Badge></a>

<Frame>
  <img alt="Sunshine Conversations Endpoint logo" />
</Frame>

The Sunshine Conversations Endpoint enables you to connect your AI Agent to the [Zendesk's Sunshine Conversations messaging platform](https://app.smooch.io/).

## Generic Endpoint Settings

Learn about the generic Endpoint settings on the following pages:

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [NLU Connectors](/ai/platform-features/nlu/external/nlu-connectors/overview)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)
* [Session Management](/ai/agents/deploy/endpoints/session-management)
* [Transformer Functions](/ai/for-developers/transformers/overview)
* [Inject and Notify](/ai/agents/deploy/endpoints/inject-and-notify)
* [Copilot](/agent-copilot/configure/copilot-section)

## Specific Endpoint Settings

<Accordion title="Sunshine Conversations Setup">
  | Parameter   | Description                                                                                                                                                                                                                 |
  | ----------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
  | API Version | The Sunshine Conversations API version Cognigy uses to communicate with your app. Cognigy supports version 1.1 and version 2. The version must match the one used to create the webhook in your Sunshine Conversations app. |
  | Key ID      | The API key ID from your Sunshine Conversations app. The format is `app_XXXXXXXXXXXXXXXXXXXXXXXX`. The key ID is paired with the secret you provide.                                                                        |
  | Secret      | The API secret generated for your Sunshine Conversations app. This secret must match the key ID you entered earlier.                                                                                                        |
  | URI         | The base URL for your Sunshine Conversations app. For example, `https://app.smooch.io`. Don't include a trailing slash (`/`).                                                                                               |
</Accordion>

## How to Set Up

<AccordionGroup>
  <Accordion title="1. Create a Sunshine Conversations Endpoint">
    1. In the left-side menu of your Project, click **Deploy > Endpoints**.
    2. On the **Endpoints** page, click **+ New Endpoint**.
    3. In the **New Endpoint** section, do the following:
       1. Select the **Sunshine Conversations** Endpoint type.
       2. Specify a unique name.
       3. Select a Flow from the list.
    4. In the **Configuration Information** section, copy the URL from the **Endpoint URL** field.
  </Accordion>

  <Accordion title="2. Create a Sunshine Conversations App">
    1. Go to the [Sunshine Conversations Dashboard](https://app.smooch.io/) and log in with your Zendesk account.
    2. If you don't have a workspace yet, create one first. Otherwise, select your existing workspace.
    3. Click **Create App**. Enter the name and description for your app. Click **Create**.
    4. Go to the **Channels** or **Integrations** tab. Add the messaging channels you want, for example, WhatsApp, Facebook Messenger, Slack. Follow the Sunshine Conversations instructions for each integration.
    5. Open the **Settings** tab of your app. Go to **API Keys** and click **+ Create new secret key**.
    6. Copy the **Key ID** and **Secret** for later use in the Cognigy Endpoint. Save Changes.
  </Accordion>

  <Accordion title="3. Complete the Endpoint Configuration">
    1. In the Cognigy interface, open the Sunshine Conversations Endpoint you created.
    2. In the **Sunshine Conversations Setup** section, fill in the required fields:
       * **API Version** - select either v1.1 or v2, matching your Sunshine Conversations app.
       * **Key ID** - paste the key ID you copied from Sunshine Conversations.
       * **Secret** - paste the secret you copied.
       * **URI** - use your app's base URI, for example, `https://app.smooch.io` (omit the trailing `/`).
    3. Save changes.
  </Accordion>

  <Accordion title="4. Add a Webhook in Sunshine Conversations">
    1. In your Sunshine Conversations app, go to **Settings > Integrations > Webhooks**.
    2. Click **Add Webhook**.
    3. Paste the Cognigy Endpoint URL you copied earlier into the **URL** field.
    4. For **Triggers**, select at least the following options:
       * `AppUser messages`
       * `Postbacks`
    5. Ensure the **API Version** matches your Cognigy Endpoint version (v1.1 or v2). Save the webhook.
  </Accordion>
</AccordionGroup>

## More Information

* [Cognigy Help Center: Sunshine Conversations Deployment](https://support.cognigy.com/hc/en-us/articles/360016307699-Sunshine-Conversations-Deploy-an-Endpoint)
* [Cognigy Help Center: Sunshine Conversations Zendesk Handovers](https://support.cognigy.com/hc/en-us/articles/360016224599-Sunshine-Conversations-Zendesk-Handovers)


# Twilio
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/twilio



<Frame>
  <img alt="Twilio Endpoint logo" />
</Frame>

The Twilio Endpoint enables you to connect your AI Agent to the [Twilio platform](https://www.twilio.com/), which allows you to handle voice calls directly through Twilio.

## Prerequisites

* Access to the [Twilio platform](https://www.twilio.com/).

## Generic Endpoint Settings

Learn about the generic Endpoint settings on the following pages:

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [NLU Connectors](/ai/platform-features/nlu/external/nlu-connectors/overview)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Session Management](/ai/agents/deploy/endpoints/session-management)
* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)
* [Transformer Functions](/ai/for-developers/transformers/overview)

## Specific Endpoint Settings

<Accordion title="Twilio Settings">
  | **Parameter**      | **Description**                                                                                                                    |
  | ------------------ | ---------------------------------------------------------------------------------------------------------------------------------- |
  | STT / TTS Language | The language used by the Speech-to-Text (STT) and Text-to-Speech (TTS) services for messages exchanged between Twilio and Cognigy. |
  | Voice              | The delivery voice for your Twilio voice bot from the list of available Twilio voices.                                             |
</Accordion>

## How to Set Up

### Configure on the Cognigy Side

<Accordion title="Create a Twilio Endpoint">
  To create a Twilio Endpoint, follow these steps:

  1. Go to **Deploy > Endpoints**.
  2. On the **Endpoints** page, click **+ New Endpoint**.
  3. In the **New Endpoint** section, do the following:
     1. Select the **Twilio** Endpoint type.
     2. Specify a unique name.
     3. Select a Flow from the list. Save changes.
  4. In the **Twilio Settings** section, configure the following settings:
     * **STT / TTS Language** — select the language used by the Speech-to-Text (STT) and Text-to-Speech (TTS) services for messages exchanged between Twilio and Cognigy.
     * **Voice** — select the delivery voice for your Twilio voice bot from the list of available Twilio voices.
  5. In the **Configuration Information** section, copy the Endpoint URL and save it for later use. Save changes.
</Accordion>

### Configure on the Twilio Side

<Accordion title="Get a Phone Number">
  1. Open the Twilio Console. Go to the Account Dashboard.
  2. In the left sidebar, select the **Develop** tab, then go to **Phone Numbers > Manage > Buy a Number**. Select the country and ensure the number supports voice messages. Activate **Voice**. You can search for area code, for example, 211, which shows numbers available for Düsseldorf, Germany. Select the number and click **Buy**.
  3. Some phone numbers require you to validate that your company operates in the same city or region. This approach is common for German numbers. To validate the number, read the [Twilio documentation](https://help.twilio.com/) and upload proof of your company registration. The validation process can take up to 3 business days. US numbers don't require this validation.
  4. Go to **Active Numbers**. Select the number, then go to **Voice Configuration**. From the **A call comes in** list, select **Webhook**. In the **URL** field, enter the Cognigy Endpoint URL. Click **Save configuration**.
</Accordion>

## Additional Configuration

<Accordion title="Trigger Twilio to Initiate a Call">
  In addition to receiving calls from Twilio in Cognigy, you can also trigger Twilio to initiate a call.

  1. Create a [TwiML Bin](https://www.twilio.com/docs/serverless/twiml-bins/getting-started) in Twilio to handle the call. Twilio can host TwiML (Twilio Markup Language) for you. This TwiML tells Twilio what to do after the call connects. Use the following TwiML, replacing `your_endpoint_url` with your valid Cognigy Endpoint URL:

     ```xml theme={null}
     <?xml version="1.0" encoding="UTF-8"?>
     <Response>
       <Redirect>your_endpoint_url</Redirect>
     </Response>
     ```

     Copy the URL of the TwiML Bin for later use.

  2. Add an HTTP Request Node to your Cognigy voice Flow. To configure this Node to trigger Twilio to make a call and connect the call back to the Flow, set the following parameters:

     | **Parameter**      | **Description**                                                                                                                                                                                                                            |
     | ------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
     | Type               | The HTTP method used to send the request. Enter `POST`.                                                                                                                                                                                    |
     | URL                | The URL pattern is `https://api.twilio.com/2010-04-01/Accounts/your_twilio_account_sid/Calls.json`. Replace `your_twilio_account_sid` with your actual Twilio account SID.                                                                 |
     | Headers            | The header is `Content-Type: application/x-www-form-urlencoded`. Required to tell Twilio the request format.                                                                                                                               |
     | Payload is JSON    | A flag that controls the payload format. Enter `false`.                                                                                                                                                                                    |
     | Payload            | The data sent in the request body. Use: `To=number_to_call&Url=your_twiml_bin_url&From=your_twilio_number`  <br />- `to` – the phone number to call. <br />- `url` – the URL of your TwiML Bin. <br />- `from` – your Twilio phone number. |
     | Authorization Type | Twilio requires basic authentication for API calls. Enter `Basic Auth`.                                                                                                                                                                    |
     | Username           | Your Twilio **Account SID**. You can find this on the Account Dashboard.                                                                                                                                                                   |
     | Password           | Your Twilio **Auth Token**. You can find this on the Account Dashboard.                                                                                                                                                                    |
</Accordion>

## More Information

* [Cognigy Help Center: Twilio Intro](https://support.cognigy.com/hc/en-us/articles/360016308619-Twilio-Intro)


# Twilio SMS
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/twilio-sms



<Frame>
  <img alt="Twilio SMS Endpoint logo" />
</Frame>

The Twilio SMS Endpoint enables you to connect your AI Agent to the [Twilio platform](https://www.twilio.com/) to handle SMS and WhatsApp messages. Via this Endpoint, your end users can interact with your AI Agent through text messaging.

## Prerequisites

* Access to the [Twilio platform](https://www.twilio.com/).

## Generic Endpoint Settings

Learn about the generic Endpoint settings on the following pages:

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [NLU Connectors](/ai/platform-features/nlu/external/nlu-connectors/overview)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Session Management](/ai/agents/deploy/endpoints/session-management)
* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)
* [Transformer Functions](/ai/for-developers/transformers/overview)

## How to Set Up

### Configure on the Cognigy Side

<Accordion title="1. Create a Twilio Endpoint">
  To create a Twilio Endpoint, follow these steps:

  1. Go to **Deploy > Endpoints**.
  2. On the **Endpoints** page, click **+ New Endpoint**.
  3. In the **New Endpoint** section, do the following:
     1. Select the **Twilio SMS** Endpoint type.
     2. Specify a unique name.
     3. Select a Flow from the list. Save changes.
  4. In the **Configuration Information** section, copy the Endpoint URL and save it for later use.
</Accordion>

### Configure on the Twilio Side

<AccordionGroup>
  <Accordion title="1. Get a Phone Number">
    1. Open the Twilio Console. Go to the Account Dashboard.
    2. In the left sidebar, select the **Develop** tab, then go to **Phone Numbers > Manage > Buy a Number**. Select the country and ensure the number supports SMS. Activate **SMS**. You can search for area code, for example, 211, which shows numbers available for Düsseldorf, Germany. Select the number and click **Buy**.
    3. Some phone numbers require you to validate that your company operates in the same city or region. This approach is common for German numbers. To validate the number, read the [Twilio documentation](https://help.twilio.com/) and upload proof of your company registration. The validation process can take up to 3 business days. US numbers don't require this validation.
    4. Go to **Active Numbers**. Select the number, then go to **Voice Configuration**. From the **A call comes in** list, select **Webhook**. In the **URL** field, enter the Cognigy Endpoint URL. Click **Save configuration**. To test the connection, send an SMS to your Twilio number.
  </Accordion>

  <Accordion title="2. Connect Your Twilio Number to WhatsApp">
    1. Go to **Explore Products > Messaging**.
    2. Select **WhatsApp**. To connect the Twilio number to WhatsApp, your Twilio account must first be approved by WhatsApp.
  </Accordion>
</AccordionGroup>

## More Information

* [Cognigy Help Center: Twilio Intro](https://support.cognigy.com/hc/en-us/articles/360016308619-Twilio-Intro)


# Userlike
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/userlike



<Frame>
  <img alt="Userlike Endpoint logo" />
</Frame>

The Userlike Endpoint enables you to connect your AI Agent to the [Userlike platform](https://www.8x8.com/).

## Prerequisites

* Access to [Userlike](https://www.userlike.com/).

## Generic Endpoint Settings

Learn about the generic Endpoint settings on the following pages:

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [NLU Connectors](/ai/platform-features/nlu/external/nlu-connectors/overview)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)
* [Transformer Functions](/ai/for-developers/transformers/overview)

## Specific Endpoint Settings

<Accordion title="Userlike Settings">
  | **Parameter**  | **Description**                                                                                                     |
  | -------------- | ------------------------------------------------------------------------------------------------------------------- |
  | API Version    | Select the API version you want to use for the Userlike integration: **Live Chat** or **Unified Messaging**.        |
  | Webhook URL    | If you select **Unified Messaging**, specify the URL that Userlike calls to send events and messages.               |
  | Security Token | If you select **Unified Messaging**, specify the token used to authenticate requests from Userlike to your webhook. |
</Accordion>

## How to Set Up

### Setup on the Cognigy Side

<Accordion title="Create a Userlike Endpoint">
  1. In the left-side menu of your Project, click **Deploy > Endpoints**.
  2. On the **Endpoints** page, click **+ New Endpoint**.
  3. In the **New Endpoint** section, do the following:
     1. Select the **Userlike** Endpoint type.
     2. Specify a unique name.
     3. Select a Flow from the list and save changes.
  4. Go to the **Configuration Information** section and copy the URL from the **Endpoint URL** field. Save changes.
</Accordion>

### Setup on the Userlike Side

<Accordion title="Create a Chatbot in Userlike">
  1. Log in to the Userlike Dashboard.
  2. From the left-side menu, select **Settings > Chatbots**.
  3. Click **+ Add** to create a new chatbot.
  4. On the **New chatbot** page, configure the following settings:
     * **Name** — enter the chatbot name.
     * **Operator group** — select the group for the chatbot.
     * **Bot behavior mode** — select **Human**.
     * **Bot type** — select **API**.
  5. Save your changes. When the chatbot appears in the **Chatbots** list, go to the **Actions** column, click the three-dot button, then select **Edit**.
  6. In the **Webhook URL** field, paste the Endpoint URL from the Cognigy Userlike Endpoint.
  7. If you selected the **Unified Messaging API** in the Cognigy Userlike Endpoint, copy the following fields:
     * **API URL** — copy and paste this URL into the **Webhook URL** field in the Cognigy Endpoint.
     * **API security token** — copy and paste this token into the **Security Token** field in the Cognigy Endpoint.
  8. Save your changes. Go to **Settings > Channels**. Select **Website**, then click **+ Add widget**. Copy the code from the **Your Userlike Widget** field and click **Configure widget**.
  9. When the widget appears in the **Widgets** list, go to the **Actions** column, click the three-dot button, then select **Edit**.
  10. In the **Messenger URL**, enter the URL where you want to display your widget. Open your website to see the widget and test the workflow.
</Accordion>

## More Information

* [Help Center: Userlike Deployment](https://support.cognigy.com/hc/en-us/articles/360016310079)


# Voice Copilot
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/voice-copilot



<a href="/release-notes/4.66"><Badge>Updated in 4.66</Badge></a>

<Frame>
  <img alt="Voice Copilot Endpoint logo" />
</Frame>

The Voice Copilot Endpoint handles transcripts from voice calls via webhooks, processes them in real time and sends them to [Agent Copilot](/agent-copilot/overview) Flows. After receiving a transcript, the Endpoint passes the `userId` and `sessionId` variables to Cognigy to identify the Agent Copilot workspace to be updated.

The Voice Copilot Endpoint works with any [handover provider](/ai/escalate/handover-reference/overview) supported by Cognigy. This Endpoint also supports integration with SIP recording (SIPREC). SIPREC is an open SIP-based protocol for call recording in contact centers. Session Border Controllers (SBC) integrate SIPREC with the contact center.

You can [configure](#configure-the-voice-copilot-endpoint) the Voice Copilot Endpoint in the following setups:

* Cognigy in front of a contact center
* A SIPREC contact center integration in front of Voice Gateway

## Key Features

* Processes real-time speech transcripts.
* Manages user and session identification.
* Supports human agent and user tagging in channels.
* Implements output buffering for sequential message processing.
* Handles SIPREC INVITE requests for call setup.

## Prerequisites

* You have configured an [Agent Copilot Flow](/agent-copilot/configure/overview).

## Transcript Data Flow

The data flow shows how the Endpoint receives call transcripts, identifies the speaker, sends the data to Agent Copilot for processing, and then processes the responses to send back.

```mermaid theme={null}
flowchart TD
    A["Receives call transcript<br/>via webhook"] --> B["Processes speech data<br/>with source identification"]
    B --> C["Transforms data for<br/>processing in Cognigy"]
    C --> D["Sends the data to<br/>the Agent Copilot Flow"]
    D --> E["Buffers and processes<br/>responses"]
```

## Generic Endpoint Settings

Find out about the generic endpoint settings available with this endpoint on the following pages:

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [Transformer Functions](/ai/for-developers/transformers/overview)

## Set up a Voice Copilot Endpoint

1. In **Deploy > Endpoints**, click **+ New Endpoint**.
2. In the **New Endpoint** section, configure the following:
   * **Endpoint Type** — select **Voice Copilot**.
   * **Name** — enter a unique name.
   * **Flow** — select a Flow from the list.
3. In the **Configuration Information** section, copy the Endpoint URL for later use in the Transfer Node if you use [Cognigy in front](#configure-the-voice-copilot-endpoint) of your contact center.

## Configure the Voice Copilot Endpoint

<Tabs>
  <Tab title="Cognigy in Front">
    Configure the Voice Copilot Endpoint in a Transfer Node:

    1. In **Build > Flows**, select a voice Flow from the list.
    2. In the Transfer Node, select the **Dial** transfer type.
       1. Activate the **Enable Copilot** setting.
       2. Enter the **Copilot Headers Key**. The key is provider-dependent. For example, for Genesys, the key is `User-to-User`. The generated value becomes accessible after engaging the Transfer Node, located in the Input object as `{{input.UUIValue}}`. You don't need to handle this value manually. It is automatically generated and added to the header.
       3. In the **Transcribe** section, enter the Voice Copilot Endpoint URL in the **Transcript Webhook** field. You need to copy this Endpoint URL from the **Endpoint URL** field in the Endpoint settings. For example, `https://endpoint.cognigy.ai/<your-endpoint-id>`. This setup ensures that the Voice Copilot Endpoint receives the transcript and the corresponding session and user IDs, ensuring the transcript is correctly matched with the Agent Copilot workspace.
  </Tab>

  <Tab title="SIPREC Contact Center Integration in Front">
    When a call is transferred to a human agent, [Voice Gateway](/voice-gateway/overview) transcribes the audio from a SIPREC call and sends it to the Voice Copilot Endpoint for processing. This approach replaces the [user ID and session ID](/ai/agents/analyze/user-and-session-identification) for contact center-specific identifiers, for example, agent ID and conversation ID.

    **Configure a SIPREC Contact Center Integration**

    On the SIPREC contact center side, configure the following:

    1. In the SIPREC call settings, set the Voice Gateway IP addresses as the target of the audio stream. You can find the Voice Gateway IP addresses in the Voice Gateway Self-Service Portal under [Carriers](/voice-gateway/webapp/carriers). Make sure you allow your carriers to send data to and receive data from Voice Gateway's SIP signaling IPs.
    2. In the SIP headers, you need to provide identifiers that match the identifiers of the Agent Copilot workspace. This approach ensures that the transcripts are correctly transferred from the SIPREC call to the Agent Copilot workspace. For example, if you use agent ID and conversation ID in the contact center, configure the SIP headers as follows:

    ```text theme={null}
    X-Session-ID = {conversationId}
    X-User-ID = {agentId}
    ```

    **Configure a Voice Gateway Application**

    1. In the Voice Gateway Self-Service Portal, configure an [application](/voice-gateway/webapp/applications) to receive the call transcripts.
    2. In **Accounts**, select the account you want to use for SIPREC calls and select the application you've just configured in the **Application for SIPREC calls** list.

    **Configure Agent Copilot**

    To ensure that the [Agent Copilot workspace](/agent-copilot/overview) receives the transcripts from the SIPREC call, the identifiers of the Agent Copilot workspace URL and of the SIP headers must match. For example, if you use agent ID and conversation ID in the SIP headers, the Agent Copilot URL should be as follows:

    ```text theme={null}
    ${AgentCopilotBaseUrl}/?userId=${agentId}&sessionId=${conversationId}&URLToken=${URLToken}
    ```
  </Tab>
</Tabs>

## More Information

* [Agent Copilot](/agent-copilot/overview)
* [Voice Gateway](/voice-gateway/overview)


# Cognigy Voice Gateway
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/voice-gateway



<a href="/release-notes/2026.16"><Badge>Updated in 2026.16.0</Badge></a>

<Frame>
  <img alt="Voice Gateway Endpoint configuration in Cognigy" />
</Frame>

The Cognigy Voice Gateway Endpoint enables you to connect your AI Agent to your contact center or other phone numbers using [Cognigy Voice Gateway](/voice-gateway/index).
This Endpoint allows your customers to speak with your AI Agents instead of using a chat.

## Prerequisites

* Configured Flow with [Voice Gateway Nodes](/ai/agents/develop/node-reference/voice/overview).
* To route your contact center or phone number to your Voice Gateway Endpoint, contact Cognigy support.

## Generic Endpoint Settings

Learn about the generic Endpoint settings available for this Endpoint on the following pages:

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Transformer Functions](/ai/for-developers/transformers/overview)
* [NLU Connectors](/ai/platform-features/nlu/external/nlu-connectors/overview)
* [Session Management](/ai/agents/deploy/endpoints/session-management)
* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)
* [Copilot](/agent-copilot/configure/copilot-section)

## Specific Endpoint Settings

<AccordionGroup>
  <Accordion title="Click To Call Widget Settings">
    The Voice Gateway Endpoint includes the Click To Call embedding HTML code and the **Click To Call Widget Settings** section for configuring [Click To Call](/click-to-call/overview). Click To Call lets you embed a multimodal widget into your website, customer portal or mobile app so customers can be connected through one click to your voice AI Agent.

    | Parameter                   | Type   | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
    | --------------------------- | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Enable Click To Call Widget | Toggle | Activates the Click To Call widget for the Endpoint. This setting is active by default.                                                                                                                                                                                                                                                                                                                                                                                         |
    | AI Agent Name               | String | Sets the name displayed on the Click To Call widget.                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | Tagline                     | String | Sets the subtext displayed under the AI Agent name.                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | AI Agent Avatar Logo URL    | String | Sets the URL to the image displayed on the Click To Call widget. If left empty, the widget uses the Cognigy logo.                                                                                                                                                                                                                                                                                                                                                               |
    | Theme                       | String | The color theme of the Click To Call widget. You can select the following options:<ul><li>**Clean white** — the widget's background is white and the image background is light gray.</li><li>**Dark mode** — the widget's background is black and the image background is dark gray. This option is selected by default.</li><li>**AI Purple** — the widget's background is dark purple and the image background has a lighter purple shade.</li></ul>                          |
    | Display Live Transcription  | Toggle | Displays conversation transcript in messages in the Click To Call widget. By default, the toggle is activated.                                                                                                                                                                                                                                                                                                                                                                  |
    | Enable Privacy Notice       | Toggle | Displays a privacy notice at the beginning of each chat session to inform users about data handling practices and prompts the users to accept or decline the privacy policy. Accepting the privacy policy is required to start the conversation, ensuring compliance and user trust. If the user declines the privacy policy, the conversation doesn't start.                                                                                                                   |
    | Privacy Notice Title        | String | This parameter is displayed only if you toggle on **Enable Privacy Notice**. Sets the heading that appears at the top of the privacy notice screen. By default, it is set to `Privacy Notice`.                                                                                                                                                                                                                                                                                  |
    | Privacy Notice Text         | String | This parameter is displayed only if you toggle on **Enable Privacy Notice**. Sets the message that users see in the privacy notice prompt. This parameter supports Markdown syntax. The default text is `Please accept our privacy policy to start your chat`. You can customize this message to provide more detailed information about your data collection, usage practices, and users' rights, ensuring clarity and building trust from the outset of the chat interaction. |
    | Submit Button               | String | This parameter is displayed only if you toggle on **Enable Privacy Notice**. Sets the text displayed on the button users click to accept the Privacy Notice. By default, the button is labeled `Submit`.                                                                                                                                                                                                                                                                        |
    | Cancel Button Text          | String | This parameter is displayed only if you toggle on **Enable Privacy Notice**. Sets the text displayed on the button users click to decline the Privacy Notice. By default, the button is labeled `Cancel`.                                                                                                                                                                                                                                                                       |
    | Privacy Policy Link Title   | String | This parameter is displayed only if you toggle on **Enable Privacy Notice**. Sets the text that will be displayed as the title for the Privacy Policy link. The link title allows you to tailor the label of the link to better match the tone and context of your brand.                                                                                                                                                                                                       |
    | Privacy Policy Link URL     | String | This parameter is displayed only if you toggle on **Enable Privacy Notice**. Sets the URL that directs users to your privacy policy.                                                                                                                                                                                                                                                                                                                                            |

    <Accordion title="Customization">
      | Parameter                             | Type         | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
      | ------------------------------------- | ------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
      | Live Transcription Background         | Selector     | Sets whether a background is displayed for the live transcription messages. This parameter is only available if you toggle on **Display Live Transcription**. You can select the following options:<ul><li>**Transparent** — the live transcription messages pop and hover above the Click To Call widget without a chat box. This option is selected by default. If you use this option, make sure the website where the Click To Call widget and the transcription messages are visually adequate. </li><li>**Custom** — provides a solid background for the live transcription messages.</li></ul> |
      | Transcription Background Color        | Color picker | This parameter is only available if you select **Custom** in the **Live Transcription Background** selector. Sets the background color for the live transcription messages. By default, the background is white.                                                                                                                                                                                                                                                                                                                                                                                      |
      | Demo Widget Page Background           | Selector     | Sets the parameter to fill the page background when using the Demo Widget page. You can use this field to test your website's background against the Demo Widget. You can select the following options:<ul><li>**Color** — provides the **Demo Widget Page Background Color**. This option is selected by default. </li><li>**Image URL** — displays an image as the background for the Demo Widget.</li></ul>                                                                                                                                                                                        |
      | Demo Widget Page Background Color     | Color picker | This parameter is only available if you select **Color** in the **Demo Widget Page Background** selector. Sets the background color for the Demo Widget. By default, the background is white.                                                                                                                                                                                                                                                                                                                                                                                                         |
      | Demo Widget Page Background Image URL | String       | This parameter is only available if you select **Image URL** in the **Demo Widget Page Background** selector. Sets the URL to the image displayed as the background for the Demo Widget.                                                                                                                                                                                                                                                                                                                                                                                                              |
      | Center Click To Call Widget           | Selector     | Sets the parameter to position the Click To Call widget on the page. You can select the following options:<ul><li>**Center** — the Click To Call widget is centered on the page. This option is selected by default. </li><li>**Bottom Right** — the Click To Call widget is positioned at the bottom right of the page.</li></ul>                                                                                                                                                                                                                                                                    |
    </Accordion>
  </Accordion>

  <Accordion title="Generic Settings">
    These settings will be valid for every session of this Endpoint.

    | Parameter                           | Type   | Description                                                                                                                                                                                                                                                                                      |
    | ----------------------------------- | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Prosody Settings                    | Toggle | If the parameter is enabled, this configuration will be used to specify changes to speed, pitch, and volume for the text-to-speech output. The Set Session Config or Say Nodes with **Activity Parameters** enabled can override the **Prosody Settings** option configured within the Endpoint. |
    | Output Speed                        | Number | This parameter is active only when the **Prosody Settings** toggle is enabled. The change in the speaking rate for the contained text in percentage. Value range: -100 to 200.                                                                                                                   |
    | Output Pitch                        | Number | This parameter is active only when **Prosody Settings** toggle is enabled. The baseline pitch for the contained text in percentage. Value range: -100 to 200.                                                                                                                                    |
    | Output Volume                       | Number | This parameter is active only when **Prosody Settings** toggle is enabled. The volume for the contained text in percentage. Value range: 10 to 200.                                                                                                                                              |
    | Show Best Transcripts Only          | Toggle | If the parameter is enabled, only the best transcript in the input object is accessible, instead of all variations.                                                                                                                                                                              |
    | Enable Call Event: Call in Progress | Toggle | If the parameter is enabled, the [Call in Progress](/voice-gateway/references/events/CALL_IN_PROGRESS) toggle appears in the Call Events section.                                                                                                                                                |

    <Note>As with all Speech Provider settings, there are certain differences in the use of this feature. Refer to your Speech Provider's documentation first to understand which values to add. An example would be the difference between Microsoft Azure and Google. For Microsoft Azure, users need to specify the `%` by which they would like to increase or decrease from the default settings. For instance, an example value could be `Output Speed -20%`, indicating a `20%` reduction from the default output speed of `100%`, resulting in a new output speed of `80%`. To achieve the same effect with Google, users need to add the final output speed in `%` to the Prosody Settings. For example, `Output Speed 80%` directly displays the new output speed, as with Google; the final output speed is directly added to the settings.</Note>
  </Accordion>

  <Accordion title="AudioCodes Compatibility Mode">
    Allows using both AudioCodes Nodes and Voice Gateway Nodes simultaneously in a Flow.
    When this mode is enabled, the system treats Flows with AudioCodes Nodes as if they were Flows with Voice Gateway Nodes.
    This mode helps ensure that voice AI Agents operate smoothly without interruptions during the transition from AudioCodes to Cognigy Voice Gateway.

    To activate the **AudioCodes Compatibility Mode** section, add the `FEATURE_VG_AC_COMPATIBILITY_MODE` feature flag to the Cognigy `values.yaml` file.

    For more information, read [Migration from AudioCodes to Voice Gateway](/voice-gateway/migrate-from-ac-to-vg).
  </Accordion>

  <Accordion title="Call Events">
    Allows activating call events for a Flow.
    Select a call event from the [Voice Gateway Events](/voice-gateway/references/events/overview) list.
    This event will trigger the action.

    If you have configured the same call event in both the Endpoint and the [Lookup](/ai/agents/develop/node-reference/logic/lookup) Node, the Endpoint settings will overwrite the Node settings.

    **Call Event Settings**

    <Note>As with all other Endpoint settings, you can't test the Call Events settings within the Interaction Panel.</Note>

    | Parameter    | Type     | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | ------------ | -------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Action       | Dropdown | Select the action to be performed when the call event is detected: <br /> - **Inject into current Flow** — inject the defined text and data payload into the current flow. <br /> - **Execute Flow** — trigger a selected Flow when the call event is detected. <br /> - **None** — no action will be taken when the call event is detected. <br /> - **Transfer** — transfer a call in case [TRANSFER\_DIAL\_ERROR](/voice-gateway/references/events/TRANSFER_DIAL_ERROR) or [TRANSFER\_REFER\_ERROR](/voice-gateway/references/events/TRANSFER_REFER_ERROR) call events occur. This action for call events is configured similarly to Call Failover. |
    | Text Payload | String   | Enter the text that will be sent to your Flow. The parameter is available only for the **Inject into current Flow** action.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | Data Payload | JSON     | Provide the data that will be sent into your Flow in JSON format. The parameter is available only for the **Inject into current Flow** action.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         |
    | Execute Flow | Dropdown | Execute the selected Flow. The parameter is available only for the **Execute Flow** action. <br /><br /> Note that you can't use [Go To](/ai/agents/develop/node-reference/logic/go-to) or [Execute Flow](/ai/agents/develop/node-reference/logic/execute-flow) Nodes in the executed Flow. When the Flow is fully processed, with all Nodes from Start to End executed, any dependent Flows won't be triggered.                                                                                                                                                                                                                                       |
  </Accordion>

  <Accordion title="Call Failover">
    The **Call Failover** section is intended to handle runtime and speech provider authentication errors that occur on the Cognigy side during the execution of a Flow.

    <Note>
      For 2026.14.1 or earlier versions:

      * If you use a SaaS Cognigy installation, contact the support team to activate this feature.
      * If you use an on-premises Cognigy installation, activate this feature by adding `FEATURE_ENABLE_ENDPOINT_CALL_FAILOVER` in `values.yaml`.
    </Note>

    <Note>
      For Voice Gateway 2025.19 and later, the Anchor Media parameter has been renamed to Media Path and includes more options. The previously available options correspond to:

      * **Full Media** — the toggled on Anchor Media parameter in earlier versions.
      * **Partial Media** — the toggled off Anchor Media parameter in earlier versions.

      The Media Path parameter is hidden behind the feature flag `FEATURE_DISABLE_VG_MEDIA_PATH`, with a default value of `false`. It is available only in environments where the feature flag is set to `false`.
    </Note>

    | Parameter                           | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           | Transfer Type |
    | ----------------------------------- | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------- |
    | Flow execution failover enabled     | Toggle        | If enabled, the configuration below will be used to perform a transfer in case of a runtime error. If the **Speech provider failover** toggle is also enabled, applies to cases of runtime errors or speech provider failures.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        | -             |
    | Speech provider failover enabled    | Toggle        | If enabled, the configuration below will be used to perform a transfer only if speech providers fail:<ul><li>If only a primary speech provider is configured, the call failover occurs after the primary speech provider fails.</li><li>If both primary and fallback speech providers are configured, the call failover occurs only after both speech providers fail.</li></ul> If the **Flow execution failover** toggle is also enabled, the configuration below applies to cases of runtime errors or speech provider failures.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    | -             |
    | Transfer Type                       | Dropdown      | There are two transfer types: <br /> - **Refer** — forwarding an existing call. <br /> - **Dial** — creating a new outgoing call. If you want to use this type and still have the old Node version, add a new Voice Gateway Transfer Node in the Flow Editor and manually transfer the required settings from the old Node.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           | -             |
    | Reason                              | CognigyScript | The reason for the handover. It is shown in Voice Gateway logs.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       | All           |
    | Target                              | CognigyScript | E.164 syntax or a SIP URI are supported.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              | All           |
    | Caller ID                           | Number        | The caller ID. Some carriers, like Twilio, require a registered number for outgoing calls.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            | Dial          |
    | Dial Music                          | URL           | Custom audio or ring-back which plays to the caller while the outbound call is ringing. Only `.wav` or `.mp3` files are supported. The URL doesn't need to include the `.mp3` or `.wav` extension. For example, `https://abc.xyz/music.mp3` or `https://audio.jukehost.co.uk/N5pnlULbup8KabGRE7dsGwHTeIZAwWdr`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       | Dial          |
    | Timeout                             | Number        | The amount of time (in seconds) that the AI Agent will ring before a [no-answer](/voice-gateway/references/events/NO_ANSWER) timeout. The default value is 60 seconds.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                | Dial          |
    | Enable Duration Limit for Transfer  | Toggle        | Automatically disconnects transferred calls after a set time to prevent excessive call durations. This setting is useful in the following cases:<br />- Limit call duration by automatically ending transferred calls that exceed the predefined time, helping manage system performance.<br />- Free up resources by disconnecting calls if the caller forgets to hang up, making lines available for other callers.<br />- Reduce costs by preventing unnecessary charges from long, unattended calls.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              | Dial          |
    | Duration Limit                      | Number        | This parameter is active only when **Enable Duration Limit for Transfer** is selected. Set the maximum duration in seconds. Transferred calls end automatically after this time, even if the caller is on the line. The default value is `600`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       | Dial          |
    | Enable Copilot                      | Toggle        | Creates the UUIValue which will be sent to the Contact Center through SIP Headers. Will collect information from the Transcription Webhook field as well as from the VG Endpoint Copilot Config field to create the UUIValue. This setting requires a configured [Voice Copilot](/ai/agents/deploy/endpoint-reference/voice-copilot) Endpoint.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        | Dial          |
    | Failover Transcribe Enabled         | Toggle        | If enabled, transcriptions will be attempted in case of a failed call transfer.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       | Dial          |
    | Media Path                          | Dropdown      | Controls the routing of RTP traffic through a media platform, such as FreeSwitch, for monitoring, transcoding, and security purposes. You can also [change this value during an active call](#changing-the-media-path-during-an-active-call) using the `updateCall` API. Select one of the following options:<ul><li>**Full Media** — routes all RTP traffic through a media platform. This option is used when complete media capture is required, for example, for logging, monitoring, or other processing workflows.</li><li>**Partial Media** — this option is selected by default. Routes through a media platform only when necessary. For example, to enable transcriptions or other optional processing.</li><li>**No Media** – routes directly from the caller's SBC to the far SBC or SIP trunk, without passing any media through the media platform. This option is primarily used for compliance purposes. When the option is selected, the system is disconnected from the media stream entirely, ensuring that sensitive data, such as credit card information, can't be captured.</li></ul> When you select **Full Media** or **Partial Media**, media anchoring process changes the addresses and ports of incoming and outgoing data:<ul><li>Outgoing data — the source IP address and port become the IP address and port of Voice Gateway. The destination IP address and port become the IP address and port of the public network. </li><li>Incoming data — the source IP address and port become the IP address and port of Voice Gateway. The destination IP address and port become the IP address and port of the PBX system or local telephone.</li></ul> | Dial          |
    | STT Vendor                          | Dropdown      | Defines the STT vendor. You can select a custom vendor.<br /><br />The **Default** option's behavior in the Set Session Config Node depends on whether you select this option in the first or in a subsequent Node of this type:<ul><li>**First Set Session Config Node** — switches the STT vendor to the default configuration set in [Voice Gateway Self-Service Portal](/voice-gateway/webapp/applications#add-additional-tts-and-stt-vendor)</li><li>**Subsequent Set Session Config Node** — keeps the TTS vendor set in the previous Set Session Config Node, allowing you to change other parameters without overwriting the STT vendor.</li></ul> The ElevenLabs STT vendor is hidden behind a feature flag. Activate ElevenLabs on the Cognigy side: <ul><li>If you have a SaaS installation, contact [Cognigy technical support](/help/get-help). </li><li>If you have an on-premises installation, specify the `FEATURE_ENABLE_ELEVENLABS_STT_VENDOR=true` in the `values.yaml`.</li></ul>The Speechmatics STT vendor is hidden behind a feature flag. Activate Speechmatics on the Cognigy and Voice Gateway sides:<ul><li>If you have a SaaS installation, contact [Cognigy technical support](/help/get-help).</li><li>If you have an on-premises installation, specify the [feature flags](/voice-gateway/webapp/speech-services#speechmatics) in `values.yaml`.</li></ul>                                                                                                                                                                                                                                                                                            | Dial          |
    | STT Language                        | Dropdown      | Select the desired STT Language. For custom languages, use the following format: de-DE, fr-FR, en-US.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 | Dial          |
    | Disable STT Punctuation             | Toggle        | This parameter is active only when Google or Deepgram is selected from the **STT Vendor** list.<br /><br /> Prevents the transcription returned to the AI Agent from including punctuation marks.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     | Dial          |
    | Deepgram Model                      | Dropdown      | This parameter is active only when Deepgram is selected from the **STT Vendor** list. <br /><br /> Choose a model for processing submitted audio. Each model is associated with a tier. Ensure that the selected tier is available for the chosen STT language. For detailed information about Deepgram models, refer to the [Deepgram documentation](https://developers.deepgram.com/docs/model).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    | Dial          |
    | Endpointing                         | Toggle        | This parameter is active only if Deepgram or Deepgram Flux is selected from the **STT Vendor** list: <ul><li>**Deepgram** – detects the end of speech by monitoring user pauses. When a pause is detected, Deepgram finalizes the transcript with `speech_final = true`. The duration for detecting the end of speech is preconfigured with the default value of 10 milliseconds. If you want to change this value, use the **Endpointing Time** parameter.</li><li>**Deepgram Flux** – detects the end of speech using a confidence threshold. When a user's speech ends or the optional timeout is reached, the transcript is finalized. This approach ensures accurate, responsive multiturn transcription.</li></ul> For detailed information about the endpointing configuration, refer to the documentation for [Deepgram](https://developers.deepgram.com/docs/endpointing) or [Deepgram Flux](https://developers.deepgram.com/docs/flux/configuration).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       | Dial          |
    | Endpointing Time                    | Number        | This parameter is active only when Deepgram is selected from the **STT Vendor** list and the **Endpointing** toggle is enabled. <br /><br /> Customize the duration (in milliseconds) for detecting the end of speech. The default is 10 milliseconds of silence. After detecting silence, the system waits until either the speaker resumes or the required silence duration is reached. Once one of these conditions is met, the transcript is sent with `speech_final` set to `true`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              | Dial          |
    | Smart Formatting                    | Toggle        | This parameter is active only when Deepgram is selected from the **STT Vendor** list. <br /><br /> Deepgram's Smart Format feature applies additional formatting to transcripts to optimize them for human readability. Smart Format capabilities vary between models. When Smart Formatting is turned on, Deepgram will always apply the best-available formatting for your chosen model, tier, and language combination. For detailed examples, refer to the [Deepgram documentation](https://developers.deepgram.com/docs/smart-format). <br /><br /> Note that when Smart Formatting is turned on, punctuation will be activated, even if you have the Disable STT Punctuation setting enabled.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   | Dial          |
    | Google Model                        | Dropdown      | This parameter is active only when Google is selected from the **STT Vendor** list.<br /><br />Uses one of Google Cloud Speech-to-Text transcription models. The default value is `latest_short`. For a detailed list of Google models, refer to the [Transcription models](https://cloud.google.com/speech-to-text/docs/transcription-model#transcription_models) section in the Google Documentation. Keep in mind that the `default` value is a Google model type that can be used if other models don't suit your use case.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       | Dial          |
    | End of Turn Threshold               | Slider        | This parameter is active only if Deepgram Flux is selected from the **STT Vendor** list. Defines the confidence level for detecting when a user's speech ends. Higher values improve accuracy but may increase latency. The default value is `0.7`, meaning the system waits until it is 70% confident that the user has finished speaking before finalizing the transcript.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          | Dial          |
    | End of Turn Timeout in Milliseconds | Number        | This parameter is active only if Deepgram Flux is selected from the **STT Vendor** list. Defines the maximum number of milliseconds to wait before finalizing a turn. The timeout acts as a backup if the user stops speaking unexpectedly. The default value is `5000` (5 seconds), meaning the turn will be closed after 5 seconds of silence.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      | Dial          |
    | Transcription Webhook               | CognigyScript | The webhook is triggered with an HTTP POST whenever an interim or final transcription is received. If the **STT Vendor** and **STT Language** fields are empty, the system will use the default STT from the [Set Session Config Node](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details) (if it exists) or from the [Voice Gateway Self-Service Portal](/voice-gateway/webapp/speech-services). The parameter supports [CognigyScript](/ai/platform-features/cognigyscript), allowing it to accept dynamic content. For example, you can specify the URL as follows: `https://test-hook.com?contact={{ci.contact_name}}`. <br /><br /> Note that if the [Voice Copilot](/ai/agents/deploy/endpoint-reference/voice-copilot) Endpoint is inactive, you can use any Webhook URL to receive voice call transcripts. However, when the Voice Copilot Endpoint is enabled, ensure that the specified Webhook URL is associated with it for processing.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              | Dial          |
    | STT Label                           | CognigyScript | The alternative name of the vendor is the one you [specify in the Voice Gateway Self-Service Portal](/voice-gateway/webapp/applications#add-additional-tts-and-stt-vendor). If you have created multiple speech services from the same vendor, use the label to specify which service to use.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         | Dial          |
    | Audio Stream Selection              | Dropdown      | Select the source of the audio stream: <br /> - **Caller/Called** - both the incoming and outgoing audio streams of the caller and the called party. <br /> - **Caller** - the incoming and outgoing audio stream of the caller. <br />- **Called** - the incoming and outgoing audio stream of the called party. <br /> <br /> Ensure that the selected audio stream matches the language specified for transcription. If no audio stream is provided, the system will use the one set in the beginning, which should also match the language specified for transcription.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           | Dial          |
    | Referred By                         | String        | This parameter is optional.<br /><br /> This setting allows you to change the original Referred By value, which can be a SIP URI or a user identifier such as a phone number. To define the Referred By value, you can use the following patterns:<br />- **SIP URI** - `sip:[referred-by]@custom.domain.com`. In this case, the entire SIP URI will be sent as the Referred-By header. Example: `"Referred-by": "sip:CognigyOutbound@custom.domain.com"`.<br />- **User Identifier** - `sip:[referred-by]@[SIP FROM Domain from carrier config]`. Example, `"Referred-By": "sip:CognigyOutbound@sip.cognigy.ai"`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    | Refer         |
    | Custom Transfer SIP Headers         | Toggle        | Data that needs to be sent as SIP headers in the generated SIP message.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               | All           |

    ### Changing the Media Path during an Active Call

    Besides setting the media path in the Flow, you can also change the media path during an active transferred call using the Voice Gateway API.

    This approach is useful for PCI compliance. For example, when a caller shares sensitive payment information, Voice Gateway, as an intermediate system, must be removed from the media stream. Audio is then routed directly between the caller and the contact center, bypassing Voice Gateway, reducing PCI data exposure and helping protect sensitive data.

    To remove Voice Gateway from the media path, send the [PUT /Accounts//Calls/](https://api-vg-trial.cognigy.ai/docs/#/Accounts/updateCallPut) API request with `{ "media_path": "noMedia" }` from the contact center to Voice Gateway after the call has been transferred. With this setting, Voice Gateway stops processing the call's audio and DTMF entirely, including for call transcription. After Voice Gateway confirms the API request with a `200` or `202` response, you can't revert to processing media for the transferred call. A `400`, `404`, or `422` response means the request failed, which keeps the call on its previous media path, and Voice Gateway continues processing audio and DTMF.

    Capture the call SID before the transfer takes place. During a transfer, a new call SID is created for the new call leg, but the original call SID remains unchanged. Pass the original call SID along during the transfer, for example, in SIP headers. This way, the contact center can reference the original call when triggering the API request at the correct moment.
  </Accordion>
</AccordionGroup>

### Additional Information

<AccordionGroup>
  <Accordion title="SIP Headers">
    The SIP headers, including any custom headers, are available within the [Input](/ai/agents/develop/ai-agent-memory/input) object.
    You can find them in `input.data` or `input.data.payload.sip.headers`.
  </Accordion>

  <Accordion title="Call Metadata">
    Voice Gateway identifies information about the caller and adds it to the [Input](/ai/agents/develop/ai-agent-memory/input) object as `input.data.numberMetaData`.

    | Parameter          | Type    | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            | Example            |
    | ------------------ | ------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------ |
    | number             | String  | The caller's phone number, including country code. The phone number has E.164 format.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  | +4921154591991     |
    | country            | String  | The 2-character country code.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          | DE                 |
    | countryCallingCode | String  | The country calling code associated with the phone number.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             | 49                 |
    | nationalNumber     | String  | The national number part of the phone number. It excludes the country code and any leading zero.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       | 21154591991        |
    | ext                | String  | The extension number, if present. It consists of a numerical sequence, such as 123 or 568, used in PBX systems for streamlined internal phone communications within organizations. Users can dial these numbers for internal calls without needing to dial the full external phone number.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             | 123                |
    | valid              | Boolean | The validity of the phone number.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      | true               |
    | type               | String  | The phone number's classification. Can be any of: <br />- `PREMIUM_RATE` — a number associated with premium rate services, incurring higher charges. <br />- `TOLL_FREE` — a toll-free number where the recipient pays for the call. <br />- `SHARED_COST` — a number where call costs are shared between the caller and the recipient. <br />- `VOIP` — a number associated with Voice over Internet Protocol (VoIP) services. <br />- `PERSONAL_NUMBER` — a non-geographic number typically associated with individuals. <br />- `PAGER` — a number associated with a pager device for receiving short messages. <br />- `UAN` — a universal Access Number for nationwide business services. <br />- `VOICEMAIL` — a number used for accessing voicemail services. <br />- `FIXED_LINE_OR_MOBILE` — a fixed-line or mobile number. <br />- `FIXED_LINE` — a traditional landline phone number. <br />- `MOBILE` — a number associated with a mobile/cellular device. | FIXED\_LINE        |
    | uri                | String  | The URI associated with the phone number, if applicable.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               | tel:+4921154591991 |

    <Note>The data in `input.data.numberMetaData` is available as [Tokens](/ai/platform-features/tokens) inside Cognigy Text fields.</Note>
  </Accordion>

  <Accordion title="Handover">
    You can use any [handover provider](/ai/escalate/handover-reference/overview) within the Voice Gateway integration.
    Handover enables your voice agent to route calls to the contact center, ensuring efficient communication and swift access to human assistance.

    In contrast to chatting, communication between actors during a voice call occurs using speech recognition technologies (text-to-speech and speech-to-text).
    Using speech-to-text technology, the voice agent sends the conversation script from the end user to the human agent.
    Once the script is received, the human agent can respond using text.
    After that, the voice agent converts the human agent's message to speech using text-to-speech technology.
    Finally, the converted message is sent to the end user.

    **Configuration**

    To enable handover, add a [Handover to Human Agent](/ai/agents/develop/node-reference/service/handover-to-agent) Node in your voice Flow. This Node transfers the conversation from the voice agent to the human agent.

    In this example, an end user seeks billing assistance and interacts with agents in a customer service system via text-to-speech and speech-to-text technologies.

    **End User:** Calls the customer service line. Speaks to the voice agent. Example: `Hello, I need assistance with my billing inquiry`.<br />
    **Voice Agent:** Transcribes user's speech into text using speech-to-text technology. Example: `Hello, I need assistance with my billing inquiry`.<br />
    **Voice Agent:** Determines the need for human assistance. Initiates handover to the human agent via handover provider.<br />
    **Human Agent:** Receives a conversation script containing user's query.<br />
    **Human Agent:** Types response on interface. Example: `Hello, thank you for reaching out. Let me check your billing details`.<br />
    **Voice Agent:** Receives human agent's typed response. Converts the text message into speech using text-to-speech technology.<br />
    **End User:** Hears human agent's response spoken by voice agent. Example: `Hello, thank you for reaching out. Let me check your billing details`.<br />
    **End User / Human Agent:** The conversation between the end user and the human agent continues until the issue is resolved or the interaction is concluded.

    **Multilingual Support**

    During a handover, you can use translation into other languages.
    Use the translation feature to overcome language barriers between end users and support staff. This feature is particularly useful in situations where there are no support staff proficient in the end user's language, and urgent issue resolution is necessary.

    For example, if a user communicates in German, the voice agent responds accordingly.
    However, after transferring the call to the contact center, the dialogue is translated into English.
    The human agent responds in English, and the response is translated into German for the end user via text-to-speech technology.
    This method ensures seamless communication despite language differences.

    Note that depending on the language and provider selected,
    translations may experience delays while they reach the contact center.
    If the speech-to-text fails to recognize speech for any reason, the message won't appear on the contact center's side.
    Plan for such scenarios within your Flow and configure the voice agent's behavior accordingly.
    This approach will ensure that the end user is aware that their message may not be delivered.
    You can achieve this by adjusting timeout settings and error handling policies.

    To set up a multilingual Flow for the handover use case, follow these steps:

    1. In your voice Flow, navigate to the [Set Session Config](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details) Node.
    2. In the Node editor, go to the **Recognizer STT** section and activate the **Recognize Language** toggle.
    3. From the **Alternative Language** list, select the additional languages, for example, `German` and `Italian`.
    4. Below the Set Session Config Node, add a [Lookup](/ai/agents/develop/node-reference/logic/lookup) Node.
    5. In the parent Lookup Node, select **CognigyScript** from the **Type** list and specify `input.data.payload.speech.language_code` in the **Operator** field.
    6. In the case Nodes of the Lookup Node, specify the language codes, for example, `de-DE`, `it-IT`.
    7. Below each case Node, add the second [Set Session Config](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details) Nodes.
    8. Configure the TTS and STT settings for each of the Set Session Config Nodes. This means that one Node should have German language selected, while the other should have Italian. These Nodes will manage the voice agent's communication in the end user's language.
    9. Below each Set Session Config Node, add the [Set Translation](/ai/agents/develop/node-reference/other-nodes/set-translation) Nodes. These Nodes will perform translation of a text from the end user's language to the human agent's language.
    10. In the Set Translation Node, activate the **Translation Enabled** toggle.
    11. In the **User Input Language** field, specify the language of the user input, for example, `de`, `it`.
    12. In the **Flow Language** field, specify the language to translate to, for example, `en`.
    13. At the end of the Flow, place the [Handover to Human Agent](/ai/agents/develop/node-reference/service/handover-to-agent) Node.

    <Frame>
      <img alt="Voice Gateway handover Flow configuration example" />
    </Frame>

    In this example, an end user seeks billing assistance and interacts with actors in a customer service system via text-to-speech and speech-to-text technologies. The end user and human agent speak different languages, but this doesn't hinder their understanding, as the voice agent translates their speech into the required language for each party.

    **End User:** Calls the customer service line. Speaks to the voice agent. Example: `Hallo, ich brauche Hilfe bei meiner Rechnungsanfrage`.<br />
    **Voice Agent:** Determines the language and answers using the same language. Example: `Danke für die Informationen. Ich werde Sie an einen menschlichen Agenten weiterleiten, der Ihnen weiterhelfen kann`.<br />
    **Voice Agent:** Transcribes user's speech into text using speech-to-text technology. Example: `Hello, I need assistance with my billing inquiry`.<br />
    **Voice Agent:** Determines the need for human assistance and initiates the handover to the human agent via the handover provider.<br />
    **Human Agent:** Receives a conversation script containing user's query.<br />
    **Human Agent:** Types response on interface. Example: `I've found the issue. It seems there was an error in the billing statement. I'll correct it for you`.<br />
    **Voice Agent:** Receives human agent's typed response. Converts the text message into speech using text-to-speech technology.<br />
    **End User:** Hears human agent's response spoken by voice agent. Example: `Ich habe das Problem gefunden. Es scheint, dass es einen Fehler in der Rechnungsstellung gab. Ich werde es für dich korrigieren`.<br />
    **End User/Human Agent:** The conversation between the end user and the human agent continues until the issue is resolved or the interaction is concluded.
  </Accordion>
</AccordionGroup>

## More Information

* [Voice Gateway](/voice-gateway/index)
* [Voice Gateway Nodes](/ai/agents/develop/node-reference/voice/voice-gateway/overview)


# Cognigy Webchat
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/webchat



Cognigy Webchat is an effective communication tool that assists businesses in interacting with their customers in a seamless manner.
Its main function is to facilitate real-time communication through text or voice messages, which aids businesses in providing immediate support, collecting feedback, and engaging with users effectively. Webchat can be easily embedded into websites, enabling visitors to communicate with conversational AI Agents powered by Cognigy.

Select the Endpoint version you want to use:

* [Webchat v3](/webchat/v3/configuration)
* [Webchat v2](/webchat/v2/configuration)


# Webhook
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/webhook



<a href="/release-notes/2026.16"><Badge>Updated in 2026.16</Badge></a>

<Frame>
  <img alt="Webhook Endpoint logo" />
</Frame>

The Webhook Endpoint connects your AI Agent to external systems by sending real-time events, such as user or system messages, to a specified callback (webhook) URL. Additionally, you can configure API key-based authentication for requests to this Endpoint.

The Endpoint receives `POST` requests at the Cognigy Endpoint URL, processes them with the assigned Flow, and sends results asynchronously to your webhook URL.

```mermaid theme={null}
flowchart LR
A[External System] -->|POST Request| B[Cognigy Webhook Endpoint]
B -->|Processes| C[Assigned Flow]
C -->|Sends Results| D[External Webhook URL]
```

If you use [Agent Copilot for voice](/agent-copilot/getting-started/voice) with the Webhook Endpoint, you can switch to a specific [Voice Copilot](/ai/agents/deploy/endpoint-reference/voice-copilot) Endpoint. This Endpoint includes all webhook logic, so you don't need to use a Code Node.

## Prerequisites

* To receive `POST` requests from Cognigy at your webhook URL, run a web server on your side.
* *(Optional)* Set up basic authentication for your web server.

## Restrictions

* [Snapshots](/ai/agents/deploy/snapshots) and [Packages](/ai/platform-features/packages) don't support API keys generated in this Endpoint. If you export a Webhook Endpoint that includes an API key, generate a new API key in the target Project. Then update your code to include the new API key.

## Generic Endpoint Settings

Learn about the generic Endpoint settings on the following pages:

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [NLU Connectors](/ai/platform-features/nlu/external/nlu-connectors/overview)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)
* [Inject and Notify](/ai/agents/deploy/endpoints/inject-and-notify)
* [Copilot](/agent-copilot/configure/copilot-section)

## Specific Endpoint Settings

<AccordionGroup>
  <Accordion title="API Key Authentication">
    Use this section to configure how incoming requests to this Endpoint are authenticated.

    | Parameter             | Type   | Description                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | --------------------- | ------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Authentication Method | List   | Sets an authentication step for requests sent to this Endpoint. Select one of the options:<ul><li>**None** — no authentication step. This option is selected by default.</li><li>**API Key** — sets API key-based authentication. If you select this option, you need to enter an Endpoint-specific API key in the header of your requests to this Endpoint. To generate an Endpoint-specific API key, use the **Generate API Key** button.</li></ul> |
    | Generate API Key      | Button | This button appears in the **Endpoint API Keys** section when you select **API Key** in **Authentication Method**. Opens a dialog to generate an Endpoint-specific API key. You can generate multiple Endpoint-specific API keys.                                                                                                                                                                                                                     |
  </Accordion>

  <Accordion title="Basic Auth Credentials">
    Use this section to provide the webhook URL and, optionally, credentials Cognigy uses to authenticate outgoing requests to the webhook URL.

    | Parameter | Type | Description                                                                                                     |
    | --------- | ---- | --------------------------------------------------------------------------------------------------------------- |
    | User      | Text | This parameter is optional. Sets the username to authenticate against when sending requests to the webhook URL. |
    | Password  | Text | This parameter is optional. Sets the password to authenticate against when sending requests to the webhook URL. |
    | Webhook   | Text | Sets the webhook URL to send requests to.                                                                       |
  </Accordion>
</AccordionGroup>

## How to Set Up

### Setup on the Cognigy Side

<Accordion title="1. Create a Webhook Endpoint">
  1. In the left-side menu of your Project, go to **Deploy > Endpoints**, and click **+ New Endpoint**.
  2. In the **New Endpoint** section, do the following:
     1. Select the **Webhook** Endpoint type.
     2. Specify a unique name.
     3. Select a Flow from the list.
  3. *(Optional)* To set up Endpoint-specific API authentication, select **API Key** from the **Authentication Method** list in the **API Key Authentication** section. The **Endpoint API Keys** section appears. Follow these steps:
     1. Click **+ Generate API Key**.
     2. In the dialog box, enter a unique name for your API key and click **Generate API Key**.
     3. Click the **Key** field to copy the generated API key for later use in the header of requests to the Endpoint URL. Save the API key because you can't retrieve it after closing the dialog. Alternatively, you can create and manage API keys for this Endpoint using the `/v2.0/endpoints/{endpointId}/apikeys` routes via [Cognigy API](https://api-trial.cognigy.ai/openapi#tag--Endpoints).
  4. In the **Basic Auth Credentials** section, enter the external webhook URL in the **Webhook** field. This URL is where Cognigy sends output data.
  5. *(Optional)* If your webhook uses basic authentication, fill in the **User** and **Password** fields.
  6. Save changes and go to the **Configuration Information** section. For sending `POST` requests to the Cognigy Webhook Endpoint, copy the URL from the **Endpoint URL** field.
</Accordion>

### Setup on the Third-Party Provider Side

<AccordionGroup>
  <Accordion title="1. Send a Request">
    Send a `POST` request to the Cognigy Webhook Endpoint. Your web server must accept `POST` requests and process the JSON payload sent by Cognigy. For testing purposes, you can use [webhook.site](https://webhook.site) as a temporary web server.

    <AccordionGroup>
      <Accordion title="No Authentication">
        <Tabs>
          <Tab title="cURL">
            Replace `https://<your-endpoint-url>` with the Endpoint URL from the Endpoint settings.

            ```bash theme={null}
            curl -X POST https://<your-endpoint-url> \
              -H "Content-Type: application/json" \
              -d '{
                "userId": "user123",
                "sessionId": "session123",
                "text": "Hello, I need help with my order",
                "data": {
                  "exampleKey": "exampleValue"
                }
              }'
            ```
          </Tab>

          <Tab title="Postman">
            1. Open the Postman collection, select **Add a request**, then set the request type to `POST`.
            2. Enter the Endpoint URL as the request URL.
            3. Go to the **Headers** tab and add `Content-Type: application/json`.
            4. Go to the **Body** tab, select **raw**, then select **JSON** as the format.
            5. Paste the request body:

            ```json theme={null}
            {
              "userId": "user123",
              "sessionId": "session123",
              "text": "Hello, I need help with my order",
              "data": {
                "exampleKey": "exampleValue"
              }
            }
            ```

            | Parameter | Type   | Description                                                                                                                                                                                                 | Required                                             |
            | --------- | ------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------- |
            | userId    | String | The ID of the end user.                                                                                                                                                                                     | Yes                                                  |
            | sessionId | String | The ID used to track the current session and maintain its state. Generate a new unique ID for each new session. For testing, you can use any string and change it whenever you want to start a new session. | Yes                                                  |
            | text      | String | The text that the assigned Flow processes.                                                                                                                                                                  | No, if `data` is specified<sup>[1](#footnote1)</sup> |
            | data      | Object | The data that the assigned Flow processes.                                                                                                                                                                  | No, if `text` is specified<sup>[1](#footnote1)</sup> |
          </Tab>
        </Tabs>
      </Accordion>

      <Accordion title="With API Key Authentication">
        The API key authentication header name depends on your Cognigy version.

        <AccordionGroup>
          <Accordion title="For Cognigy 2026.16.0 and later version">
            Use the `x-cognigy-endpoint-key` header. The Endpoint also accepts the legacy `X-Webhook-Key` header for backward compatibility, but `x-cognigy-endpoint-key` takes precedence if both are provided.

            <Tabs>
              <Tab title="cURL">
                Replace `https://<your-endpoint-url>` with the Endpoint URL and `<your-api-key>` with an active API key from the Endpoint settings.

                ```bash theme={null}
                curl -X POST https://<your-endpoint-url> \
                  -H "Content-Type: application/json" \
                  -H "x-cognigy-endpoint-key: <your-api-key>" \
                  -d '{
                    "userId": "user123",
                    "sessionId": "session123",
                    "text": "Hello, I need help with my order",
                    "data": {
                      "exampleKey": "exampleValue"
                    }
                  }'
                ```
              </Tab>

              <Tab title="Postman">
                1. Open the Postman collection, select **Add a request**, then set the request type to `POST`.
                2. Enter the Endpoint URL as the request URL.
                3. Go to the **Headers** tab and configure:
                   * `Content-Type` — select `application/json`
                   * `x-cognigy-endpoint-key` — the API key you generated in the [API Key Authentication](#api-key-authentication) section.
                4. Go to the **Body** tab, select **raw**, then select **JSON** as the format.
                5. Paste the request body:

                ```json theme={null}
                {
                  "userId": "user123",
                  "sessionId": "session123",
                  "text": "Hello, I need help with my order",
                  "data": {
                    "exampleKey": "exampleValue"
                  }
                }
                ```

                | Parameter | Type   | Description                                                                                                                                                                                                 | Required                                             |
                | --------- | ------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------- |
                | userId    | String | The ID of the end user.                                                                                                                                                                                     | Yes                                                  |
                | sessionId | String | The ID used to track the current session and maintain its state. Generate a new unique ID for each new session. For testing, you can use any string and change it whenever you want to start a new session. | Yes                                                  |
                | text      | String | The text that the assigned Flow processes.                                                                                                                                                                  | No, if `data` is specified<sup>[1](#footnote1)</sup> |
                | data      | Object | The data that the assigned Flow processes.                                                                                                                                                                  | No, if `text` is specified<sup>[1](#footnote1)</sup> |
              </Tab>
            </Tabs>
          </Accordion>

          <Accordion title="For Cognigy 2026.13 to 2026.15">
            <Tabs>
              <Tab title="cURL">
                Replace `https://<your-endpoint-url>` with the Endpoint URL and `<your-api-key>` with an active API key from the Endpoint settings.

                ```bash theme={null}
                curl -X POST https://<your-endpoint-url> \
                  -H "Content-Type: application/json" \
                  -H "X-Webhook-Key: <your-api-key>" \
                  -d '{
                    "userId": "user123",
                    "sessionId": "session123",
                    "text": "Hello, I need help with my order",
                    "data": {
                      "exampleKey": "exampleValue"
                    }
                  }'
                ```
              </Tab>

              <Tab title="Postman">
                1. Open the Postman collection, select **Add a request**, then set the request type to `POST`.
                2. Enter the Endpoint URL as the request URL.
                3. Go to the **Headers** tab and configure:
                   * `Content-Type` — select `application/json`
                   * `X-Webhook-Key` — the API key you generated in the [API Key Authentication](#api-key-authentication) section.
                4. Go to the **Body** tab, select **raw**, then select **JSON** as the format.
                5. Paste the request body:

                ```json theme={null}
                {
                  "userId": "user123",
                  "sessionId": "session123",
                  "text": "Hello, I need help with my order",
                  "data": {
                    "exampleKey": "exampleValue"
                  }
                }
                ```

                | Parameter | Type   | Description                                                                                                                                                                                                 | Required                                             |
                | --------- | ------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------- |
                | userId    | String | The ID of the end user.                                                                                                                                                                                     | Yes                                                  |
                | sessionId | String | The ID used to track the current session and maintain its state. Generate a new unique ID for each new session. For testing, you can use any string and change it whenever you want to start a new session. | Yes                                                  |
                | text      | String | The text that the assigned Flow processes.                                                                                                                                                                  | No, if `data` is specified<sup>[1](#footnote1)</sup> |
                | data      | Object | The data that the assigned Flow processes.                                                                                                                                                                  | No, if `text` is specified<sup>[1](#footnote1)</sup> |
              </Tab>
            </Tabs>
          </Accordion>
        </AccordionGroup>
      </Accordion>
    </AccordionGroup>
  </Accordion>

  <Accordion title="2. Get a Response">
    The Webhook Endpoint sends the following JSON response to your external system. This response contains information about the user, session, and the AI Agent output:

    ```JSON theme={null}
    {
       "userId": "user123",
       "sessionId": "session123",
       "AIOutput": {
          "text": "I’d be happy to help. Could you please provide your order number?",
          "data": {},
          "traceId": "endpoint-httpIncomingMessage-83b52cb7-1452-4c2d-a57d-4a81e6adb92c",
          "disableSensitiveLogging": false,
          "source": "bot"
       }
    }
    ```

    | Parameter                 | Type    | Description                                                                                                                             |
    | ------------------------- | ------- | --------------------------------------------------------------------------------------------------------------------------------------- |
    | `userId`                  | String  | The ID of the user who sent the original request.                                                                                       |
    | `sessionId`               | String  | The session ID used to track the conversation context.                                                                                  |
    | `AIOutput.text`           | String  | The response message generated by the AI Agent.                                                                                         |
    | `AIOutput.data`           | Object  | The message data returned from the Flow.                                                                                                |
    | `AIOutput.traceId`        | String  | The ID used for tracing and debugging purposes.                                                                                         |
    | `AIOutput.source`         | String  | The message source. Always `"bot"` for AI Agent replies.                                                                                |
    | `disableSensitiveLogging` | Boolean | The flag indicating if logging is disabled. If the value is `true`, this interaction won't be logged for privacy or compliance reasons. |
  </Accordion>
</AccordionGroup>

***

<sup>1</sup>: You must provide at least one of `text` or `data`. You can send either, or both. If both are missing or invalid, the Webhook Endpoint throws an error.


# WhatsApp
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/whatsapp



<a href="/release-notes/2026.16"><Badge>Updated in 2026.16.0</Badge></a>

<Frame>
  <img alt="WhatsApp Endpoint logo" />
</Frame>

The WhatsApp Endpoint enables you to connect AI Agents to the [WhatsApp Cloud API](https://developers.facebook.com/docs/whatsapp/cloud-api/) via the [Facebook
(Meta) App Dashboard](https://developers.facebook.com/docs/development/create-an-app/app-dashboard/).
This way, the AI Agent can send messages on WhatsApp Messenger and engage in conversations with your end users.

## Prerequisites

* [Registered as a Meta Developer](https://developers.facebook.com/docs/development/register).
* [Created a Facebook (Meta) Business account](https://business.facebook.com/).

## Generic Endpoint Settings

Learn about the generic Endpoint settings on the following pages:

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Transformer Functions](/ai/for-developers/transformers/overview)
* [Session Management](/ai/agents/deploy/endpoints/session-management)
* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)
* [File Storage](/ai/agents/deploy/endpoints/file-storage)
* [Inject and Notify](/ai/agents/deploy/endpoints/inject-and-notify)

## Specific Endpoint Settings

<AccordionGroup>
  <Accordion title="WhatsApp Setup">
    | Parameter       | Description                                                                                                                                                                                                            |
    | --------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | App ID          | A unique App ID used by the Meta/Facebook App Dashboard to identify the WhatsApp app that communicates with the Cloud API.                                                                                             |
    | Access Token    | A token provided by the WhatsApp app on the Meta/Facebook App Dashboard that enables communication between Cognigy and the WhatsApp Cloud API.                                                                         |
    | Phone Number ID | An ID for a phone number registered on the WhatsApp app or provided by the Facebook App Dashboard during setup for testing. This number will be used as an ID in the requests made to the WhatsApp Cloud API.          |
    | Verify Token    | A verification token for connecting the Cognigy WhatsApp Endpoint to the Meta/Facebook App Dashboard. This token needs to be generated on the Cognigy side and can consist of letters, numbers, and punctuation marks. |
    | App Secret      | This parameter can be optional. It adds extra security by requiring both the Access Token and the App Secret for each request.                                                                                         |
  </Accordion>

  <Accordion title="WhatsApp Settings">
    | Parameter     | Description                                                                            |
    | ------------- | -------------------------------------------------------------------------------------- |
    | Message Delay | The amount of time to wait between sending AI Agent replies. Measured in milliseconds. |
  </Accordion>

  <Accordion title="Session Management">
    | Parameter          | Description                                                                                                   |
    | ------------------ | ------------------------------------------------------------------------------------------------------------- |
    | Session Expiration | The amount of inactivity after which a session expires. Measured in minutes. The default value is 30 minutes. |
  </Accordion>
</AccordionGroup>

## How to Set Up

To set up the WhatsApp Endpoint, follow these steps:

<AccordionGroup>
  <Accordion title="1. Configure a WhatsApp App">
    1. Go to the [Create an App Dashboard](https://developers.facebook.com/apps/) page. Click **Create app**, then select **Use Cases > Other**.
    2. On the **Create an app** page, select **Business app type** and click **Next**.
    3. Fill in the following fields:
       * **App name** — enter the app name that will show on your **My Apps** page and is associated with your app ID.
       * **Contact email** — enter the email address to contact you about your app. Make sure it is an address you check regularly. Meta/Facebook may contact you about policies, app restrictions, or recovery if your app is deleted or compromised.
       * **Business portfolio** — at this stage, this parameter is optional. Connecting a business portfolio to your app is only required for certain products and permissions.
    4. Enter your account password for confirmation. On the **Add products to your app** page, go to the **WhatsApp** card and click **Set up**.
    5. You will be asked to add a Meta/Facebook existing Business Account. Select an account and click **Continue**. Confirm the account by going to your email and clicking the received link.
    6. In the left-side menu of the [App Dashboard](https://developers.facebook.com/docs/development/create-an-app/app-dashboard) interface, select **WhatsApp > Quickstart**.
    7. In the **WhatsApp Business Platform API** window, select your portfolio from the **Select business portfolio** list. If you don't have one, you need to create it. Click **Continue**.
    8. In the upper-right corner, go to **WhatsApp > API Setup**.
    9. In the **Temporary access token** field, copy the value and save it for later use. Note that this token is useful for testing purposes, but it expires after 23 hours. To get a permanent token, refer to the Get a Permanent Access Token section.
    10. Go to the **Step 1: Select phone numbers** section. Copy the Phone Number ID that is placed under the **Test number** field and save it for later use.
    11. Go to the **To** field. Click **Select a recipient phone number > Manage phone number list**. Specify your WhatsApp number to which you want to send messages.
    12. Go to **Step 2: Send messages with the API**, and click **Send message**. You will receive a welcome message on WhatsApp Messenger. The sender will be specified as the number selected in **Step 1: Select phone numbers**.
    13. In the left-side menu, go to **App settings > Basic**. Copy the ID from the **App ID** field and save it for later use.
    14. Next to the **App secret** field, click **Show**, confirm the action by entering your password, then copy the secret and save it for later use.
  </Accordion>

  <Accordion title="2. Configure a WhatsApp Endpoint">
    1. In the left-side menu of the Cognigy Project, go to **Deploy > Endpoints**.
    2. On the **Endpoints** page, click **+ New Endpoint**.
    3. In the **New Endpoint** section, do the following:
       1. Select the **WhatsApp** Endpoint type.
       2. Add a unique name.
       3. Select a relevant Flow from the list.
    4. In the **Configuration Information** section, copy the Endpoint URL and save it for later use on the Meta/Facebook App Dashboard.
    5. Activate the **Enable Endpoint** setting.
    6. In the **WhatsApp Setup** section, in the **Verify Token** field, create a verification token for connecting the WhatsApp Endpoint to the WhatsApp app. The token can consist of letters, numbers, and punctuation marks. Save this token for later use on the Meta/Facebook App Dashboard.
    7. Fill in the following fields:
       * **App ID** — enter the App ID that you copied and saved previously.
       * **Access Token** — enter the temporary access token that you copied and saved previously.
       * **Phone Number ID** — enter the Phone Number ID that you copied and saved previously.
       * **App Secret** — enter the App Secret that you copied and saved previously.
         Save changes.
  </Accordion>

  <Accordion title="3. Complete the WhatsApp Integration">
    1. Open your WhatsApp app that you configured previously. In the left-side menu, go to **WhatsApp > Configuration**.
    2. In the **Webhook** section, click **Edit**.
    3. In the **Edit webhook's callback URL** window, fill in the following fields:
       * **Callback URL** — enter the Webchat Endpoint URL in the Callback URL field.
       * **Verify Token** — enter the token you created in the **WhatsApp Setup** section in the **Verify Token** field.
    4. Click **Verify and Save**. The connection between Cognigy and WhatsApp app will be established.
    5. Subscribe your webhook to the message event. In the **Webhook fields** section, click **Manage**.
    6. In the **Name** column, find `messages`, and in the **Subscribe** column, activate the option for this event.
  </Accordion>

  <Accordion title="4. Get a Permanent Access Token">
    To prevent your connection from expiring after 23 hours, replace the temporary access token with a permanent access token:

    1. On the Meta/Facebook App Dashboard, go to the Business settings of your app. To do so, click the blue link next to the business name.
    2. In the left-side menu of the app, go to **Users > System Users**. Click **Add**.
    3. Specify a name in the **System Username** field and select a role from the **System User Role** list. Click **Create System User**.
    4. On the **Assigned assets** page, click **Add Assets**.
    5. Go to **Apps**, select the app, enable the **Manage App** option, and then save changes.
    6. Click **Generate Token** and select the app from the **Select App** list for which you want to generate a token.
    7. From the available permissions, select **whatsapp\_business\_messaging** and **whatsapp\_business\_management**, then click **Generate token**.
    8. Go to the WhatsApp Endpoint on the Cognigy platform.
    9. In the **WhatsApp Setup** section, go to the **Access Token** field. Replace the temporary access token with the generated access token. Save changes.
  </Accordion>
</AccordionGroup>

## More Information

* [Cognigy Help Center: WhatsApp Endpoint: Messages Types and Payload Examples](https://support.cognigy.com/hc/en-us/articles/5326454286108)


# Workplace by Facebook
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/workplace-by-fb



<Frame>
  <img alt="Workplace by Facebook Endpoint logo" />
</Frame>

<Warning>
  From September 1, 2025, to May 31, 2026, Workplace will be in read-only mode for viewing and downloading data. Workplace will be shut down on [June 1, 2026](https://www.workplace.com/help/work/1167689491269151). Ensure you migrate to an alternative service and a different Cognigy endpoint before that date.
</Warning>

The Workplace Endpoint enables you to connect AI Agents to the [Workplace](https://www.workplace.com/) platform.

## Prerequisites

* A Workplace account at [workplace.com](https://www.workplace.com/).
* Admin access to the Workplace Admin Panel.

## Restrictions

* The Workplace by Facebook Endpoint is available only in SaaS environments.

## Generic Endpoint Settings

Learn about the generic Endpoint settings on the following pages:

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [NLU Connectors](/ai/platform-features/nlu/external/nlu-connectors/overview)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Session Management](/ai/agents/deploy/endpoints/session-management)
* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)
* [Transformer Functions](/ai/for-developers/transformers/overview)
* [Handover Providers](/ai/escalate/handovers)
* [Inject and Notify](/ai/agents/deploy/endpoints/inject-and-notify)

## Specific Endpoint Settings

### Workplace Settings

<Accordion title="Workplace Setup">
  | Parameter    | Description                                                                                                                                  |
  | ------------ | -------------------------------------------------------------------------------------------------------------------------------------------- |
  | Recipient ID | The ID from the URL of a conversation with your Workplace bot. It must be added after the bot has been installed in your Workplace instance. |
</Accordion>

## How to Set Up

### Setup on the Workspace Side

<Accordion title="Create a Workplace">
  1. Go to workplace.com and set up a Workplace.
  2. After adding the Cognigy bot, go to the Admin Panel in Workplace.
  3. In the left-side menu, click **Integrations**.
  4. Confirm the bot appears under **Added to Workplace > Added By Your Company**.
  5. On the Workplace Home screen, scroll to the Bots section to locate your Cognigy bot.
  6. Start a conversation with the bot and copy the `Recipient Id` from the top bar.
</Accordion>

### Setup on the Cognigy Side

<Accordion title="Create a Workplace Endpoint">
  1. In the left-side menu of the Cognigy Project, go to **Deploy > Endpoints**.
  2. On the **Endpoints** page, click **+ New Endpoint**.
  3. In the **New Endpoint** section, do the following:
     1. Select the **Workplace** Endpoint type.
     2. Add a unique name.
     3. Select a relevant Flow from the list.
  4. In the **Workplace Settings** section, click **Install**.
  5. In the window, review the Cognigy Terms of Service and Privacy Policy.
  6. *(Optional)* Click **Customize** to modify **Integration Name** and **Description**.
  7. *(Optional)* Select **Specific groups** to restrict the bot to certain group chats.
  8. Click **Add to Workplace** to integrate the bot.
  9. Paste the copied `Recipient Id` into the **Workplace Recipient Id** field. Save changes.
</Accordion>

## More Information

* [Help Center: Workplace by Facebook Deploy an Endpoint](https://support.cognigy.com/hc/en-us/articles/360016310379)


# Zoom Contact Center
Source: https://docs.cognigy.com/ai/agents/deploy/endpoint-reference/zoom



<a href="/release-notes/2025.12"><Badge>Updated in 2025.12</Badge></a>

<Frame>
  <img alt="Zoom Contact Center Endpoint logo" />
</Frame>

The Zoom Contact Center Endpoint enables you to connect your AI Agent to the [Zoom Contact Center platform](https://www.zoom.com/en/products/contact-center/).

## Prerequisites

* Access to [the Zoom Contact Center platform](https://www.zoom.com/en/products/contact-center/).

## Generic Endpoint Settings

Learn about the generic Endpoint settings on the following pages:

* [Endpoints Overview](/ai/agents/deploy/endpoints/overview)
* [NLU Connectors](/ai/platform-features/nlu/external/nlu-connectors/overview)
* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)
* [Transformer Functions](/ai/for-developers/transformers/overview)

## Specific Endpoint Settings

<Accordion title="Zoom Contact Center Settings">
  | **Parameter** | **Description**                                                                                                                                                                                                               |
  | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
  | Verify Token  | A secret token defined in the Zoom Contact Center integration to ensure secure communication between Zoom and Cognigy. You can obtain this token from your Zoom admin or support representative during the integration setup. |
</Accordion>

## How to Set Up

### Setup on the Zoom Contact Center Side

Contact your Zoom representative or support to provide you with the Verify Token (secret token) and further instructions on how to set up the integration on the Zoom Contact Center side.

### Setup on the Cognigy Side

<Accordion title="Create a Zoom Contact Center Endpoint">
  1. In the left-side menu of your Project, click **Deploy > Endpoints**.
  2. On the **Endpoints** page, click **+ New Endpoint**.
  3. In the **New Endpoint** section, do the following:
     1. Select the **Zoom Contact Center** Endpoint type.
     2. Specify a unique name.
     3. Select a relevant Flow from the list and save changes.
  4. In the Endpoint settings, go to the **Zoom Contact Center Settings** section.
  5. In the **Verify Token** field, enter the token that Zoom support provided your previously.
  6. Save changes, then verify the integration with your Zoom representative or support to finalize the setup.
</Accordion>

## More Information

* [Overview](/ai/agents/deploy/endpoint-reference/zoom)
* [Zoom Contact Center Documentation](https://support.zoom.com/hc/en/category?id=kb_category\&kb_category=1e16b2d68720391089a37408dabb3553)
* [Zoom Marketplace](https://marketplace.zoom.us/)


# Content Conversion
Source: https://docs.cognigy.com/ai/agents/deploy/endpoints/content-conversion



Cognigy provides a wide range of Endpoints and components, such as the Interaction Panel, xApp: Show Adaptive Card Node, and Live Agent, for its conversational AI platform. However, not all of these Endpoints and components can correctly convert your content to the intended output type.

If the Endpoint or component doesn't support the selected output type, Cognigy falls back to text as a default output.

The following table shows the [output types](/ai/agents/develop/node-reference/basic/say) that Cognigy can provide for each Endpoint or component with these notations:

* <img alt="supported" /> — supports and converts the output type.
* <img alt="not supported" /> — doesn't support the output type and falls back to plain text.
* N/A — the output type isn't applicable for the Endpoint or component.

<div>
  | <span>Component, Endpoint</span> | Text                    | Text with Quick Replies     | Gallery                     | Text with Buttons           | List                        | Audio                       | Image                       | Video                       | JSON/XML                    | Adaptive Card (including supported versions) |
  | -------------------------------- | ----------------------- | --------------------------- | --------------------------- | --------------------------- | --------------------------- | --------------------------- | --------------------------- | --------------------------- | --------------------------- | -------------------------------------------- |
  | **Components**                   |                         |                             |                             |                             |                             |                             |                             |                             |                             |                                              |
  | Demo Webchat                     | <img alt="supported" /> | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="not supported" /> | 1.5                                          |
  | Simulator                        | <img alt="supported" /> | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="not supported" /> | <img alt="supported" />     | <img alt="not supported" /> | <img alt="not supported" /> | 1.5                                          |
  | Interaction Panel                | <img alt="supported" /> | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="not supported" /> | 1.2                                          |
  | Live Agent                       | <img alt="supported" /> | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="not supported" /> | 1.5                                          |
  | Live Agent (LA) Agent Copilot    | <img alt="supported" /> | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="not supported" /> | 1.5                                          |
  | Webchat Adaptive Card plugin     | N/A                     | N/A                         | N/A                         | N/A                         | N/A                         | N/A                         | N/A                         | N/A                         | <img alt="not supported" /> | 1.2                                          |
  | Copilot: Adaptive Card Tile      | N/A                     | N/A                         | N/A                         | N/A                         | N/A                         | N/A                         | N/A                         | N/A                         | <img alt="not supported" /> | 1.6                                          |
  | **Endpoints**                    |                         |                             |                             |                             |                             |                             |                             |                             | <img alt="not supported" /> | 1.5                                          |
  | Webchat                          | <img alt="supported" /> | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="not supported" /> | 1.5                                          |
  | Voice Gateway                    | <img alt="supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" />                  |
  | WhatsApp                         | <img alt="supported" /> | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | JSON                        | <img alt="not supported" />                  |
  | 8x8                              | <img alt="supported" /> | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | JSON                        | 1.3<sup>[1](#footnote1)</sup>                |
  | Alexa (deprecated)               | <img alt="supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" />                  |
  | Avaya CPaaS                      | <img alt="supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | XML                         | <img alt="not supported" />                  |
  | Facebook (Meta) Messenger        | <img alt="supported" /> | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="not supported" /> | <img alt="not supported" />                  |
  | Azure Bot Services               | <img alt="supported" /> | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | JSON                        | 1.5<sup>[1](#footnote1)</sup>                |
  | Line                             | <img alt="supported" /> | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | JSON                        | <img alt="not supported" />                  |
  | Microsoft Teams                  | <img alt="supported" /> | <img alt="not supported" /> | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | JSON                        | 1.5<sup>[1](#footnote1)</sup>                |
  | RingCentral Engage               | <img alt="supported" /> | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | JSON                        | <img alt="not supported" />                  |
  | Slack                            | <img alt="supported" /> | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | JSON                        | <img alt="not supported" />                  |
  | Sunshine Conversations           | <img alt="supported" /> | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | JSON                        | <img alt="not supported" />                  |
  | Twilio                           | <img alt="supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | TwiML                       | <img alt="not supported" />                  |
  | Twilio SMS                       | <img alt="supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | TwiML                       | <img alt="not supported" />                  |
  | Userlike                         | <img alt="supported" /> | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | <img alt="supported" />     | JSON                        | <img alt="not supported" />                  |
  | AudioCodes                       | <img alt="supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" />                  |
  | Genesys                          | <img alt="supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | JSON                        | <img alt="not supported" />                  |
  | Bandwidth                        | <img alt="supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | JSON                        | <img alt="not supported" />                  |
  | NiCE CXone                       | <img alt="supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | JSON                        | <img alt="not supported" />                  |
</div>

<span>
  <sup>1</sup>: For third-party Endpoints supporting Adaptive Cards, the current version may differ from what is indicated in the table. We recommend verifying the latest Adaptive Card version with your provider or contacting their support service for information on supported versions.
</span>


# Data Protection & Analytics
Source: https://docs.cognigy.com/ai/agents/deploy/endpoints/data-protection-and-analytics



<a href="/release-notes/4.92"><Badge>Updated in 4.92</Badge></a>

The *Data Protection & Analytics* section includes Endpoint settings for analytics and system logging. Here you can toggle whether you want to create and store data about contacts, and whether you want to collect analytics data for the Endpoint. If you choose to collect data, you can also configure whether you're masking sensitive data.

| Setting                                              | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
| ---------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Contact Profiles                                     | Collects individual [Contact Profiles](/ai/agents/analyze/contact-profiles) data to store user data across sessions.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
| Collect Analytics                                    | Collects Analytics data for the Endpoint. The [Voice Gateway Endpoint](/ai/agents/deploy/endpoint-reference/voice-gateway) has [additional settings for collecting analytics data](#store-extensive-data-payloads-in-analytics).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
| Mask IP Address                                      | Masks the IP address from the user in the [Input object](/ai/agents/develop/ai-agent-memory/input), [Insights](/insights/overview), and [Analytics](/ai/agents/analyze/overview). When this setting is activated, IP addresses are replaced by asterisks and aren't available through the [OData Analytics Endpoint](/ai/agents/analyze/odata).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
| Mask Sensitive Analytics                             | Masks sensitive information in the [Input object](/ai/agents/develop/ai-agent-memory/input), [Insights](/insights/overview), and [Analytics](/ai/agents/analyze/overview). Specifically, the `inputText`, `inputData`, and `userLanguageText` properties, and the user IP address are replaced with asterisks and aren't available through [Cognigy OData Analytics Endpoint](/ai/agents/analyze/odata). If you activate this setting, you can't use the [Intent Trainer](/ai/platform-features/nlu/intents/intent-trainer) because Cognigy doesn't store the user's input text.<br /><br />Additionally, you can use the Code Node to [ignore specific analytics](#ignore-specific-analytics-data).                                                                                                                                           |
| Mask Sensitive Logging                               | Masks potential sensitive information in system logs on the [Logs](/ai/agents/test/logs) page. If you activate this setting, Cognigy replaces the following log entries with asterisks:<ul><li>`userId`</li><li>`text`</li><li>`data`</li></ul>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
| Collect Conversations                                | Stores the conversation history in Contact Profiles.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
| Enable Input Sanitization                            | Activates server-side input sanitization. If you activate this setting, Cognigy removes potentially harmful HTML content from user inputs before processing them in other parts of the system.<br /><br />If you use [Input Transformers](/ai/for-developers/transformers/input-transformer), the sanitization takes place after their execution. This process involves validating and cleaning the data on the server to make sure the data is safe and complies with the expected format.<br /><br />For [Webchat](/webchat/index), client-side input [sanitization](https://github.com/Cognigy/WebchatWidget/blob/master/docs/embedding.md#endpoint-settings) (`disableHtmlContentSanitization: false`) works by default. In this case, activating server-side input sanitization provides an additional layer of security for your server. |
| Hide References to External Resources in Transcripts | This setting removes tags that reference third-party APIs, such as `<a>` for links and `<img>` for images, from user inputs in conversation transcripts to protect users from risks associated with third-party content. In the Conversations explorer in Insights, conversation transcripts keep references to the content instead of rendering the actual links or images. For example, `<img src="url-to-an-image">` or `<a href="url">`.<br /><br />For the Interaction Panel, set `DISABLE_SKIP_URI_TAGS_IP_CONVERSATIONS: "true"` in `values.yaml` to hide references to external resources.                                                                                                                                                                                                                                             |
| Dashbot                                              | Allows you to integrate Cognigy with [Dashbot](https://www.dashbot.io) using a Dashbot API key. Refer to the [Help Center Dashbot article](https://support.cognigy.com/hc/en-us/articles/360016311379) for a detailed description of how to set up the Dashbot integration.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |

## Endpoint-Specific Settings

### Voice Gateway Endpoint

#### Store Extensive Data Payloads in Analytics

If you activate **Store Extensive Data Payloads in Analytics** in the [Voice Gateway](/ai/agents/deploy/endpoint-reference/voice-gateway) Endpoint, all the information from [Voice Gateway event](/voice-gateway/references/events/overview) payloads is stored in the Analytics database. You can access this information via [OData](/ai/agents/analyze/odata) or the [Conversations explorer](/insights/explorers/conversations) in Insights. You can activate this setting after you toggle on **Collect Analytics**. By default, the setting is disabled.

This setting allows you to store the following information:

* `trace_id`
* `account_sid`
* `application_sid`
* `call_id`
* `sip_status`
* `sip_reason`
* `originating_sip_ip`
* `local_sip_address`
* `service_provider_sid`
* `fs_sip_address`
* `api_base_url`
* `originating_sip_trunk_name`

## Ignore Specific Analytics Data

You can deactivate the logging of specific analytics data fields with a [Code Node](/ai/for-developers/code/analytics-data) using the `analyticsdata` object. For example:

```js theme={null}
  analyticsdata.text = "overwritten text";
  delete analyticsdata.data.token;
```

In this example, the Code Node deletes `token` from the analytics data before it's written to the database.

## More Information

* [Endpoints](/ai/agents/deploy/endpoints/overview)


# File Storage
Source: https://docs.cognigy.com/ai/agents/deploy/endpoints/file-storage



<a href="/release-notes/2026.13"><Badge>Updated in 2026.13.1</Badge></a>

By default, users can't attach files to chat messages. However, you can configure *File Storage* in the Endpoint settings that allows users to upload files in the chat interface. To configure file storage, you need to create a connection to a third-party storage provider.

After uploading a file to the chat interface, the file is saved in third-party storage and can be viewed by both the sender and the recipient.

## Key Features

* **Improved Conversational Experience**. Users can upload screenshots to illustrate technical issues and share various documents, such as contracts, invoices, order confirmations, and photos of purchased goods. File uploads enable richer interactions and additional capabilities for your Flow.
* **File Manipulation**. You can retrieve file data, resize images, and validate files.
* **Security Measures**. Cognigy performs a virus scan on the file before uploading it to the cloud storage provider, checking for potential malware or viruses, and preventing harmful content from being stored or shared through the cloud.

## Prerequisites

* You have access to one of the following third-party storage providers:
  * [Azure Blob Storage](https://learn.microsoft.com/en-us/azure/storage/blobs/storage-blobs-introduction)
  * [Amazon S3](https://docs.aws.amazon.com/AmazonS3/latest/userguide/Welcome.html)
  * [Google Cloud Storage](https://cloud.google.com/storage/docs/)
* Files are accessed using the configured storage credentials, so no anonymous or public read access is needed. Configure the storage container or bucket as private:
  * For Azure Blob Storage, set the container access level to **Private (no anonymous access)**.
  * For Amazon S3, enable **Block all public access** in the bucket's permissions settings.
  * For Google Cloud Storage, set **Public access prevention** to **Enforced** in the bucket configuration.

## Restrictions

* Only the following Endpoints support file uploads:
  * [8x8](/ai/agents/deploy/endpoint-reference/8x8)
  * [Webchat v2](/webchat/v2/overview)
  * [Webchat v3](/webchat/v3/overview)
  * [WhatsApp](/ai/agents/deploy/endpoint-reference/whatsapp)

## How to Use the File Storage

### Create Connections to File Storage Providers

To enable file uploads, connect to one of the following file storage providers:

<Tabs>
  <Tab title="Azure Blob Storage Container">
    <AccordionGroup>
      <Accordion title="1. On the Azure Side">
        1. Go to the [Azure portal](https://portal.azure.com/#home) and select **Storage Accounts** in the **Azure services** section.
        2. On the **Storage accounts** page, select the account from the **Name** column.
        3. In the navigation menu, go to **Data storage > Containers**. Copy the name of the target container from the **Name** column. This container will store user-uploaded files. Save the container name for later use.
        4. Select the container, then click **Change access level**. Set the access level to **Private (no anonymous access)** and click **OK**. The service uses storage credentials to access files, so anonymous access isn't needed.
        5. In the navigation menu, go to **Security + networking** and select **Networking**. Under **Public network access**, ensure the setting is **Enabled from all networks**. This keeps the storage account accessible over the internet so the service can authenticate using credentials.
        6. In **Security + networking**, select **Access Keys**. On the **Access Keys** page, copy the name from the **Storage account name** field and the key from the **Key** field for later use.
      </Accordion>

      <Accordion title="2. On the Cognigy Side">
        1. Open the Project that contains the Endpoint for which you want to configure file storage. Go to **Deploy > Endpoints** and select an Endpoint or create one.
        2. Depending on the Endpoint you select, proceed as follows:
           * **Webchat v3** — activate the **Allow Attachment Upload** setting in **Webchat Behavior** in the **Attachment Upload** section and select **Azure** from the **File Storage Provider** list.
           * **Other Endpoints** — go to the **File Storage** section and select **Azure** from the **File Storage Provider** list.
        3. Next to the **Azure Connection**, click **+** and fill in the following fields:
           * **Container Name** — enter the name of the container that you copied and saved previously.
           * **Account Name** — enter the storage account name that you copied and saved previously.
           * **Account Key** — enter the key that you copied and saved previously.
        4. Click **Create**, then **Save**.
      </Accordion>
    </AccordionGroup>
  </Tab>

  <Tab title="Amazon S3 Bucket or S3-compatible Storage">
    <AccordionGroup>
      <Accordion title="1. On the AWS or S3-Compatible Storage Side">
        1. Make sure the bucket has the required upload access. AWS uses [IAM policies](https://docs.aws.amazon.com/IAM/latest/UserGuide/reference_policies_elements_statement.html) for access control. For Amazon S3-compatible cloud storage, use an equivalent access control mechanism. Configure the following AWS IAM policy for AWS buckets:

        ```json theme={null}
        {
          "Version": "2012-10-17",
          "Statement": [
            {
              "Effect": "Allow",
              "Action": "s3:ListBucket",
              "Resource": "arn:aws:s3:::BUCKET-NAME"
            },
            {
              "Effect": "Allow",
              "Action": [
                "s3:PutObject",
                "s3:PutObjectAcl",
                "s3:GetObject"
              ],
              "Resource": "arn:aws:s3:::BUCKET-NAME/*"
            }
          ]
        }
        ```

        2. Make sure that you have access to an [Amazon S3 bucket](https://docs.aws.amazon.com/AmazonS3/latest/userguide/Welcome.html) or to an Amazon S3-compatible cloud, for example, [Open Telekom Cloud](https://docs.otc.t-systems.com/), [DigitalOcean Spaces](https://docs.digitalocean.com/products/spaces/), [Backblaze B2](https://www.backblaze.com/docs). You need at least one bucket, or [create a new one](https://docs.aws.amazon.com/AmazonS3/latest/userguide/create-bucket-overview.html). For a cloud storage provider compatible with Amazon S3, you need information similar to that described in the following steps.
        3. In the [Amazon S3 console](https://console.aws.amazon.com/s3/), go to **Buckets**, select the bucket, then go to **Permissions** tab > **Block public access (bucket settings)** > **Edit**. Enable **Block all public access** and save changes. New buckets have this enabled by default.
        4. In the [Amazon S3 console](https://console.aws.amazon.com/s3/), go to the **General purpose buckets** tab.
        5. From the **Name** column, copy the name of the target bucket. This bucket will store user-uploaded files. Save the bucket name for later use.
        6. From the **AWS Region** column, copy the region ID, for example, `us-east-1`, and save it for later use.
        7. In the upper-right corner, go to your account, then select **Security credentials**.
        8. On the **My Security Credentials** page, go to **Access keys** and click **Create access key**.
        9. On the **Alternatives to root user access keys** page, review the security recommendations. To continue, select the option, and then click **Create access key**. On the **Retrieve access key** page, your access key ID is displayed.
        10. Under **Secret access key**, click **Show** and then copy the access key ID and secret key from your browser window and save them for later use. Click **Done**. If you want to use temporary access with MFA instead of a secret key, you need a [session token](https://docs.aws.amazon.com/STS/latest/APIReference/API_GetSessionToken.html#:~:text=Returns%20a%20set%20of%20temporary,key%2C%20and%20a%20security%20token). To set up a session token, obtain this value via AWS API and specify the token in the `AWS_SESSION_TOKEN` variable in the `values.yaml` file. For more information, read [Using temporary credentials with AWS resources](https://docs.aws.amazon.com/IAM/latest/UserGuide/id_credentials_temp_use-resources.html).
      </Accordion>

      <Accordion title="2. On the Cognigy Side">
        1. Open the Project that contains the Endpoint for which you want to configure file storage. Go to **Deploy > Endpoints** and select an Endpoint or create one.
        2. Depending on the Endpoint you select, proceed as follows:
           * **Webchat v3** — activate the **Allow Attachment Upload** setting in **Webchat Behavior** in the **Attachment Upload** section and select **AWS** from the **File Storage Provider** list.
           * **Other Endpoints** — go to the **File Storage** section and select **AWS** from the **File Storage Provider** list.
        3. Next to **AWS Connection**, click **+** and fill in the following fields:
           * **Access Key ID** — enter the access key ID that you copied and saved previously.
           * **Secret Access Key** — enter the secret key that you copied and saved previously.
           * **Region** — enter the AWS region that you copied and saved previously.
           * **Bucket Name** — enter the name of the bucket that you copied and saved previously.
           * **Custom URL** — enter the custom URL of your S3-compatible storage. This parameter applies only to Amazon S3-compatible cloud storage.
        4. Click **Create**, then **Save**.
      </Accordion>
    </AccordionGroup>
  </Tab>

  <Tab title="Google Cloud Storage Bucket">
    <AccordionGroup>
      <Accordion title="1. On the Google Cloud Console Side">
        1. In the [Google Cloud console](https://console.cloud.google.com/), select **Cloud Storage > Buckets** in the navigation menu.
        2. On the **Buckets** page, copy the name of the target bucket from the **Bucket** column. This bucket will store user-uploaded files. Save the bucket name for later use.
        3. Select the bucket, go to the **Configuration** tab, then set **Public access prevention** to **Enforced**. This prevents the bucket from being exposed publicly.
        4. In the navigation menu, select **IAM & Admin > Service Accounts**.
        5. From the **Email** field, copy the email address and save it for later use.
        6. In the **Actions** column, click **Actions > Manage Keys**.
        7. On the **Keys** tab, click **Add Key > Create New Key** and select the JSON key type. The key will be downloaded to your computer. Open the JSON file and copy the entire value of the `private_key` parameter, including the lines:

        ```txt theme={null}
        -----BEGIN PRIVATE KEY-----
        -----END PRIVATE KEY-----
        ```

        Save the key for later use.
      </Accordion>

      <Accordion title="2. On the Cognigy Side">
        1. On the Cognigy side, open the Project that contains the Endpoint for which you want to configure file storage. Go to **Deploy > Endpoints** and select an Endpoint or create one.
        2. Depending on the Endpoint you select, proceed as follows:
           * **Webchat v3** — activate the **Allow Attachment Upload** setting in **Webchat Behavior** in the **Attachment Upload** section and select **Google Cloud** from the **File Storage Provider** list.
           * **Other Endpoints** — go to the **File Storage** section and select **Google Cloud** from the **File Storage Provider** list.
        3. Next to **Google Cloud Connection**, click **+** and fill in the following fields:
           * **Bucket Name** — enter the bucket name that you copied and saved previously.
           * **Client Email** — enter the email address that you copied and saved previously.
           * **Private Key** — enter the key that you copied and saved previously.
        4. Click **Create**, then **Save**.
      </Accordion>
    </AccordionGroup>
  </Tab>
</Tabs>

To test file uploads for the Webchat v2 and Webchat v3 Endpoints, use [Demo Webchat](/webchat/demo). Upload a file in the Demo Webchat interface by either dragging it from your computer or clicking <img alt="attach file" />.

For other Endpoints, testing only works in the production environment. Upload a file to the chat interface and go to the file storage on your provider's side. If the file appears in the storage, the file was successfully uploaded.

### Access the File Storage through the Input Object

By default, the data of uploaded files is stored in the `input.data.attachments` array. Each file object in this array contains a file name, type, and URL pointing to the uploaded file:

```json theme={null}
{
  "name": "example_file.png",
  "url": "https://example.com/uploads/example_file.png",
  "type": "image"
}
```

<Accordion title="Use Cases">
  | Use Case                    | Description                                                                                                                        | CognigyScript                           | Example Output                                 |
  | --------------------------- | ---------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------- | ---------------------------------------------- |
  | Confirm uploaded files      | To ensure the user uploaded the correct file, display the file name for confirmation.                                              | `{{input.data.attachments[0].name}}`    | `example_file.png`                             |
  | Provide the direct file URL | Share the direct URL of the uploaded file with the user.                                                                           | `{{input.data.attachments[0].url}}`     | `https://example.com/uploads/example_file.png` |
  | Validate attachments        | To check if a valid file has been uploaded in the most recent user input, verify the length of the `input.data.attachments` array. | `{{input.data.attachments.length}} > 0` | -                                              |
</Accordion>

### Resize Images

For image file types, such as JPEG, PNG, GIF, WebP, and others, you can automatically resize images based on the query parameters in the URL.
The available parameters are in the table.

| Query Parameter      | Description                                             |
| -------------------- | ------------------------------------------------------- |
| `maxWidth`           | The maximum width of the image.                         |
| `maxHeight`          | The maximum height of the image.                        |
| `maxFileSizeInBytes` | The maximum file size of the image, specified in bytes. |

<Accordion title="Use Cases">
  | Use Case                      | Description                                                                                                                                        | Example                                                                                                 | URL Example                                                             |
  | ----------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------- |
  | Attach image without changes  | An attachment is uploaded with the original URL.                                                                                                   | The image has dimensions of 3840 x 2160 and a size of 8 MB.                                             | `https://files-trial.cognigy.ai/123/456/789`                            |
  | Resize to specific dimensions | If you provide the `maxWidth` and `maxHeight` parameters, the image is resized to fit these parameters while maintaining the aspect ratio.         | To resize the image to a maximum of 640x480, append the `maxWidth=640` and `maxHeight=480` parameters.  | `https://files-trial.cognigy.ai/123/456/789?maxWidth=640&maxHeight=480` |
  | Limit file size               | If you provide the `maxFileSizeInBytes` parameter, the image quality is gradually reduced until the image file size meets the specified file size. | To ensure the image is less than 100 KB (102,400 bytes), use the `maxFileSizeInBytes=102400` parameter. | `https://files-trial.cognigy.ai/123/456/789?maxFileSizeInBytes=102400`  |

  **HTTP Error Codes**

  | Error Code | Description                                                                                                                             |
  | ---------- | --------------------------------------------------------------------------------------------------------------------------------------- |
  | `500`      | Returns if the image can't be resized according to the specified parameters. For example, if the `maxFileSizeInBytes` can't be reached. |
</Accordion>

## More Information

* [Azure Blob Storage](https://learn.microsoft.com/en-us/azure/storage/blobs/storage-blobs-introduction)
* [Amazon S3](https://docs.aws.amazon.com/AmazonS3/latest/userguide/Welcome.html)
* [Google Cloud Storage](https://cloud.google.com/storage/docs)
* [Create a Bucket in Azure Blob Storage](https://learn.microsoft.com/en-us/azure/storage/blobs/storage-quickstart-blobs-portal#create-a-container)
* [Create a Bucket in Amazon S3](https://docs.aws.amazon.com/AmazonS3/latest/userguide/create-bucket-overview.html)
* [Create a Bucket in Google Cloud Storage](https://cloud.google.com/storage/docs/creating-buckets)


# Inject and Notify API
Source: https://docs.cognigy.com/ai/agents/deploy/endpoints/inject-and-notify



The *Inject* and *Notify* API requests handle external processes that need to communicate back to an active session in Cognigy. You can use these API requests when an AI Agent needs to run external processes, for example, document analysis, outside of Cognigy before continuing the conversation. After the external process has concluded, it needs to trigger Cognigy.

## Limitations

<Warning>
  As of April 2025, the AudioCodes Endpoint is deprecated. The full removal of the AudioCodes Endpoint is planned for Q3 2026. Migrate to the [Voice Gateway Endpoint](/ai/agents/deploy/endpoint-reference/voice-gateway) before the removal date.
</Warning>

Only the following Endpoints support the inject and notify API requests:

* [Facebook (Meta) Messenger](/ai/agents/deploy/endpoint-reference/facebook-messenger)
* [LINE Messenger](/ai/agents/deploy/endpoint-reference/line-messenger)
* [Azure Bot Services](/ai/agents/deploy/endpoint-reference/azure-bot-services)
* [RingCentral Engage](/ai/agents/deploy/endpoint-reference/ringcentral-engage)
* [Slack](/ai/agents/deploy/endpoint-reference/slack)
* [Socket](/ai/agents/deploy/endpoint-reference/socketio)
* [Sunshine Conversations](/ai/agents/deploy/endpoint-reference/sunshine-conversations)
* [Voice Gateway](/ai/agents/deploy/endpoint-reference/voice-gateway)
* [AudioCodes](/ai/agents/deploy/endpoint-reference/audiocodes)
* [WhatsApp](/ai/agents/deploy/endpoint-reference/whatsapp)
* [Webchat](/ai/agents/deploy/endpoint-reference/webchat)
* [Webhook](/ai/agents/deploy/endpoint-reference/webhook)

## Inject

An inject API request sends text and data from an external service to a Flow. This approach allows you to convert them into complex information, for example, into lists and galleries, and then send it to the user through the Endpoint. You can send a simple message to your Flow, for example, `makeList`, along with the data from an external API request, and use the Nodes in the Flow to recognize the message and send a list to the user with information from an external API.

## Notify

A notify API request sends text or data to the user through the Endpoint. With this approach, the user receives the text or data from the external API without any additional Flow logic.

## How to Use Inject and Notify

You can use the [HTTP Request Node](/ai/agents/develop/node-reference/service/http-request) or [Code Node](/ai/for-developers/code/overview) to trigger external processes through the Cognigy API and pass the necessary data to the inject and notify API requests in addition to the specific external payload.

### Inject and Notify API URLs

For the inject and notify API methods, use the Endpoint service for the API URL. You can find your Endpoint URL and URL token in the [Endpoint settings](/ai/agents/deploy/endpoints/overview) under the **Endpoint URL** field or **Config URL** field, for Webchat Endpoints. The API routes are:

* Notify: `{EndpointURL}/notify/{URLToken}`
* Inject: `{EndpointURL}/inject/{URLToken}`

For example, if your Endpoint URL is `https://endpoint-trial.cognigy.ai/4e5e2f01110c935fff81ec54bbce12c175b831d8105dc69face36a8983feec17`, the base URL is `https://endpoint-trial.cognigy.ai` and URL token is `4e5e2f01110c935fff81ec54bbce12c175b831d8105dc69face36a8983feec17`. In this case, the API routes are:

* Notify: `https://endpoint-trial.cognigy.ai/notify/4e5e2f01110c935fff81ec54bbce12c175b831d8105dc69face36a8983feec17`
* Inject: `https://endpoint-trial.cognigy.ai/inject/4e5e2f01110c935fff81ec54bbce12c175b831d8105dc69face36a8983feec17`

### Payload

You can retrieve the necessary data for the inject and notify payload from the [Input object](/ai/agents/develop/ai-agent-memory/input) with [CognigyScript](/ai/platform-features/cognigyscript), as follows:

| Payload Parameter | CognigyScript |
| ----------------- | ------------- |
| text              | ci.text       |
| data              | ci.data       |
| sessionId         | ci.sessionId  |
| userId            | ci.userId     |
| URLToken          | ci.URLToken   |

### HTTP Error Codes

When inject or notify API requests fail, the API returns one of the following HTTP error codes:

| HTTP Error Code | Description                                                                                                                      |
| --------------- | -------------------------------------------------------------------------------------------------------------------------------- |
| 400             | The request is malformed. For example, a required parameter, such as `sessionId`, is missing.                                    |
| 404             | The Endpoint associated with the URL token wasn't found. Verify that the URL token is correct and that the Endpoint exists.      |
| 500             | An internal server error occurred. Try again later or contact [Cognigy technical support](/help/get-help) if the issue persists. |

## Inject and Notify in Transformers

There are cases where you don't have control over the payload that the external process sends to the inject or notify API requests. In these cases, you can use the [inject](/ai/for-developers/transformers/inject-transformer) or [notify](/ai/for-developers/transformers/notify-transformer) transformers to change the request and extract the necessary parameters from the request body before calling the actual notify or inject API.

## More Information

* [Inject Transformer](/ai/for-developers/transformers/inject-transformer)
* [Notify Transformer](/ai/for-developers/transformers/notify-transformer)


# Overview
Source: https://docs.cognigy.com/ai/agents/deploy/endpoints/overview

Configure Endpoints to connect AI Agents with external platforms like Messenger, and Slack for chat and voice.

<a href="/release-notes/2026.16"><Badge>Updated in 2026.16.0</Badge></a>

*Endpoints* are connectors between AI Agents and [external systems or platforms](/ai/agents/deploy/endpoint-reference/overview), such as Facebook (Meta) Messenger, Slack, and others.

Endpoints support chat and voice conversations in contact centers and integrate with [handover provider](/ai/escalate/handovers) to manage these interactions.

Endpoints convert [channel-specific](/ai/agents/develop/nodes/channels) input and output, including images, galleries, and voice, into the structure of the [Input object](/ai/agents/develop/ai-agent-memory/input). This approach allows the development of platform-agnostic [Flows](/ai/agents/develop/projects-and-flows/overview) that require no modification when you add new channels.

## Working with Endpoints

<Tabs>
  <Tab title="GUI">
    You can create, view, edit, and delete Endpoints in **Deploy > Endpoints**.
  </Tab>

  <Tab title="API">
    You can view, create, and edit Endpoints using the [Cognigy
    API](https://api-trial.cognigy.ai/openapi#tag--Endpoints-v2.0).
  </Tab>

  <Tab title="CLI">
    You can manage Endpoints using the [Cognigy
    CLI](https://github.com/Cognigy/Cognigy-CLI).
  </Tab>
</Tabs>

### General Endpoint Settings

For every Endpoint, you have the following basics:

| Setting         | Description                                                                                                                                                                                                                                                                                                                                                                                                                |
| --------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Snapshot        | This setting is optional. Select the [Snapshot](/ai/agents/deploy/snapshots) of the [Project](/ai/agents/develop/projects-and-flows/projects) you want to use.<br /><br />By default, Endpoints use the [Connections](/ai/platform-features/connections) from the Snapshot. If you select a Snapshot, the [Override Snapshot Connections](#override-snapshot-connections) toggle is displayed under the **Snapshot** list. |
| Locale          | This setting is optional. Select the [Locale](/ai/platform-features/translation-and-localization/localization) for the Endpoint. The Locale list only appears if you've created a Locale beforehand.                                                                                                                                                                                                                       |
| Flow            | Select the [Flow](/ai/agents/develop/projects-and-flows/overview) to use for the Endpoint.<br /><br />You can enter manually by clicking **Manual Input** below the Flow list. Click <img alt="jump to Flow" /> next to the **Flow** list to open the selected Flow in the Flow editor.                                                                                                                                    |
| Endpoint URL    | The Endpoint URL serves as the interface to the Endpoint. For most third-party platforms, you need to copy it into the channel's settings to configure the connection to Cognigy.                                                                                                                                                                                                                                          |
| NLU Connector   | Select the [NLU Connector](/ai/platform-features/nlu/external/nlu-connectors/overview) for the Endpoint. To disable Intent and Slot recognition, select the [No NLU](/ai/platform-features/nlu/external/nlu-connector-reference/no-nlu) option.                                                                                                                                                                            |
| Enable Endpoint | Toggle to activate or deactivate the Endpoint. This setting is enabled by default.                                                                                                                                                                                                                                                                                                                                         |

#### Override Snapshot Connections

If you activate the **Override Snapshot Connections** toggle, the Endpoint overrides the [Connections](/ai/platform-features/connections) in the Snapshot with Connections from the Project. If only the Snapshot has a Connection, the Connection isn't overridden. If necessary, you need to update such Connections manually.

To override the Connection of the AI Agent in production, you can:

* Make the necessary changes in a staging environment, create a Snapshot of the same AI Agent, and select the Snapshot in the Endpoint settings. Deactivate the **Override Snapshot Connections** toggle.
* Import Connections from another AI Agent using a [Package](/ai/platform-features/packages). Activate the **Override Snapshot Connections** toggle.

### Mocking

In the **Advanced Settings** section, you can toggle on the **Enable Mocking** setting to activate [mocking](/ai/agents/test/mocking) for the Endpoint. Mocking lets you change Node behaviors and test how they perform through the Endpoint without changing the original Node settings. Make sure that you also activate mocking at the [Node level](/ai/agents/test/mocking#at-the-node-level).

## Specific Endpoint Settings

You can use the following additional settings for the Endpoint configuration:

* [Data Protection & Analytics](/ai/agents/deploy/endpoints/data-protection-and-analytics)
* [Transformer Functions](/ai/for-developers/transformers/overview)
* [Session Management](/ai/agents/deploy/endpoints/session-management)
* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)
* [Agent Copilot](/agent-copilot/overview)
* [File Storage](/ai/agents/deploy/endpoints/file-storage)
* [Copilot Section](/agent-copilot/configure/copilot-section)

## More Information

* [Localization](/ai/platform-features/translation-and-localization/localization)


# Real-Time Translation Settings
Source: https://docs.cognigy.com/ai/agents/deploy/endpoints/real-time-translation-settings



In the *Real-Time Translation Settings* section, you can configure the [Real-Time Translation](/ai/platform-features/translation-and-localization/real-time-translation) feature. Before using this feature, you need to add a machine translation provider.

## Add a Machine Translation Provider

1. In the left-side menu of the Project, go to **Manage > Settings**.
2. In the **Translation settings** section, select one of the following providers and configure the settings:

<Tabs>
  <Tab title="Microsoft Translator">
    | Setting                       | Description                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | ----------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Translation Provider API Key  | Enter your unique API key provided by Microsoft Translator. It's necessary for authentication and allows you to access translation services.                                                                                                                                                                                                                                                                                                           |
    | Request Retries               | Enter how often Cognigy should attempt to call the translation provider if an error occurs during the initial request. For example, if you set it to `3`, Cognigy tries three times before giving up on a translation request.                                                                                                                                                                                                                         |
    | Request Timeout               | Enter how long Cognigy should wait for a response from the translation provider after making a request, in milliseconds. If the provider doesn't respond within this time frame, it's considered a timeout.                                                                                                                                                                                                                                            |
    | Sentence Cache Expiry Timeout | Enter how long translations should be kept in the cache before they are considered outdated and removed. The default value is `84,600` seconds, which corresponds to 1 day. For no caching, set it to `0`.                                                                                                                                                                                                                                             |
    | Custom API Base URL           | This setting is optional. If you have specific requirements or need to connect to a custom instance of the Microsoft Translator API different from the default one `https://api.cognitive.microsofttranslator.com/`, you can use this field to specify a custom base URL. It allows you to define the URL, including the protocol scheme, for example `https://api-eur.cognitive.microsofttranslator.com`, to tailor the API connection to your needs. |
    | Custom Subscription Region    | This setting is optional. It represents the location or region of your Azure AI Translator resource. You may need to use this field when calling this API, especially if you have specified a specific region in your Azure account.                                                                                                                                                                                                                   |
  </Tab>

  <Tab title="Google Cloud Translation">
    | Setting                       | Description                                                                                                                                                                                                                    |
    | ----------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Translation Provider API Key  | Enter your unique API key provided by Google Cloud Translation. It's necessary for authentication and allows you to access translation services.                                                                               |
    | Request Retries               | Enter how often Cognigy should attempt to call the translation provider if an error occurs during the initial request. For example, if you set it to `3`, Cognigy tries three times before giving up on a translation request. |
    | Request Timeout               | Enter how long Cognigy should wait for a response from the translation provider after making a request, in milliseconds. If the provider doesn't respond within this time frame, it's considered a timeout.                    |
    | Sentence Cache Expiry Timeout | Enter how long translations should be kept in the cache before they are considered outdated and removed. The default value is `84,600` seconds, which corresponds to 1 day. For no caching, set it to `0`.                     |
  </Tab>

  <Tab title="DeepL Translate Pro">
    | Setting                       | Description                                                                                                                                                                                                                                                                                                                                     |
    | ----------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Translation Provider API Key  | Enter your unique API key provided by DeepL Translate Pro. It's necessary for authentication and allows you to access translation services.                                                                                                                                                                                                     |
    | Request Retries               | Enter how often Cognigy should attempt to call the translation provider if an error occurs during the initial request. For example, if you set it to `3`, Cognigy tries three times before giving up on a translation request.                                                                                                                  |
    | Request Timeout               | Enter how long Cognigy should wait for a response from the translation provider after making a request, measured in milliseconds. If the provider doesn't respond within this time frame, it's considered a timeout.                                                                                                                            |
    | Sentence Cache Expiry Timeout | Enter how long translations should be kept in the cache before they are considered outdated and removed. The default value is `84,600` seconds, which corresponds to 1 day. For no caching, set it to `0`.                                                                                                                                      |
    | Custom API Base URL           | This setting is optional. If you have specific requirements or need to connect to a custom instance of the DeepL Translate Pro API, you can use this field to specify a custom base URL. It allows you to define the URL, including the protocol scheme, for example `https://api-free.deepl.com/`, to tailor the API connection to your needs. |
  </Tab>
</Tabs>

3. Click **Save**.

## Real-Time Translation Settings

<a href="/release-notes/4.63"><Badge>Updated in 4.63</Badge></a>

| Parameter                                  | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         | Providers                              |
| ------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------- |
| Enable Translation                         | Activate real-time translation for the Endpoint.                                                                                                                                                                                                                                                                                                                                                                                                                                                                    | Microsoft, Google, DeepL Translate Pro |
| User Input Language                        | Select the language of the user input or the **Auto-Detect** option. This option detects the language of the user automatically in their first messages.                                                                                                                                                                                                                                                                                                                                                            | Microsoft, Google, DeepL Translate Pro |
| Flow Language                              | Select the AI Agent's language.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     | Microsoft, Google, DeepL Translate Pro |
| Set user input language on execution count | This setting applies if you've selected the **Auto-Detect** option in the **User Input Language** list. Set the number of user inputs the AI Agent uses to determine the language of the user. After this number of user inputs, the AI Agent stops trying to detect the user language.                                                                                                                                                                                                                             | Microsoft, Google, DeepL Translate Pro |
| Glossary ID Output                         | Enter the glossary ID for the AI Agent's output translations. To ensure that the translations work correctly, the source and target languages of the glossary must match the selected language in the **Flow Language** and **User Input Language** fields.                                                                                                                                                                                                                                                         | DeepL Translate Pro                    |
| Glossary ID Input                          | Enter the glossary ID for user input translations. To ensure that the glossary works correctly, the source and target languages of the glossary must match the selected language in the **Flow Language** and **User Input Language** fields. If you've selected **Auto-Detect** in the **User Input Language** field, the glossary isn't used for input translations.                                                                                                                                              | DeepL Translate Pro                    |
| Formality                                  | Select the level of formality for the translated text. This setting applies only to the following target languages: <ul><li>DE (German)</li><li>FR (French)</li><li>IT (Italian)</li><li>ES (Spanish)</li><li>NL (Dutch)</li><li>PL (Polish)</li><li>PT-BR (Portuguese)</li><li>PT-PT (Portuguese)</li><li>JA (Japanese)</li><li>RU (Russian)</li></ul>                                                                                                                                                             | DeepL Translate Pro                    |
| No-Translation Marker                      | Enter the marker for excluding text segments from translation. Any text enclosed in the no-translation marker isn't translated. By default, the no-translation marker is `**`. In this case, if you enter, for example, `Willkommen, **Moin**!` in the **Text** field of a Say Node and the detected user language is English, the AI Agent outputs `Welcome, Moin!` to the user. `Moin` isn't translated as `Good Morning`. This setting can be useful, for example, to avoid translating product and brand names. | Microsoft, Google, DeepL Translate Pro |
| Always remove No-Translation Markers       | When you activate this setting, the AI Agent always removes the no-translation markers before outputting any text, even if you deactivate real-time translation.                                                                                                                                                                                                                                                                                                                                                    | Microsoft, Google, DeepL Translate Pro |
| Prevent payloads from being translated     | When you activate this setting, user inputs received via payload aren't translated.                                                                                                                                                                                                                                                                                                                                                                                                                                 | Microsoft, Google, DeepL Translate Pro |

## More Information

* [Real-Time Translation](/ai/platform-features/translation-and-localization/real-time-translation)
* [Automatic Translation](/ai/platform-features/translation-and-localization/auto-translation)


# Session Management
Source: https://docs.cognigy.com/ai/agents/deploy/endpoints/session-management



The **Session Management** section exists on some Endpoints. Here you can set the session expiration. Once a session has expired, the conversation will restart with the next user input.

The following Endpoints support session expiration:

* [Facebook (Meta) Messenger](/ai/agents/deploy/endpoint-reference/facebook-messenger)
* [Line](/ai/agents/deploy/endpoint-reference/line-messenger)
* [Slack](/ai/agents/deploy/endpoint-reference/slack)
* [TwilioSMS](/ai/agents/deploy/endpoint-reference/twilio)
* [Azure Bot Services](/ai/agents/deploy/endpoint-reference/azure-bot-services)
* [WhatsApp](/ai/agents/deploy/endpoint-reference/whatsapp)

<AccordionGroup>
  <Accordion title="Default Session Expiration">
    The default session expiration is set to 30 minutes.
  </Accordion>

  <Accordion title="Disable Session Expiration">
    You can disable session expiration by setting it to 0. This will prevent a user from having multiple sessions in the **Contact Profile**, which reduces the **Analytics** effectiveness.
  </Accordion>
</AccordionGroup>

## Limitations

* By default, the limit is 30 minutes. For on-premises installations, this value can be overridden if a lower value is set in the [Management UI](/ai/administer/access/management-ui#change-expiration-values) or via the `MAX_SESSION_STATE_TTL_IN_MINUTES` environment variable.


# Test Mode for Endpoints
Source: https://docs.cognigy.com/ai/agents/deploy/endpoints/test-mode



*Test Mode* allows you to send messages to Endpoints without affecting the [billable conversation count](/ai/administer/billing#billable-conversations). The messages you send are processed as regular messages between a user and an AI Agent.

## Key Benefits

* **Availability Check**. You can periodically check the availability of your AI Agent.
* **End-To-End Tests**. You can run end-to-end tests in the actual conditions in which AI Agents interact with users and analyze their performance.

## Restrictions

The following Endpoint types support test mode:

* [REST](/ai/agents/deploy/endpoint-reference/rest), for Cognigy 4.27.0 and later versions.
* [Socket.IO](/ai/agents/deploy/endpoint-reference/socketio), for Cognigy 4.42.0 and later versions.

## How to Use Test Mode for Endpoints

### REST

To send a message in test mode, prepend `/test/` to the `<URL_TOKEN>` in the `POST` request, as follows:

```text theme={null}
POST https://<ENDPOINT_BASE_URL>/test/<URL_TOKEN>
```

### Socket.IO

To send a message in test mode, establish a Socket.IO connection with the Socket.IO Endpoint and pass the `testMode` parameter with `true` to `SocketClient`.

#### Use Constructor Options

```js theme={null}
const { SocketClient } = require("@cognigy/socket-client");

const client = new SocketClient("https://socket.url", "socket-token", {
  testMode: true,
});
```

#### Use Manual Socket.IO Setup

```js theme={null}
socketClient = new SocketClient({ /* provide your config parameters */ })
//...

if (config.testMode) {
	socketClient.io.opts.query["testMode"] = "true";
}

socketClient.connect();
//...
```

The examples may differ depending on which Socket.IO client library you use.

## Impact on Billing

The messages you send in test mode don't increase the billable message count in a conversation. However, in addition to the billable conversation count, Cognigy also tracks the total conversation count, which includes test mode messages. Both counts are included in Cognigy's [billing reports](/ai/administer/billing) to prevent misuse. You can send up to 600 test messages per hour. Any rate above this limit may be considered misuse.


# Snapshots
Source: https://docs.cognigy.com/ai/agents/deploy/snapshots



<a href="/release-notes/2026.16"><Badge>Updated in 2026.16</Badge></a>

A *Snapshot* is an immutable version of a Project that you can use to publish AI Agents through [Endpoints](/ai/agents/deploy/endpoints/overview). Snapshots are intended for production. They offer maximum performance and aren't affected by changes in the Project configuration.

You can also download a Snapshot as a Cognigy Snapshot file. Snapshots files have a `.csnap` extension and can be used only in Cognigy. Also, they are encrypted because they may include sensitive information, such as data in [Connections](/ai/platform-features/connections).

You can use Snapshots for the following purposes:

* [Deploy an AI Agent version through Endpoints](#deploy-a-snapshot) and use it as the production version while your developers work on the next version.
* [Back up](#restore-a-snapshot) your AI Agent in case you need to return to the back up one.
* [Transfer](#deploy-ai-agents-to-production-with-snapshots) an AI Agent between different CI/CD environments.

A Snapshot includes the following resources:

* [AI Agents](/ai/agents/overview)
* [Connections](/ai/platform-features/connections)
* [Extensions](/ai/for-developers/extensions)
* [Flows](/ai/agents/develop/projects-and-flows/overview)
* [Functions](/ai/for-developers/functions)
* [Handover Providers](/ai/escalate/handovers)
* [Large Language Models](/ai/agents/develop/gen-ai-and-llms/llms)
* [Lexicons](/ai/platform-features/nlu/slots/user-defined/lexicon)
* [Localization](/ai/platform-features/translation-and-localization/localization)
* [Natural Language Understanding (NLU) Connectors](/ai/platform-features/nlu/external/nlu-connectors/overview)
* [Playbooks](/ai/agents/test/playbooks)
* [Settings](/ai/agents/develop/projects-and-flows/projects#working-with-projects)
* [Tokens](/ai/platform-features/tokens)
* [Trained NLU Intent models](/ai/platform-features/nlu/intents/overview)

## Restrictions

* Snapshots don't include the following resources:

  * Endpoints
  * Intent Trainer records
  * Analytics data
  * Contact Profiles
  * Logs
  * Knowledge AI
  * Other Snapshots

* The `text-embedding-3-model` [model](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) increases the size of Snapshots, leading to longer download and upload times.

* Earlier versions of Cognigy might not support Snapshots created in newer versions and can cause unexpected errors.

* Snapshot creation fails with an error if any JSON key in the Project's data contains a period (`.`). Periods are reserved for dot notation used to access nested properties in Cognigy, for example, `context.user.name`. Avoid using periods in JSON keys across all Cognigy resources and other data included in the Snapshot.

## Limitations

* You can create up to 10 Snapshots per AI Agent. For on-premises installations, you can change this value using the `MAX_AMOUNT_SNAPSHOTS_IN_AGENT` environment variable in `values.yaml`.
* The maximum Snapshot size depends on how you create or upload it:
  * **Cognigy UI**: You can create and upload Snapshots of up to 8 GB. For on-premises installations, you can increase this limit by setting the `RESUMABLE_UPLOAD_MAX_FILE_SIZE` environment variable in `values.yaml`. The same limit also applies to [Packages](/ai/platform-features/packages).
  * **Cognigy API**: You can create and upload Snapshots of up to 2 GB. Snapshots that exceed this limit can't be created or uploaded. For on-premises installations, you can increase this limit by setting the `SNAPSHOT_MAX_FILE_SIZE` environment variable in `values.yaml`.
* You have up to 1 day to download a Snapshot after you've prepared the download. After this period, you need to prepare the download again.

## Working with Snapshots

<Danger>
  If you delete a Snapshot that you use with an Endpoint in production, the AI Agent connected to that Endpoint stops working.
</Danger>

<Tabs>
  <Tab title="GUI">
    You can create, delete, and [restore](#restore-a-snapshot) Snapshots in **Deploy > Snapshots**. Additionally, you can download and upload Snapshots:

    <Accordion title="Download a Snapshot">
      To download a Snapshot, follow these steps:

      1. On the **Snapshots** page, select <img alt="menu" /> **> Prepare Download**. A new [task](/ai/agents/develop/projects-and-flows/projects) named **Prepare Download** is created.
      2. Once the task is completed, click <img alt="snapshot-download" /> to download the Snapshot.
    </Accordion>

    <Accordion title="Upload a Snapshot">
      To upload a Snapshot, follow these steps:

      1. On the **Snapshots** page, click **Upload Snapshot**.
      2. Select a Snapshot from your computer. When the file starts uploading, you will see a dialog window with a progress bar.
      3. Once the file has been uploaded, a [task](/ai/agents/develop/projects-and-flows/projects) named **Upload Snapshot** is created. After the task is completed, the Snapshot is successfully uploaded.
    </Accordion>
  </Tab>

  <Tab title="API">
    You can view, create, delete, and [restore](#restore-a-snapshot) Snapshots using the [Cognigy API](https://api-trial.cognigy.ai/openapi#get-/v2.0/snapshots) as well as download and upload them.
  </Tab>

  <Tab title="CLI">
    You can create Snapshots using the [Cognigy CLI](https://github.com/Cognigy/Cognigy-CLI).
  </Tab>
</Tabs>

### Deploy a Snapshot

To deploy a Snapshot, select it from the **Snapshot** list in the [Endpoint settings](/ai/agents/deploy/endpoints/overview).

While an Endpoint is active, you can replace a Snapshot or remove it in the Endpoint settings. Active sessions continue to use the old Snapshot until they time out or you delete the old Snapshot. If you replace a Snapshot and delete it, the session is reset using the new Snapshot.

If you use the [Persistent History](https://github.com/Cognigy/Webchat/blob/master/docs/persistent-history.md) for Webchat, the user's browser retains the chat history. Even after the session expires, users still see old messages, but any new message starts a new session.

### Restore a Snapshot

You can use Snapshots to back up AI Agents versions. With this approach, you can restore a back up version if the new AI Agent doesn't perform as intended.

Before restoring a Snapshot, be aware of the following:

* **Restoring a Snapshot overwrites all the current AI Agent's resources**. All the AI Agent's resources are deleted before they're recreated from the Snapshot you restore. In case you have active Endpoints pointing to the Flows of the AI Agent, the conversations stop working because the Flows are deleted.
* **Restoring a Snapshot might require manual changes to the Endpoints**. Restoring a Snapshot automatically updates only the Endpoints that are assigned to the primary Locale. You need to update Endpoints assigned to non-primary Locales manually. These Endpoints are marked with a red dot next to the Endpoint name after you restore the Snapshot, meaning that the settings are not complete.

## Deploy AI Agents to Production with Snapshots

The following three scenarios show the recommended methods for deploying AI Agents into production.

<AccordionGroup>
  <Accordion title="Develop an AI Agent with a Single Project">
    With Snapshots, you can make your AI Agents available through Endpoints while you work on the next version of your AI Agent:

    1. Develop and test Flows, Lexicons, Intents, and all other resources as usual. Eventually, you reach an AI Agent version that you want to publish to the users.
    2. Create a Snapshot of the AI Agent. You don't have to prepare a downloadable Package for the Snapshot because everything remains in the current Project.
    3. To release the AI Agent version of the Snapshot, go to the Endpoint settings, for example, in the Webchat Endpoint settings, and select the new Snapshot and the Flow.<br />Your production environment now uses the AI Agent from the Snapshot.

    Since you are using resources from the Snapshot in production, developers can change the Flows in the AI Agent without breaking the current conversations with your customers.
  </Accordion>

  <Accordion title="Develop an AI Agent with Multiple Projects">
    You want to build an AI Agent for FAQs on your website. In this case, you have two Projects:

    * **FAQ Bot development** — a Project for developers to build and test the AI Agent.
    * **FAQ Bot production** — a Project for the production environment.

    1. Developers only have access to the FAQ Bot development Project, in which they can develop and test AI Agents. Eventually, they reach an AI Agent version that you want to publish to the users.
    2. Create a Snapshot of the AI Agent in the FAQ Bot development Project and download it.
    3. To release the AI Agent version of the Snapshot, upload the Snapshot to the FAQ Bot production Project. Then, select the uploaded Snapshot in the Endpoints settings of your production Endpoint.
  </Accordion>

  <Accordion title="Develop an AI Agent with a CI/CD Environments">
    You have the following environments:

    * **Dev** — an environment for development purposes.
    * **Staging** — an environment for staging purposes, where acceptance tests are conducted for newer versions.
    * **Production** — an environment for publishing AI Agents to users.

    1. Developers only have access to Projects in the Dev installation, in which they build AI Agents, design new experiences, or improve already existing solutions. Eventually, they reach an AI Agent version that you want to publish to the users.
    2. Create Snapshots of the AI Agents in the Dev installation, download the Snapshots, and provide them to the teams working on the Staging installation.
    3. People working on the Staging installation receive the Snapshots and upload them to their Projects. They select the uploaded Snapshots in their Endpoint settings and run acceptance tests.
    4. If they find errors, they inform their development teams, who fix the problems and provide new Snapshots. If tests are successful, the Snapshots are moved to the Production installation.
    5. To release the AI Agent version of the Snapshots, upload them to the Production installation and select the uploaded Snapshots in the production Endpoints.
  </Accordion>
</AccordionGroup>

## More Information

* [Endpoints](/ai/agents/deploy/endpoints/overview)
* [Flows](/ai/agents/develop/projects-and-flows/overview)


# Context
Source: https://docs.cognigy.com/ai/agents/develop/ai-agent-memory/context



<a href="/release-notes/2026.16"><Badge>Updated in 2026.16</Badge></a>

The *Context Object* is a JSON object that persists throughout the entire chat or call session. If a user disconnects from Cognigy, the Context object can be retrieved upon reconnection with the same session ID. By default, the Context object is empty. You need to set a [Default Context](#default-context) or [configure](#storing-information-in-the-context-object) what data is stored in this object.

[Nodes](/ai/agents/develop/nodes/overview) can access the Context object and store information in it, for example, user messages, [Slots](/ai/platform-features/nlu/slots/overview), [Intents](/ai/platform-features/nlu/intents/overview), or any other data available throughout the session.

## Limitations

* By default, you can store up to 256 kB in the Context object, and this limit can be raised up to 1.5 MB. The limit is based on JavaScript object size (UTF-16 encoding), so the actual raw text size may be smaller than expected. For example, 1 MB of text can take up about 2 MB in memory. For on-premises installations, you can change this limit using the `MAX_MEMORY_OBJECT_SIZE` variable in the `values.yaml` file. Increasing the limit raises memory usage and may slow performance.

## Working with the Context Object

<Tabs>
  <Tab title="GUI">
    **Interaction Panel**

    You can view the Context object by navigating to **Info > Context** in the [Interaction Panel](/ai/agents/test/interaction-panel/overview). In the **Context** tab, you can save, cancel, or reset the Context object by clicking the respective buttons at the bottom of the Interaction Panel. The changes to the values in the Context object apply only to the current session.

    You can copy the exact JSON path of a Context object value by right-clicking it and selecting **Copy JSON Path**.

    **Flow Editor**

    In the Flow editor, you can set a [Default Context](#default-context) in the **Settings** tab.
  </Tab>

  <Tab title="API">
    You can inject data into and reset the Context object of a session with the [Cognigy API](https://api-trial.cognigy.ai/openapi#tag--Session-v2.0).
  </Tab>
</Tabs>

### Default Context

The *Default Context* is the Flow's Context object when the user triggers the Flow execution for the first time. By default, the Default Context is empty, and you can customize it in the **Settings** tab of the Flow editor.

A session inherits the Default Context from the Flow in which this session starts. If the user switches to another Flow during the session, the Context object remains the same. However, you can apply the Default Context of the target Flow with the [Go To Node](/ai/agents/develop/node-reference/logic/go-to). To do so, activate the Absorb Context parameter in the Go To Node settings that switches to the target Flow.

### Storing Information in the Context Object

Nodes, such as the [Add To Context](/ai/agents/develop/node-reference/logic/add-to-context), [AI Agent](/ai/agents/develop/node-reference/ai/ai-agent), [Question](/ai/agents/develop/node-reference/other-nodes/search-extract-output), and others, include different parameters to store information in the Context object.

### Accessing the Context object

Use Nodes to dynamically access Context properties with [CognigyScript](/ai/platform-features/cognigyscript), for example, `{{context.property}}`, or [Tokens](/ai/platform-features/tokens). The CognigyScript expression used to access the Context object follows the dot notation `property.child.child`.

#### Example

Consider the following Context object:

<Accordion title="Context Object Example">
  ```json theme={null}
  {
      "order": {
          "id": "ORD-2023-1234",
          "items": [
              {
                  "name": "Margherita Pizza",
                  "category": "food",
                  "size": "large",
                  "quantity": 1,
                  "price": 15.99
              },
              {
                  "name": "Caesar Salad",
                  "category": "food", 
                  "size": "regular",
                  "quantity": 1,
                  "price": 8.99
              },
              {
                  "name": "Coca Cola",
                  "category": "drink",
                  "size": "medium",
                  "quantity": 2,
                  "price": 2.99
              }
          ],
          "subtotal": 30.96,
          "tax": 2.48,
          "deliveryFee": 5.00,
          "total": 38.44,
          "deliveryAddress": "123 Main St, Apt 4B",
          "orderTime": "2023-10-25T19:30:00Z",
          "status": "preparing"
      }
  }
  ```

  * `{{context.order.items[0].price}}` returns the price of the first item in the order.
  * `{{context.order.items[1].name}}` returns the name of the second item in the order.
  * `{{context.order.items[2].quantity}}` returns the quantity of the third item in the order.
</Accordion>

## More Information

* [Profile](/ai/agents/develop/ai-agent-memory/profile)


# Input
Source: https://docs.cognigy.com/ai/agents/develop/ai-agent-memory/input



<a href="/release-notes/4.90"><Badge>Updated in 4.90</Badge></a>

The *Input Object* is a JSON object that serves as the short-term AI Agent memory. The Input object is generated each time the user sends a message or provides a voice input to an AI Agent.

The Input object stores:

* General data about the user input, such as time and type of input
* Endpoint-specific data
* Results from the NLU Intent mapping process, such as the recognized [Intent](/ai/platform-features/nlu/intents/overview) and [Slots](/ai/platform-features/nlu/slots/overview)

## Limitations

* You can store up to 1.5 MB in the Input object. For on-premises installations, you can configure the limit using the `MAX_MEMORY_OBJECT_SIZE` variable in the `values.yaml` file. Increasing the object size could lead to performance issues.

## Properties

<Note>
  The data in the Input object depends on the Endpoint through which the user input is received. This article doesn't include all possible property variations.
</Note>

<Warning>
  As of Cognigy 4.91.0, the Understood Messages concept is deprecated. Use the [Goals and Tasks](https://docs.cognigy.com/ai/analyze/goals-and-tasks/overview/) concepts to evaluate whether your AI Agent understood a particular phrase. In Cognigy 2025.19, the charts related to this concept were removed from Insights reports.

  As of Cognigy 4.91.0, the `understood` field in the Cognigy OData endpoint is deprecated. The `understood` field has been removed from the Cognigy OData endpoint. Use the [Goals collection](https://docs.cognigy.com/ai/analyze/odata#goals) in the Cognigy OData endpoint.

  Refer to the list of all upcoming removals on the [Deprecations and Removals](https://docs.cognigy.com/release-notes/deprecations-and-removals/) page.
</Warning>

The Input object contains the following properties:

<Accordion title="Input Object Properties">
  | Key                                                          | Type                      | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
  | ------------------------------------------------------------ | ------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
  | `text`                                                       | String                    | The text of the user input.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
  | `intent`<sup>[1](#footnote1)</sup>                           | String                    | The detected Intent name. If no Intent is detected, the property value is `null`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
  | `intentScore`<sup>[1](#footnote1)</sup>                      | Number                    | The detected Intent score. If no Intent is detected, the property value is `null`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
  | `intentFlow`                                                 | String                    | The Flow ID in which the Intent is detected. If no Intent is detected, the property value is `null`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |
  | `slots`                                                      | Array                     | Contains the detected [system-defined](/ai/platform-features/nlu/slots/system-defined) and [user-defined Slots](/ai/platform-features/nlu/slots/user-defined/overview).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
  | `nlu`<sup>[1](#footnote1)</sup>                              | [Object](#nested-objects) | Contains the detailed NLU processing results.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
  | `mode`                                                       | String                    | The type of input received by Cognigy. Can be either **TextOnly** or **TextData**.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
  | `type`<sup>[1](#footnote1)</sup>                             | String                    | The type of input sentence as determined by the NLU engine (Statement, Command, Greeting, BGreeting, whQuestion, howQuestion, ynQuestion, pAnswer, nAnswer).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
  | `intentOutOfState`                                           | String                    | Contains the detected Intent that has been excluded in the current Flow state. If no detected Intent was excluded in the current Flow state, the value is `null`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
  | `currentTime`                                                | [Object](#nested-objects) | Contains the timestamp data of the user input reception.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
  | `language`                                                   | String                    | The locale language that was used to process the user input. For example, `en-US`, `de-DE`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
  | `ip`                                                         | String                    | The IP address of the user. If the IP address isn't available, for example, due to [privacy or masking settings](/ai/agents/deploy/endpoints/data-protection-and-analytics), the value is `null`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
  | `hashedIp`                                                   | String                    | The hashed IP address of the user.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
  | `state`                                                      | String                    | The current state of the Flow. By default, the value is `default`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
  | `channel`                                                    | String                    | The channel through which the user input was received. This property is also available in the OData records to allow channel-based analytics filtering.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
  | `endpointType`                                               | String                    | The type of Endpoint through which the message was received, for example, Alexa or Facebook (Meta) Messenger. You can't change this value.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
  | `entrypoint`                                                 | String                    | The ID of this property varies based on Snapshot usage:<ul><li>The Project ID when no Snapshot is used.</li><li>The Snapshot ID when a Snapshot is used.</li></ul>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
  | `userId`                                                     | String                    | The current user's ID. Cognigy uses the user ID to identify their individual contact profile.<br />See [Contact Profiles](/ai/agents/analyze/contact-profiles) for more information.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |
  | `inputId`                                                    | String                    | A unique ID that is generated for each user input received by Cognigy.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
  | `sessionId`                                                  | String                    | The current session ID. The session ID either provided by the channel or generated by Cognigy. Most Endpoints allow defining an individual session expiration (TTL).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |
  | `flowName`                                                   | String                    | The name of the Flow in execution.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
  | `URLToken`                                                   | String                    | The [Endpoint URL Token](/ai/agents/deploy/endpoints/overview).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
  | `frustration`                                                | Number                    | Indicates user frustration by representing how often the user repeatedly takes the same path in the Flow.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         |
  | `completedGoals`                                             | Array                     | The [Goals](/insights/reports/goals) that have been at the current stage in the Flow.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
  | `execution`                                                  | Number                    | The number of messages received in the current session.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
  | `traceId`                                                    | String                    | A unique identifier for the user input.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
  | `activeQuestion`                                             | [Object](#nested-objects) | Contains information about an active question.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
  | `apps`                                                       | [Object](#nested-objects) | Contains information about the xApps that are active in the current session.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
  | `intentLevel`<sup>[1](#footnote1)</sup>                      | [Object](#nested-objects) | Contains information about the Intent Hierarchy level of the detected Intent. If no Intent was detected, the object is `{"level1": null, "level2": null, "level3": null}`                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         |
  | `understood`<sup>[1](#footnote1)</sup>                       | Boolean                   | An input is considered `understood` if any of the following conditions are met:<ul><li>An Intent or Slot is identified.</li><li>The input is marked as understood through the Nodes: Code, Overwrite Analytics, Search Extract Output, or AI Agent.</li><li>The sentence type is `pAnswer`, `nAnswer`, or `Greeting`, if the [Confirmation Word logic](/ai/platform-features/nlu/intents/confirmation-words) is activated.</li></ul><br />An input is considered `not understood` if any of the following conditions are met:<ul><li>There is an active handover without triggering an Agent Copilot Whisper Flow.</li><li>The message is explicitly marked as `Don't count` or `Null`.</li></ul> |
  | `data`                                                       | Object                    | The data payload of the input. This object depends on the Endpoint through which the user input is received.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
  | `attachments`                                                | Array                     | An array of URLs pointing to the attachments uploaded to the chat.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
  | `localeId`<sup>[1](#footnote1)</sup>                         | String                    | The unique ID for the current locale used to process the input message.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
  | `result`                                                     | String                    | The answer from the user to a question from a [Question](/ai/agents/develop/node-reference/basic/question) or [Optional Question](/ai/agents/develop/node-reference/basic/optional-question).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
  | `conditionalEntrypointWasExecuted`<sup>[1](#footnote1)</sup> | Boolean                   | Indicates if the Flow was executed on a conditional Node.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         |
  | `transcript`                                                 | [Object](#nested-objects) | Contains information about the conversation transcript generated via the Get Transcript Node.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
</Accordion>

### Nested Objects

<AccordionGroup>
  <Accordion title="nlu Object">
    | Key                   | Property                            |                          | Type    | Description                                                                                                                                                                                                                            |
    | --------------------- | ----------------------------------- | ------------------------ | ------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | `intentMapperResults` |                                     |                          | Object  | Provides details about the NLU Intent mapping results.                                                                                                                                                                                 |
    |                       | `finalIntentName`                   |                          | String  | The name of the detected Intent.                                                                                                                                                                                                       |
    |                       | `finalIntentScore`                  |                          | Number  | The score of the detected Intent.                                                                                                                                                                                                      |
    |                       | `finalIntentNegated`                |                          | Boolean | Indicates whether the detected Intent was negated.                                                                                                                                                                                     |
    |                       | `finalIntentConfirmationSentence`   |                          | String  | The confirmation sentence of the detected Intent.                                                                                                                                                                                      |
    |                       | `finalIntentDisambiguationSentence` |                          | String  | The disambiguation sentence of the detected Intent.                                                                                                                                                                                    |
    |                       | `highestFlow`                       |                          | String  | Reference ID of the Flow in which the highest-scoring Intent was detected.                                                                                                                                                             |
    |                       | `intentPath`                        |                          | String  | The path of the Intent.                                                                                                                                                                                                                |
    |                       | `intentPathIds`                     |                          | Array   | The IDs of the Intents in the path.                                                                                                                                                                                                    |
    |                       | `intentPathFlowReferenceIds`        |                          | Array   | The reference IDs of the Flows in which the Intents in the path were found.                                                                                                                                                            |
    |                       | `scores`                            |                          | Array   | Contains the results for the highest-scoring Intents for the current user input.                                                                                                                                                       |
    |                       |                                     | `id`                     | String  | The ID of the Intent.                                                                                                                                                                                                                  |
    |                       |                                     | `name`                   | String  | The name of the Intent.                                                                                                                                                                                                                |
    |                       |                                     | `score`                  | Number  | The score of the Intent.                                                                                                                                                                                                               |
    |                       |                                     | `negated`                | Boolean | Indicates whether the Intent was negated.                                                                                                                                                                                              |
    |                       |                                     | `confirmationSentence`   | String  | The confirmation sentence of the Intent.                                                                                                                                                                                               |
    |                       |                                     | `confirmationSentences`  | Array   | The confirmation sentences of the Intent.                                                                                                                                                                                              |
    |                       |                                     | `disambiguationSentence` | String  | The disambiguation sentence of the Intent.                                                                                                                                                                                             |
    |                       |                                     | `flow`                   | String  | The reference ID of the Flow in which the Intent was detected.                                                                                                                                                                         |
    |                       |                                     | `description`            | String  | The description of the detected Intent.                                                                                                                                                                                                |
    | `intentFlow`          |                                     |                          | String  | A unique ID assigned to the Flow in which the Intent was detected.                                                                                                                                                                     |
    | `uniqueIntentFlowId`  |                                     |                          | String  | A unique identifier for the Flow in which the Intent was detected.                                                                                                                                                                     |
    | `intentId`            |                                     |                          | String  | A unique ID assigned to the Intent that was selected by the NLU processing.                                                                                                                                                            |
    | `uniqueIntentId`      |                                     |                          | String  | A unique identifier for the selected Intent.                                                                                                                                                                                           |
    | `detailedSlots`       |                                     |                          | Array   | Contains the detected [system-defined](/ai/platform-features/nlu/slots/system-defined) and [user-defined Slots](/ai/platform-features/nlu/slots/user-defined/overview), with more detailed information about the system-defined Slots. |
    | `tokens`              |                                     |                          | Array   | Contains the user input tokens.                                                                                                                                                                                                        |
    | `yesNoIntentResults`  |                                     |                          | Object  | Contains the results of the Yes/No Intents model evaluation. See [Yes/No Intents](/ai/platform-features/nlu/intents/yes-no-intents) for more information.                                                                              |
    |                       | `finalIntentName`                   |                          | String  | The name of the detected Yes/No Intent.                                                                                                                                                                                                |
    |                       | `finalIntentScore`                  |                          | Number  | The score of the detected Yes/No Intent.                                                                                                                                                                                               |
    |                       | `scores`                            |                          | Array   | Contains the results for the highest-scoring Yes/No Intents.                                                                                                                                                                           |
    |                       |                                     | `id`                     | String  | The ID of the Intent.                                                                                                                                                                                                                  |
    |                       |                                     | `name`                   | String  | The name of the Intent.                                                                                                                                                                                                                |
    |                       |                                     | `score`                  | Number  | The score of the Intent.                                                                                                                                                                                                               |
    |                       |                                     | `negated`                | Boolean | Indicates whether the Yes/No Intent was negated.                                                                                                                                                                                       |
    |                       |                                     | `confirmationSentence`   | String  | The confirmation sentence of the Yes/No Intent.                                                                                                                                                                                        |
    |                       |                                     | `confirmationSentences`  | Array   | The confirmation sentences of the Yes/No Intent.                                                                                                                                                                                       |
    |                       |                                     | `disambiguationSentence` | String  | The disambiguation sentence of the Yes/No Intent.                                                                                                                                                                                      |
    |                       |                                     | `flow`                   | String  | The reference ID of the Flow in which the Yes/No Intent was detected.                                                                                                                                                                  |
    |                       |                                     | `description`            | String  | The description of the detected Yes/No Intent.                                                                                                                                                                                         |
  </Accordion>

  <Accordion title="currentTime Object">
    | Key              | Type   | Description                              |
    | ---------------- | ------ | ---------------------------------------- |
    | `year`           | Number | The year.                                |
    | `month`          | Number | The month of the year.                   |
    | `day`            | Number | The day of the month.                    |
    | `hour`           | Number | The hour of the day.                     |
    | `minute`         | Number | The minute of the hour.                  |
    | `second`         | Number | The second of the minute.                |
    | `milliseconds`   | Number | The milliseconds of the current time.    |
    | `weekday`        | Number | The day of the week.                     |
    | `dayOfWeek`      | String | The day of the week.                     |
    | `ISODate`        | String | The date and time in ISO 8601 format.    |
    | `plain`          | String | The current date and time in plain text. |
    | `grain`          | String | The grain of the current time.           |
    | `timezoneOffset` | String | The timezone offset of the current time. |
  </Accordion>

  <Accordion title="activeQuestion Object">
    | Key                       | Type   | Description                                                 |
    | ------------------------- | ------ | ----------------------------------------------------------- |
    | `nodeId`                  | String | The question Node's ID.                                     |
    | `type`                    | String | The type of question.                                       |
    | `lastExecutedAt`          | String | The timestamp of the last execution of the active question. |
    | `forgetQuestionThreshold` | Number | The threshold for forgetting the question.                  |
    | `repromptCount`           | Number | The number of times the question was reprompted.            |
    | `escalationCount`         | Number | The number of times the question was escalated.             |
  </Accordion>

  <Accordion title="apps Object">
    | Key             | Property | Type   | Description                                                                                                                                                             |
    | --------------- | -------- | ------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | `customization` |          | Object | Contains the style customization options for the xApp Session as configured in the [xApp: Init Session Node](/ai/agents/develop/node-reference/xApp/init-xApp-session). |
    | `session`       |          | Object | Contains xApp session-specific data.                                                                                                                                    |
    |                 | `token`  | String | The xApp session token used to start a real-time connection with the xApp Shell Page.                                                                                   |
    |                 | `pin`    | String | The xApp session PIN, if the xApp: Get Session PIN Node is used.                                                                                                        |
    | `baseUrl`       |          | String | Base URL of the xApp Shell Page.                                                                                                                                        |
    | `url`           |          | String | The xApp session URL for the current conversation.                                                                                                                      |
  </Accordion>

  <Accordion title="intentLevel Object">
    | Key      | Type   | Description                                                                              |
    | -------- | ------ | ---------------------------------------------------------------------------------------- |
    | `level1` | String | Level 1 Intent name. If no Intent was detected, the property value is `null`.            |
    | `level2` | String | Level 2 Intent name. If no child Intent was detected, the property value is `null`.      |
    | `level3` | String | Level 3 Intent name. If no grandchild Intent was detected, the property value is `null`. |
  </Accordion>

  <Accordion title="transcript Object">
    <Note>
      The `transcript` object is only available if the [Get Transcript](/ai/agents/develop/node-reference/service/get-transcript) Node is used in the Flow.
    </Note>

    | Key         | Property                         | Type   | Description                                                                                                                                                                                                                                     |
    | ----------- | -------------------------------- | ------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | `role`      |                                  | String | Defines who sent the message. Possible values:  - `assistant` — the message is sent by the AI Agent.  - `user` — the message is sent by the end user.  - `agent` — a human agent.  - `system` — a system message.  - `tool` — an external tool. |
    | `type`      |                                  | String | The type of the message. Possible values:  - `input` — a user input.  - `output` — an AI Agent output.  - `toolCall` — a request to a tool.  - `toolAnswer` — a tool response.  - `debugLog` — a system debug message.                          |
    | `source`    |                                  | String | The origin of the message. Possible values are `user`, `assistant`, `agent`, or `system`. Helps distinguish between different actors in the conversation.                                                                                       |
    | `payload`   |                                  | Object | The main content of the message. Its structure depends on the message type.                                                                                                                                                                     |
    |             | **For type `input` or `output`** |        | Represents the messages exchanged between the end user and the AI Agent.                                                                                                                                                                        |
    |             | `text`                           | String | The text of the message.                                                                                                                                                                                                                        |
    |             | `textAlternative`                | String | The alternative text shown if the main payload is non-text, for example, for images, buttons, or other rich media. Available only for messages of the `output` type.                                                                            |
    |             | `data`                           | Object | The additional structured information, such as metadata, attachments, or rich media.                                                                                                                                                            |
    |             | **For type `toolCall`**          |        | Represents a request sent by the AI Agent to an external tool.                                                                                                                                                                                  |
    |             | `name`                           | String | The name of the tool being called.                                                                                                                                                                                                              |
    |             | `id`                             | String | The unique identifier for the tool call.                                                                                                                                                                                                        |
    |             | `input`                          | Object | The parameters or data sent to the tool for processing.                                                                                                                                                                                         |
    |             | **For type `toolAnswer`**        |        | Represents the response returned from a tool call.                                                                                                                                                                                              |
    |             | `toolCallId`                     | String | References the original tool call ID that this answer corresponds to.                                                                                                                                                                           |
    |             | `name`                           | String | The name of the tool that provided the answer.                                                                                                                                                                                                  |
    |             | `content`                        | String | The output content returned by the tool.                                                                                                                                                                                                        |
    |             | **For type `debugLog`**          |        | Represents a system debug message.                                                                                                                                                                                                              |
    |             | `header`                         | String | The header describing the debug log.                                                                                                                                                                                                            |
    |             | `message`                        | String | The detailed debug message content.                                                                                                                                                                                                             |
    |             | `metadata`                       | Object | The additional structured data related to the debug log (for example, key-value pairs for internal debugging).                                                                                                                                  |
    | `id`        |                                  | String | The unique identifier for each message. Ensures that the messages are traceable, maintain order, and prevent duplicates.                                                                                                                        |
    | `traceId`   |                                  | String | The identifier used to group related messages within the same session, allowing easier tracking of conversation threads.                                                                                                                        |
    | `timestamp` |                                  | Number | The UNIX timestamp representing when the message was created. It is useful for ordering and analyzing the message history.                                                                                                                      |
  </Accordion>
</AccordionGroup>

## Working with the Input Object

You can view the Input object by navigating to **Info > Input** in the [Interaction Panel](/ai/agents/test/interaction-panel/overview). You can copy the exact JSON path of an Input object value by right-clicking it and selecting **Copy JSON Path**.

### Accessing Input Properties

Nodes can dynamically access Input properties with [CognigyScript](/ai/platform-features/cognigyscript), for example, `{{input.property}}` or [Tokens](/ai/platform-features/tokens). The CognigyScript expression you use to access the input object follows the dot notation for JSON objects: `property.child.child`.

#### Examples

* `{{input.text}}` returns the text that the user sent to the Flow.
* `{{input.nlu.tokens}}` returns the tokens of the user input in an array.
* `{{input.attachments[0].url}}` returns the URL of the first attachment uploaded to the chat.

## More Information

* [Interaction Panel](/ai/agents/test/interaction-panel/overview)

***

<sup>1</sup>: The NLU properties are computed by the NLU Connector and stored after the NLU processing. For this reason, [Rules](/ai/platform-features/nlu/intents/rules) and [Intent Conditions](/ai/platform-features/nlu/intents/conditions) can't access these properties.


# AI Agent Memory
Source: https://docs.cognigy.com/ai/agents/develop/ai-agent-memory/overview



In Cognigy, the *AI Agent Memory* refers to the data from the user, user input, or conversation the AI Agent can retrieve. The AI Agent memory is stored in Cognigy objects which you can access using [CognigyScript](/ai/platform-features/cognigyscript) and [Tokens](/ai/platform-features/tokens). You can retrieve the stored information to generate context-aware AI Agent answers and control your Flows better, for example, with [Intent Conditions](/ai/platform-features/nlu/intents/conditions).

Cognigy objects are JSON objects that store different types of information:

* [Input object](/ai/agents/develop/ai-agent-memory/input) — the user input information, such as the [Intents](/ai/platform-features/nlu/intents/overview) in the user's message, the [Channel](/ai/agents/develop/nodes/channels) through which the user interacts with the AI Agent, and other relevant metadata.
* [Context object](/ai/agents/develop/ai-agent-memory/context) — session-specific information, such as selected products or the user's current location.
* [Profile object](/ai/agents/develop/ai-agent-memory/profile) — [Contact Profile](/ai/agents/analyze/contact-profiles) data, such as user names and email addresses.

## How to Access the AI Agent's Memory

The following table shows how to access values from the Input, Context, and Profile objects with CognigyScript:

|                  | Input            | Context                     | Profile                 |
| ---------------- | ---------------- | --------------------------- | ----------------------- |
| Variable         | `input`          | `context`                   | `profile`               |
| Shortcut         | `ci`             | `cc`                        | `cp`                    |
| Example          | `{{input.text}}` | `{{context.selectedHotel}}` | `{{profile.firstname}}` |
| Shortcut Example | `{{ci.text}}`    | `{{cc.selectedHotel}}`      | `{{cp.firstname}}`      |

### Lifespan of Cognigy Objects

The lifespan of Cognigy objects varies depending on their type:

* **Input** — each new user input and Flow execution generates a new Input object.
* **Context** — kept throughout the entire chat or call session.
* **Profile** — stored persistently to preserve information, such as user name, throughout conversations.

The following image shows a comparison of the lifespans of Cognigy objects:

<Frame>
  <img alt="Comparison of Input, Context, and Profile object lifespans" />
</Frame>


# Profile
Source: https://docs.cognigy.com/ai/agents/develop/ai-agent-memory/profile



The *Profile Object* is a JSON object that serves as the long-term AI Agent memory. The Profile object stores persistent information about the user.

Each user who connects to Cognigy has a unique Contact Profile. The Contact Profile uses the Profile object to store user data in Cognigy. When the user disconnects from Cognigy, the Contact Profile retains the data and retrieves it upon reconnection with the same user ID. This approach allows multichannel conversations to take place by accessing the stored Contact Profile when the user connects through a different channel.

With [Nodes](/ai/agents/develop/nodes/overview), such as the [Say](/ai/agents/develop/node-reference/basic/say) or [Question](/ai/agents/develop/node-reference/basic/question) Nodes, you can access the Profile object and retrieve information.

## Limitations

* You can store up to 1.5 MB in the Profile object. For on-premises installations, you can configure the limit using the `MAX_MEMORY_OBJECT_SIZE` variable in the `values.yaml` file. Increasing the object size could lead to performance issues.

## Working with the Profile Object

<Tabs>
  <Tab title="GUI">
    You can view the Profile object by navigating to **Info > Profile** in the [Interaction Panel](/ai/agents/test/interaction-panel/overview). You can copy the exact JSON path of a Profile object value by right-clicking it and selecting **Copy JSON Path**.
  </Tab>

  <Tab title="API">
    You can edit a Contact Profile and the respective data with the [Cognigy API](https://api-trial.cognigy.ai/openapi#tag--Profiles-v2.0) to change the data available in the Profile object.
  </Tab>
</Tabs>

### Managing Information in the Profile Object

The data in the Profile object corresponds to the [Contact Profile data](/ai/agents/analyze/contact-profiles), which you can manage in **Settings > Contact Profiles**.

### Accessing the Profile object

Nodes can dynamically access **Profile** properties with [CognigyScript](/ai/platform-features/cognigyscript), for example, `{{profile.property}}`, or [Tokens](/ai/platform-features/tokens). The CognigyScript expression used to access the Profile object follows the dot notation `property.child.child`.

You can also use [Analytics Nodes](/ai/agents/develop/node-reference/analytics/overview), such as the [Add Memory Node](/ai/agents/develop/node-reference/analytics/add-memory), to store data in the Contact Profile.

#### Example Profile Object

Consider the following Profile object:

```json theme={null}
{
  "firstname": "Bastian",
  "lastname": "Bux",
  "email": "b.bux@neverendingstory.com",
  "age": 30
}
```

* `{{profile.firstname}}` returns `Bastian`.
* `{{profile.lastname}}` returns `Bux`.


# LLM Fallback
Source: https://docs.cognigy.com/ai/agents/develop/gen-ai-and-llms/fallback



<a href="/release-notes/2025.25"><Badge>Updated in 2025.25</Badge></a>

*LLM Fallback* is a mechanism designed to maintain continuous service when a main model fails to respond or encounters an error.
In such cases, the system automatically activates an alternative model to handle the request,
minimizing disruptions for AI Agents during [design](/ai/agents/develop/gen-ai-and-llms/generative-ai#design-time-generative-ai-features) and [run-time](/ai/agents/develop/gen-ai-and-llms/generative-ai#run-time-generative-ai-features) operations.
The fallback model is selected for its similar capabilities to the main model, providing redundancy and improving the reliability of the connection.

Using a fallback model provides the following benefits:

* **Improved User Experience**. Automatically switching to a fallback model ensures uninterrupted service, minimizing disruptions when the main model encounters issues.
* **Quick Recovery**. The fallback mechanism allows the system to recover quickly during failures, such as model outages or updates. This approach prevents extended delays and ensures operations resume smoothly without significant downtime.
* **Greater Flexibility**. Fallback models can be deployed in different regions or configured to meet varying resource requirements, providing adaptability. For example, they can serve as regional backups or handle workloads during high-demand periods.

Note that the fallback model serves as a temporary replacement for the main model.
If the system switches to the fallback model, it indicates that the main model requires attention.
This model failure can occur if the credentials are outdated or incorrectly modified, or if the LLM provider doesn't respond.

Using a fallback model complements the [default retry mechanism](/ai/agents/develop/gen-ai-and-llms/llms#retry-mechanism).
If the main model fails to respond, the system first uses the default retry mechanism. If all retry attempts are exhausted, it then switches to the fallback model.

## Restrictions

* Creating fallbacks for embedding models is not supported.
* You can add only one fallback to a main model.

## Create Fallback Models

Creating a fallback model works the same way as creating a main model.
If your Project has only one model, you can add another by choosing one [from the same LLM provider or another provider](/ai/agents/develop/gen-ai-and-llms/providers/all-providers) with similar capabilities.

Some providers, such as [Azure OpenAI](https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/models?tabs=python-secure%2Cglobal-standard%2Cstandard-chat-completions) and [Amazon Bedrock](https://docs.aws.amazon.com/bedrock/latest/userguide/models-regions.html), allow you to use the same model across different regions.\
For example, if you are using Azure OpenAI `gpt-4o` in the West US region as your main model, you can also use`gpt-4o` as a fallback model in the Sweden Central region.

## Add Fallback Models

To add a fallback model, follow these steps:

1. In the left-side menu of the Project, go to **Build > LLM**.
2. Select the model to which you want to add a fallback.
3. Go to the **Fallback** section and activate the **Enable Fallback** setting.
4. Configure the following fields:
   * **Select a Fallback Model** — select a model from the list. If no model is available, [create a new one](#create-fallback-models).
   * **Number of failed requests** — specify the number of failed requests to the main model before the fallback model is triggered. You can specify up to 5 attempts. The default value is `1`.
   * **Duration in minutes of active fallback** — define how long the system should use the fallback model instead of the main model. After the specified time, the fallback will deactivate, and the main model will resume. The duration can be up to 30 minutes. The default value is `30`.
   * **Email Notifications** — add up to 5 email addresses to receive notifications when the fallback model is triggered. Enter each email in a separate input field using the format `example@domain.com`.
5. Click **Save**.

## More Information

* [Overview](/ai/agents/develop/gen-ai-and-llms/llms)
* [LLM Providers](/ai/agents/develop/gen-ai-and-llms/providers/all-providers)
* [Model Support by Feature](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature)
* [Other Operations](/ai/agents/develop/gen-ai-and-llms/other-operations)


# Generative AI
Source: https://docs.cognigy.com/ai/agents/develop/gen-ai-and-llms/generative-ai



<a href="/release-notes/2026.9"><Badge>Updated in 2026.9</Badge></a>

<Note>
  You are subject to the terms of the Generative AI model providers to which you are connecting. Cognigy cannot take responsibility for your use of third-party services, systems, or materials.
</Note>

Generative AI refers to a type of artificial intelligence that creates new, original content, such as images, video, audio, and text, using machine learning algorithms. It works by learning from existing data and producing new content based on that learning.

Cognigy integrates with various [Large Language Models (LLM) Providers](#prerequisites) to enable Generative AI functionality. This functionality is broadly classified as using [LLMs](/ai/agents/develop/gen-ai-and-llms/llms) to:

* [Support the AI Agent builder](#design-time-generative-ai-features) (Design-Time)
* [Enhance the customer experience](#run-time-generative-ai-features) (Run-Time)

To know more about the benefits of integrating Conversational AI with Generative AI platforms, watch this webinar:

<iframe title="Cognigy Sessions: Cognigy Slot Fillers" />

## Prerequisites

Before using Generative AI, you need to have access to an LLM of the following providers:

* [OpenAI](/ai/agents/develop/gen-ai-and-llms/providers/openai). You need to have a paid account or be a member of an organization that provides you access. Open your OpenAI user profile, copy the existing API Key, or create a new one and copy it.
* [Azure OpenAI](/ai/agents/develop/gen-ai-and-llms/providers/microsoft-azure-openai). You need to have a paid account or be a member of an organization that provides you access. Ask your Azure administrator to provide API Key, resource name, and deployment model name.
* [Anthropic Claude](/ai/agents/develop/gen-ai-and-llms/providers/anthropic). You need to be a [member of an organization](https://console.anthropic.com/docs/) that provides you access.
* [Google Vertex AI](/ai/agents/develop/gen-ai-and-llms/providers/google-vertex-ai), [Google Gemini](/ai/agents/develop/gen-ai-and-llms/providers/google-gemini), or [Google GenAI](/ai/agents/develop/gen-ai-and-llms/providers/google-genai). You need to have a paid account or be a member of an organization that provides you access.
* [Aleph Alpha](/ai/agents/develop/gen-ai-and-llms/providers/aleph-alpha). You need to have an [Aleph Alpha account](https://docs.aleph-alpha.com/).
* [Amazon Bedrock](/ai/agents/develop/gen-ai-and-llms/providers/amazon-bedrock). You need to have an [AWS account](https://aws.amazon.com/resources/create-account/) and access to the [Amazon Bedrock](https://docs.aws.amazon.com/bedrock/latest/userguide/what-is-bedrock.html) service.
* [Mistral](/ai/agents/develop/gen-ai-and-llms/providers/mistral). You need to have a [Mistral AI account](https://docs.mistral.ai/getting-started/quickstart/#account-setup) and activate payments for the LLM usage.
* Cognigy's built-in LLM. A free-of-charge LLM for [design-time](#design-time-generative-ai-features) features and available only to customers with SaaS installations.

## Design-Time Generative AI Features

### Configure LLMs for Design-Time Features

1. In **Manage > Settings**, activate **Enable Generative AI Features** in the **Generative AI Settings** section.
2. From the **Design-Time Generative LLM Features** list, select:

   * **Platform-Provided LLM** to use the built-in LLM for the design-time features.
   * a third-party LLM you want to use.

### Generate Cognigy Resources

During the design phase of creating AI Agents, LLMs can be used to generate a variety of Cognigy resources:

<AccordionGroup>
  <Accordion title="Generate Lexicons">
    <Note>
      * The generation of Lexicons for primary NLU languages besides German and English isn't fully supported.
      * The model may not always generate the exact number of Lexicon entries requested, as its output can vary based on interpretation and model behavior.
    </Note>

    To use Generative AI technology for creating a new Lexicon, do the following:

    1. In the left-side menu of the Cognigy interface, click **Build > Lexicons**, then **+ New Lexicon**.
    2. In the **New Lexicon** window, specify the name covering the Lexicon's general main topic and add a relevant description. It helps generate a more accurate result. To ensure that the generated content meets the desired expectations, fill in both fields. Relying solely on the title without using the description field will not produce the intended results.
    3. Go to the **Lexicon Entry Generation** setting, and activate **Generate Lexicon Entries**.
    4. Select **Lexicon language** from the list and set the number of entries (lexicon units).
    5. *(Optional)* Add the default [Slot](/ai/platform-features/nlu/slots/overview).
    6. *(Optional)* Activate **Generate Synonyms**. Synonyms help AI Agents understand and recognize different variations of the same concept. Up to five synonyms will be generated for each keyphrase.

    When the Lexicon Editor with new keyphrases is opened, you can edit, delete, or add new ones manually.

    <Frame>
      <img alt="Lexicon editor with AI-generated keyphrases" />
    </Frame>
  </Accordion>

  <Accordion title="Generate Flows">
    Note that the generation of Flows for primary NLU languages besides German and English is not fully supported.

    To use Generative AI technology for creating a new Flow with pre-configured Nodes based on your scenario, do the following:

    1. In the left-side menu of the Cognigy interface, click **Build > Flows**, then **+ New Flow**.
    2. In the **New Flow** window, go to the **Flow Generation** section and select one of the options:

       * **None** — the Generative AI will not be applied to this Flow. This setting is activated by default.
       * **Name and Description** — the Generative AI will use the Name and Description fields for generating Flow.
       * **Name and Transcript** — the Generative AI will use the Name and Transcript fields for generating Flow. For this setting, you need to create a scenario and put it in the Transcript field. Use the Transcript field template as an example for your scenario.

    In the existing Flow, you can edit Nodes created based on your scenario. Generate new Intent sentences or responses for a chatbot.

    <Frame>
      <img alt="Flow with Nodes generated by Generative AI from a scenario" />
    </Frame>
  </Accordion>

  <Accordion title="Generate Intent Sentences">
    <Note>
      * The generation of Intent example sentences for primary NLU languages besides German and English is not fully supported.
      * The model may not always generate the exact number of sentences requested, as its output can vary based on interpretation and model behavior.
    </Note>

    To use Generative AI technology for creating Intent example sentences, do the following:

    1. Open the existing Flow.
    2. In the upper-right corner of the **Flow Editor** page, select **NLU**.
    3. On the **Intent** tab, click **Create Intent**.
    4. Specify a unique name for the Intent and add a relevant description. It helps generate a more accurate result. To ensure that the generated content meets the desired expectations, fill in both fields. Relying solely on the title without using the description field will not produce the intended results.
    5. Activate the **Generate Example Sentences** setting and set the number of generative sentences.
    6. Generate new sentences by clicking **Create**, then click **Build Model**.

    If you want to add more examples automatically, use the **Generate Sentences** button. New sentences will be marked in light blue. The system generates specific numbers of sentences. Save changes and build your model again.

    You can also use Generative AI in the Node configuration.

    <Frame>
      <img alt="AI-generated example sentences for an Intent" />
    </Frame>
  </Accordion>

  <Accordion title="Generate Adaptive Cards">
    LLMs can create Adaptive Cards based on the prompt you provide in the description field for the **Adaptive Card** output type in the [Say](/ai/agents/develop/node-reference/basic/say) or [Copilot: Adaptive Card Tile](/ai/agents/develop/node-reference/agent-copilot/set-adaptive-card-tile) Nodes.
  </Accordion>
</AccordionGroup>

## Run-Time Generative AI Features

In Cognigy, the run-time Generative AI features enrich AI Agents with dynamic interactions using LLMs.
These features include running prompts, orchestrating conversations, rephrasing outputs,
dynamic question reprompts, and generative knowledge retrieval.

### Configure LLMs for Run-Time Features

1. In **Manage > Settings**, activate **Enable Generative AI Features** in the **Generative AI Settings** section.
2. Go to the [run-time feature](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) you want to configure and select the LLM you want to use.

### Use Run-Time Features

In the [Flow editor](/ai/agents/develop/projects-and-flows/editor), you can use the following run-time Generative AI features:

<AccordionGroup>
  <Accordion title="LLM Prompt Node">
    The [LLM Prompt Node](/ai/agents/develop/node-reference/service/llm-prompt) enables you to run a prompt against an LLM and either output the message or store it in the Input or Context objects.
  </Accordion>

  <Accordion title="Rephrasing AI Agent Outputs">
    To use AI-enhanced bot output rephrasing in [Say](/ai/agents/develop/node-reference/basic/say), [Question](/ai/agents/develop/node-reference/basic/question), and [Optional Question](/ai/agents/develop/node-reference/basic/optional-question) Nodes, do the following:

    1. Open the existing Flow.

    2. Add one of the Nodes: Say, Question, or Optional Question.

    3. Go to the **AI-enhanced output** section.

    4. In the **Rephrase Output** setting, select one of the options:

       * **None** — the Generative AI will not be applied to this Node. This setting is activated by default.
       * **Based on Custom Input** — specify custom values for the Input. Use the **Custom Inputs** field that allows the bot developer to input information for contextualizing and rephrasing the output.
       * **Based on previous user inputs** — set the last `x` user Inputs considered.

    5. Set the score in the Temperature setting. The temperature range determines the extent of variation in Generative AI's response.

    6. Click **Save Node**.

    Check in the Interaction Panel if your Flow works as expected.

    <Frame>
      <img alt="AI-enhanced output rephrasing settings in a Say Node" />
    </Frame>
  </Accordion>

  <Accordion title="LLM-powered Question Reprompts">
    The [Question Node](/ai/agents/develop/node-reference/basic/question) includes a feature to output a prompt to the user when they have answered a question incorrectly. Instead of using static text, you can use LLMs to generate a more dynamic and personalized output.
  </Accordion>

  <Accordion title="Search Extract Output Node">
    The [Search Extract Output Node](/ai/agents/develop/node-reference/other-nodes/search-extract-output) uses Cognigy [Knowledge AI](/ai/agents/develop/knowledge-ai/overview)
    to execute a search within a Knowledge Store,
    extracts a relevant answer via a Generative AI model, and creates an output.
  </Accordion>

  <Accordion title="LLM Entity Extract Node">
    The [LLM Entity Extract Node](/ai/agents/develop/node-reference/other-nodes/llm-entity-extract) retrieves specific entities from user inputs,
    such as product or booking codes and customer IDs.
    For example, if a user input is `I would like to book flight AB123 from New York to London`,
    the LLM can extract the booking code `AB123` from this input.
  </Accordion>
</AccordionGroup>

<Accordion title="FAQ">
  **Q1:** Can I use my OpenAI free account for the Generative AI feature in Cognigy?

  **A1:** A paid account is required to get an API Key, which is necessary for using Generative AI. A free account does not provide this key.

  **Q2**: Why doesn't Generative AI work with AudioCodes Nodes?

  **A2:** Generative AI output supports only text messages in the AI channel.
</Accordion>

## More Information

* [Flows](/ai/agents/develop/projects-and-flows/overview)
* [Flow Nodes](/ai/agents/develop/node-reference/overview)
* [Intents](/ai/platform-features/nlu/intents/overview)
* [Lexicons](/ai/platform-features/nlu/slots/user-defined/lexicon)
* [LLMs](/ai/agents/develop/gen-ai-and-llms/llms)
* [Cognigy Settings](/ai/administer/access/project-settings)


# Large Language Model Session Token Counter
Source: https://docs.cognigy.com/ai/agents/develop/gen-ai-and-llms/llm-session-token-counter



<a href="/release-notes/4.92"><Badge>Updated in 4.92</Badge></a>

In Cognigy, you can use the `api.getLLMTokenUsageForSession()` API call in a [Code Node](/ai/for-developers/code/overview) or an [Extension Node](/ai/for-developers/extensions) to fetch the Large Language Model (LLM) total token usage in a conversation or chat session. This way, you can track your [LLM providers](/ai/agents/develop/gen-ai-and-llms/providers/all-providers) costs.

The LLM Session Token Counter retrieves the token usage information via LLM provider-specific API calls to return the exact number of tokens the LLM uses. If the information isn't available by the provider, Cognigy estimates the token usage.

## Prerequisite

* You have added a Node that uses an LLM to your Flow, for example, an [AI Agent Node](/ai/agents/develop/node-reference/ai/ai-agent), [LLM Prompt Node](/ai/agents/develop/node-reference/service/llm-prompt), [Search Extract Output Node](/ai/agents/develop/node-reference/other-nodes/knowledge-search), or [Agent Copilot Nodes](/ai/agents/develop/node-reference/agent-copilot/overview).

## Limitations

* The LLM Session Token Counter doesn't support embedding models from Aleph Alpha.

## Fetching Session-Wide Token Usage

1. Add a [Code Node](/ai/for-developers/code/overview) after the Node that uses an LLM, for example, after an [AI Agent Node](/ai/agents/develop/node-reference/ai/ai-agent).
2. Add the following code to the Code Node:

   ```js theme={null}
   const tokens = api.getLLMTokenUsageForSession();
   ```

This Code Node fetches the total token usage in the session and stores it in the `token` variable.

<Accordion title="Returned JSON Object">
  The following JSON example shows an object returned by `api.getLLMTokenUsageForSession()`:

  ```json theme={null}
  {
      "1c945b38-5dbb-4fcf-9fdd-112bd8177f9c": {
          "IlmDisplayName": "openAI - text- embedding-3-large - 1735891122507",
          "providerType": "openAI",
          "modelType" : "text-embedding-3-large",
          "usage": {
              "inputTokens": 6,
              "outputTokens": 0
          }
      },
      "ce190e1f-eff7-4f6b-9074-9f705375a3d3": {
          "IlmDisplayName": "openAI - gpt-40 - 1735891179896",
          "providerType": "openAI",
          "modelType": "gpt-40",
          "usage": {
              "inputTokens": 441,
              "outputTokens": 106
          }
      }
  }
  ```

  | Key            | Type   | Description                                                                                                                                           | Example                                |
  | -------------- | ------ | ----------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------- |
  | llmReferenceId | String | The LLM reference ID.                                                                                                                                 | `ce190e1f-eff7-4f6b-9074-9f705375a3d3` |
  | llmDisplayName | String | The name of the LLM.                                                                                                                                  | `openAI - gpt-40 - 1735891179896`      |
  | providerType   | String | The LLM provider.                                                                                                                                     | `openAI`                               |
  | modelType      | String | The LLM model.                                                                                                                                        | `gpt-40`                               |
  | usage          | String | The object containing information about the session total token usage.                                                                                | -                                      |
  | inputTokens    | Number | The number of input tokens.                                                                                                                           | 441                                    |
  | outputTokens   | Number | The number output tokens. The output tokens count for embedding models is always 0 since embedding models output embedding vectors instead of tokens. | 106                                    |

  For testing in the Interaction Panel, you can output the JSON object by adding `api.say()` at the end of your [Code Node](/ai/for-developers/code/api-functions).
</Accordion>

## More Information

* [Overview](/ai/agents/develop/gen-ai-and-llms/llms)
* [LLM Providers](/ai/agents/develop/gen-ai-and-llms/providers/all-providers)
* [Model Support by Feature](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature)


# Large Language Models (LLMs)
Source: https://docs.cognigy.com/ai/agents/develop/gen-ai-and-llms/llms

Integrate LLMs from OpenAI, Azure, Anthropic, and more to enhance AI Agent conversations in Cognigy.

<a href="/release-notes/2025.16"><Badge>Updated in 2025.16</Badge></a>

[Large Language Models (LLMs)](https://en.wikipedia.org/wiki/Large_language_model) are specific types of AI models that are designed for generating humanlike text based on the input and context provided. These models are trained on vast amounts of text data, allowing them to learn patterns, syntax, and semantic relationships between words and phrases.

With LLMs, Cognigy AI Agents can understand and respond to user input in a natural way. These models make conversations more engaging by generating relevant and contextually appropriate responses. LLMs also help manage dialogues and provide multilingual support, enhancing the overall conversational experience for users.

In the Cognigy interface, you can connect to an LLM provider of your choice via API.
Cognigy supports LLM integration with top providers such as OpenAI, Azure, Anthropic, Amazon Bedrock, Google, and Mistral AI.

To integrate an LLM into your conversational design process within Cognigy, follow these steps:

1. [Determine a use case that aligns with your business objectives](#use-cases).
2. [Select a standard or custom model](#model-types) from a preferred LLM provider. Ensure that you have all the necessary API credentials to access the specific LLM you intend to use.
3. Make sure to familiarize yourself with [Cognigy's retry mechanism](#retry-mechanism) in case the model does not respond.
4. *(beta)* Implement a [fallback strategy](#llm-fallback-beta) to enhance the reliability of the model's connection.

## Use Cases

Explore key use cases that illustrate how LLMs can be utilized in Cognigy, and learn about the features necessary to implement these use cases:

<AccordionGroup>
  <Accordion title="Automate Routine Tasks">
    In the Cognigy interface, LLMs help automate repetitive tasks that previously required manual effort from conversation designers:

    * **Generate Flows**. LLMs can generate Flows based on the provided title and description, helping save time during conversation design iterations. To generate Flows, refer to the [Generate Flows](/ai/agents/develop/gen-ai-and-llms/generative-ai) section.
    * **Generate Lexicons**. With LLMs, you can build a list of Keyphrases for the language of your choice based on the title and description of the provided Lexicon. To generate Lexicons, refer to the [Generate Lexicons](/ai/agents/develop/gen-ai-and-llms/generative-ai) section.
    * **Generate Intent sentences**. With LLMs, you can build a list of Intent sentences based on the title and description of the provided Intent. To generate Intents, refer to the [Generate Intent Sentences](/ai/agents/develop/gen-ai-and-llms/generative-ai) section.
    * **Generate adaptive cards**. LLMs can create Adaptive Cards based on the prompt you provide in the description field for the **Adaptive Card** output type in the [Say](/ai/agents/develop/node-reference/basic/say) or [Copilot: Adaptive Card Tile](/ai/agents/develop/node-reference/agent-copilot/set-adaptive-card-tile) Nodes.
  </Accordion>

  <Accordion title="Improve AI Agent's Output">
    The AI Agent's output often requires refinement to meet the high standards customers expect.
    In Cognigy, LLMs can enhance the performance of AI Agents in various ways:

    * **Generate Enhanced Output**. By using LLMs, AI Agents can produce responses that are more comprehensive, nuanced, and in line with customer needs. This enhancement is important in handling queries that require a level of empathy or complex information processing. To generate enhanced output, add [LLM Prompt Nodes](/ai/agents/develop/node-reference/service/llm-prompt) within your Flow.
    * **Rephrase AI Agent's Output**. LLMs are good at rephrasing texts while maintaining the original meaning. This capability allows for smoother communication if the original output of an AI Agent lacks clarity or tone alignment. To use this capability, enable AI-enhanced rephrasing in [Say](/ai/agents/develop/node-reference/basic/say#ai-enhanced-output), [Question](/ai/agents/develop/node-reference/basic/question), or [Optional Question](/ai/agents/develop/node-reference/basic/optional-question) Nodes by selecting the **Rephrase Output** option in the **AI-enhanced output** section. Also, you can generate dynamic reprompts for incorrect answers via [Question Nodes](/ai/agents/develop/node-reference/basic/question).
    * **Train Intents and Find the Right Intent**. To enhance AI Agents' understanding of user inputs in various languages, use an [external embedding LLM](/ai/platform-features/nlu/external/external-nlu-intent-recognition). Also, optimize intent classification accuracy by using [LLM-based intent reranking](/ai/platform-features/nlu/external/intent-reranking), which prioritizes and selects the most relevant intent.
  </Accordion>

  <Accordion title="Use Retrieval Augmented Generation (RAG) with Knowledge AI">
    [RAG (Retrieval-Augmented Generation)](https://en.wikipedia.org/wiki/Retrieval-augmented_generation) is an AI framework that combines data from information retrieval systems with LLMs. This AI framework makes AI-generated content more accurate and relevant by focusing on specific, up-to-date information.

    The [Cognigy Knowledge AI](/ai/agents/develop/knowledge-ai/overview) feature is built on the RAG framework, allowing it to retrieve information from various knowledge bases, including PDFs, Word documents, PowerPoint presentations, web pages, and images. Knowledge AI assists both AI and human agents in providing better responses:

    * **Use Knowledge AI for AI Agents**. Using LLMs integrated into your Project, AI Agents can access relevant knowledge sources, including articles from subject-matter experts, frequently asked questions, and diagrams. To create a first Knowledge Store, go to the [Knowledge AI](/ai/agents/develop/knowledge-ai/overview) documentation. Then, retrieve data from the Knowledge Store and improve the AI Agent's output by using the [Search Extract Output](/ai/agents/develop/node-reference/other-nodes/search-extract-output) Node.
    * **Use Knowledge AI for human agents**. By retrieving data from different knowledge bases, LLMs reduce the burden on human agents, making information access quicker and more efficient. To let human agents use the knowledge base for solving tasks, add a widget to the Agent Copilot workspace using a [Copilot: Knowledge Tile](/ai/agents/develop/node-reference/agent-copilot/knowledge-tile) Node. The widget serves as a search tool, enabling human agents to enter questions and retrieve information from the knowledge base.
  </Accordion>

  <Accordion title="Extract Specific Entities from User Inputs">
    By using LLMs, you can extract specific entities from user inputs, such as product or booking codes and customer IDs.
    For example, if a user input is `I would like to book flight AB123 from New York to London`,
    the LLM can extract the booking code `AB123` from this input.
    This capability allows for dynamic, context-aware responses while improving accuracy and reducing manual data entry.
    To add this feature to your AI Agent,
    add the [LLM Entity Extract](/ai/agents/develop/node-reference/other-nodes/llm-entity-extract) Node within your Flow.

    <Frame>
      <img alt="LLM Entity Extract Node tested in the Interaction Panel" />
    </Frame>
  </Accordion>

  <Accordion title="Determine the Emotional Tone of the Conversation">
    Understanding the emotional tone of customer interactions is crucial. LLMs can analyze conversations in real time to determine customer sentiment.
    With Agent Copilot sentiment analysis, LLMs can give insights, enabling human agents to adjust responses for a more effective interaction and improved service quality.
    To let human agents use sentiment analysis provided by LLM,
    add a widget to the Agent Copilot workspace
    using a [Copilot: Sentiment Analysis Tile](/ai/agents/develop/node-reference/agent-copilot/sentiment-tile) Node. The widget monitors the conversation and displays the emotional tone (positive, negative, or neutral) of the latest user input in real time.

    <Frame>
      <img alt="Sentiment Analysis Tile widget in the Agent Copilot workspace" />
    </Frame>
  </Accordion>

  <Accordion title="Ensure LLM Accuracy with Human Review">
    Even though LLMs are powerful, there are situations where human oversight is essential.
    Errors in data retrieval or sentiment analysis can occur, especially with nuanced or complex information.
    Human review can help minimize these errors.

    Agent Copilot provides [plug-and-play widgets](/agent-copilot/configure/nodes) in the [workspace](/agent-copilot/index), allowing human agents to review the output generated by LLMs before sending responses to customers.
    This human-in-the-loop approach ensures that wrong or misleading information doesn't reach customers.
  </Accordion>
</AccordionGroup>

## Model Types

In Cognigy, you can add an LLM from the offered LLM providers as either a standard model or a custom model.

### Standard Models

*Standard models* are predefined models in the Cognigy interface.
To use a standard model, select an LLM provider and model name from the list of [predefined models](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature).
By choosing a standard model, you can ensure that this model is fully compatible with Cognigy features.

After selecting a model, you must enter the API credentials from the [LLM provider](/ai/agents/develop/gen-ai-and-llms/providers/all-providers) that offers this model to get access to its functionality.

#### Deprecations and Removals

Standard models that have become outdated and are soon to be removed by the LLM provider are marked with special icons in the Cognigy interface. These icons help you stay informed about changes to the model availability and ensure your AI Agents continue to run smoothly.

| **Icon**                 | **Meaning**                                                                                                                                                                                                                                                                                                                                                                     | **Recommended Next Step**                                                                                       |
| ------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------- |
| <img alt="deprecated" /> | The LLM provider has stopped developing the model and plans to remove it in the future. Deprecation statuses may vary by provider, for example, Anthropic and OpenAI have `legacy` and `deprecated` statuses, while Microsoft Azure uses only `deprecated`. Cognigy marks a model as `deprecated` after the LLM provider announces plans to stop providing support and updates. | Select an alternative model by consulting the LLM provider's documentation and switching your AI Agent to it.   |
| <img alt="removed" />    | The model is no longer available and can't be used. If your AI Agents use this model, they might stop working.                                                                                                                                                                                                                                                                  | Immediately switch to an available model with similar capabilities to prevent AI Agent interruptions or errors. |

### Custom Models

*Custom models* are non-predefined models in the Cognigy interface. To use a custom model, select an LLM provider and manually enter the model name instead of selecting from a list. While these models may not offer the same level of Cognigy feature support as standard models, they can be helpful when you need a specific version and are familiar with its capabilities. Moreover, if LLM providers release new model versions, opting for a custom model lets you avoid waiting for Cognigy to update its list of standard models.

As with the standard model, you must enter the API credentials for your custom model from the [LLM provider](/ai/agents/develop/gen-ai-and-llms/providers/all-providers).

## Retry Mechanism

If Cognigy encounters issues while trying to connect with LLM providers that don't return a `200 OK` response code,
Cognigy will automatically attempt to reconnect up to three times using a retry mechanism.
This retry mechanism is designed to improve the reliability and stability of the system when using certain Cognigy Nodes and features.
By attempting to reconnect multiple times, the retry mechanism increases the chances of successful communication, reduces disruptions and ensures smoother operations within the system.

The following Nodes and features are affected by the retry mechanism:

* [LLM Prompt](/ai/agents/develop/node-reference/service/llm-prompt)
* [LLM Entity Extract](/ai/agents/develop/node-reference/other-nodes/llm-entity-extract)
* [Search Extract Output](/ai/agents/develop/node-reference/other-nodes/search-extract-output)
* Question Node Slot Mapping (in case of using [external NLU embedding](/ai/platform-features/nlu/external/external-nlu-intent-recognition))
* [Question Node LLM Reprompting](/ai/agents/develop/node-reference/basic/question#parameters)
* [Sentiment Analysis](/ai/agents/develop/node-reference/agent-copilot/sentiment-tile)

<Accordion title="Retry Mechanism for On-Premises">
  The retry mechanism can be customized using environment variables for on-premises installations.
  By adjusting these environment variables, system administrators can fine-tune the retry behavior
  to suit specific performance requirements and network conditions.

  | Environment Variable                                | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    | Default Value |
  | --------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------- |
  | GENERATIVE\_AI\_RETRY\_OPTIONS\_NUMBER\_OF\_RETRIES | Determines the number of retries after the first attempt before failing.                                                                                                                                                                                                                                                                                                                                                                                                                                                       | 2             |
  | GENERATIVE\_AI\_RETRY\_OPTIONS\_MIN\_TIMEOUT        | Determines the timeout between retries to avoid rate limiting.                                                                                                                                                                                                                                                                                                                                                                                                                                                                 | 50ms          |
  | GENERATIVE\_AI\_RETRY\_OPTIONS\_MAX\_TIMEOUT        | Determines the maximum timeout between retries.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                | 50ms          |
  | GENERATIVE\_AI\_RETRY\_OPTIONS\_FACTOR              | Determines how long a retry timeout will last. The timeout duration increases with each subsequent retry. If the factor is greater than 1, the timeout duration will increase exponentially. This means that with each subsequent attempt, the waiting time will not just increase by a constant amount, but rather by multiples. For example, if the initial timeout is 1 second and the factor is set to 2, then the second timeout will be 2 seconds, the third will be 4 seconds, the fourth will be 8 seconds, and so on. | 1             |
</Accordion>

## LLM Fallback

In addition to the [default retry mechanism](#retry-mechanism) that applies to all models, you can configure an [LLM fallback](/ai/agents/develop/gen-ai-and-llms/fallback) for a specific model.
Create a fallback model to ensure your AI Agents continue working if the main model fails.
Set a threshold for failed requests to trigger the fallback, specify its active duration, and enable notifications.
Note that creating fallbacks for embedding models is not supported.

## More Information

* [Supported LLM Providers](/ai/agents/develop/gen-ai-and-llms/providers/all-providers)
* [Model Support by Feature](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature)
* [Other LLM Operations](/ai/agents/develop/gen-ai-and-llms/other-operations)
* [Generative AI](/ai/agents/develop/gen-ai-and-llms/generative-ai)
* [Knowledge AI](/ai/agents/develop/knowledge-ai/overview)


# Model Support by Feature
Source: https://docs.cognigy.com/ai/agents/develop/gen-ai-and-llms/model-support-by-feature

Explore LLM providers and models supported by Cognigy, including feature compatibility, standard and custom model options.

<a href="/release-notes/2026.18"><Badge>Updated in 2026.18</Badge></a>

The table provides an overview of the LLM providers supported by Cognigy, detailing the standard models and their compatibility with Cognigy features.
You can also add a custom model that isn't listed in the table.
To do so, select the **Custom Model** parameter when creating a model and specify both the model type and name.
For more information about adding a custom or standard model for a selected LLM provider,
refer to the [All LLM Providers](/ai/agents/develop/gen-ai-and-llms/providers/all-providers) page.

<div>
  | <div>Models</div>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        | Design-Time Features<sup><a href="#footnote1">1</a></sup> |                       AI Agent Node                      | Conversation Analyzer<sup><a href="#footnote8">8</a></sup> |     AI Enhanced Outputs     |       LLM Prompt Node       |      Answer Extraction      |       Knowledge Search      |      Sentiment Analysis     |     NLU Embedding Model     |
  | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | :-------------------------------------------------------: | :------------------------------------------------------: | :--------------------------------------------------------: | :-------------------------: | :-------------------------: | :-------------------------: | :-------------------------: | :-------------------------: | :-------------------------: |
  | <img alt="Cognigy" /> **[Cognigy](/ai/agents/develop/gen-ai-and-llms/generative-ai#design-time-generative-ai-features)**                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |                                                           |                                                          |                                                            |                             |                             |                             |                             |                             |                             |
  | Platform-Provided LLM                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |                  <img alt="supported" />                  |                <img alt="not supported" />               |                 <img alt="not supported" />                | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> |
  | <img alt="Microsoft Azure OpenAI" /> <strong><a href="/ai/agents/develop/gen-ai-and-llms/providers/microsoft-azure-openai">Microsoft Azure OpenAI</a></strong>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |                                                           |                                                          |                                                            |                             |                             |                             |                             |                             |                             |
  | `gpt-5-6-sol`<sup><a href="#footnote7">7</a></sup><br />`gpt-5-6-terra`<sup><a href="#footnote7">7</a></sup><br />`gpt-5-6-luna`<sup><a href="#footnote7">7</a></sup>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |                  <img alt="supported" />                  |                  <img alt="supported" />                 |                 <img alt="not supported" />                |   <img alt="supported" />   |   <img alt="supported" />   |   <img alt="supported" />   | <img alt="not supported" /> |   <img alt="supported" />   | <img alt="not supported" /> |
  | `gpt-5-5`<sup><a href="#footnote2">2</a>,<a href="#footnote7">7</a></sup><br />`gpt-5-4`<sup><a href="#footnote2">2</a></sup><br />`gpt-5-4-mini`<sup><a href="#footnote2">2</a></sup><br />`gpt-5-4-nano`<sup><a href="#footnote2">2</a></sup><br />`gpt-5-2`<sup><a href="#footnote2">2</a></sup><br />`gpt-5-1`<sup><a href="#footnote2">2</a></sup><br />`gpt-5`<sup><a href="#footnote2">2</a></sup><br />`gpt-5-nano`<sup><a href="#footnote2">2</a></sup> ([deprecated](https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/retired-models))<br />`gpt-5-mini`<sup><a href="#footnote2">2</a></sup> ([deprecated](https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/retired-models))<br />`gpt-4.1-nano` ([deprecated](https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/retired-models))<br />`gpt-4.1-mini` ([deprecated](https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/retired-models))<br />`gpt-4.1` ([deprecated](https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/retired-models))<br />`gpt-4o-mini` ([deprecated](https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/retired-models))<br />`gpt-4o` ([deprecated](https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/retired-models)) |                  <img alt="supported" />                  |                  <img alt="supported" />                 |                   <img alt="supported" />                  |   <img alt="supported" />   |   <img alt="supported" />   |   <img alt="supported" />   | <img alt="not supported" /> |   <img alt="supported" />   | <img alt="not supported" /> |
  | `text-embedding-ada-002`<br />`text-embedding-3-small`<sup><a href="#footnote3">3</a></sup><br />`text-embedding-3-large`<sup><a href="#footnote3">3</a></sup>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |                <img alt="not supported" />                |                <img alt="not supported" />               |                 <img alt="not supported" />                | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> |   <img alt="supported" />   | <img alt="not supported" /> |   <img alt="supported" />   |
  | <img alt="OpenAI" /> <strong><a href="/ai/agents/develop/gen-ai-and-llms/providers/openai">OpenAI</a></strong>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |                                                           |                                                          |                                                            |                             |                             |                             |                             |                             |                             |
  | `gpt-5-6-sol`<sup><a href="#footnote7">7</a></sup><br />`gpt-5-6-terra`<sup><a href="#footnote7">7</a></sup><br />`gpt-5-6-luna`<sup><a href="#footnote7">7</a></sup>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |                  <img alt="supported" />                  |                  <img alt="supported" />                 |                 <img alt="not supported" />                |   <img alt="supported" />   |   <img alt="supported" />   |   <img alt="supported" />   | <img alt="not supported" /> |   <img alt="supported" />   | <img alt="not supported" /> |
  | `gpt-5-5`<sup><a href="#footnote2">2</a>,<a href="#footnote7">7</a></sup><br />`gpt-5-4`<sup><a href="#footnote2">2</a></sup><br />`gpt-5-4-mini`<sup><a href="#footnote2">2</a></sup><br />`gpt-5-4-nano`<sup><a href="#footnote2">2</a></sup><br />`gpt-5-2`<sup><a href="#footnote2">2</a></sup><br />`gpt-5-1`<sup><a href="#footnote2">2</a></sup><br />`gpt-5`<sup><a href="#footnote2">2</a></sup><br />`gpt-5-nano`<sup><a href="#footnote2">2</a></sup><br />`gpt-5-mini`<sup><a href="#footnote2">2</a></sup><br />`gpt-4.1-nano` ([deprecated](https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/retired-models))<br />`gpt-4.1-mini`<br />`gpt-4.1`<br />`gpt-4o-mini`<br />`gpt-4o`                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |                  <img alt="supported" />                  |                  <img alt="supported" />                 |                   <img alt="supported" />                  |   <img alt="supported" />   |   <img alt="supported" />   |   <img alt="supported" />   | <img alt="not supported" /> |   <img alt="supported" />   | <img alt="not supported" /> |
  | `text-embedding-ada-002`<br />`text-embedding-3-small`<sup><a href="#footnote3">3</a></sup><br />`text-embedding-3-large`<sup><a href="#footnote3">3</a></sup>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |                <img alt="not supported" />                |                <img alt="not supported" />               |                 <img alt="not supported" />                | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> |   <img alt="supported" />   | <img alt="not supported" /> |   <img alt="supported" />   |
  | <img alt="OpenAI-Compatible" /><strong><a href="/ai/agents/develop/gen-ai-and-llms/providers/openai-compatible">OpenAI-Compatible</a></strong>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |                                                           |                                                          |                                                            |                             |                             |                             |                             |                             |                             |
  | OpenAI-compatible LLMs                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |                <img alt="not supported" />                |                  <img alt="supported" />                 |                   <img alt="supported" />                  | <img alt="not supported" /> |   <img alt="supported" />   |   <img alt="supported" />   |   <img alt="supported" />   | <img alt="not supported" /> | <img alt="not supported" /> |
  | <img alt="Anthropic" /> <strong><a href="/ai/agents/develop/gen-ai-and-llms/providers/anthropic">Anthropic</a></strong>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |                                                           |                                                          |                                                            |                             |                             |                             |                             |                             |                             |
  | `claude-sonnet-5`<sup><a href="#footnote2">2</a>,<a href="#footnote9">9</a></sup><br />`claude-opus-4-7`<sup><a href="#footnote2">2</a>,<a href="#footnote9">9</a></sup><br />`claude-opus-4-8`<sup><a href="#footnote2">2</a>,<a href="#footnote9">9</a></sup>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |                  <img alt="supported" />                  |                  <img alt="supported" />                 |                   <img alt="supported" />                  |   <img alt="supported" />   |   <img alt="supported" />   |   <img alt="supported" />   | <img alt="not supported" /> |   <img alt="supported" />   | <img alt="not supported" /> |
  | `claude-sonnet-4-6`<sup><a href="#footnote2">2</a></sup>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |                  <img alt="supported" />                  |                  <img alt="supported" />                 |                 <img alt="not supported" />                |   <img alt="supported" />   |   <img alt="supported" />   |   <img alt="supported" />   | <img alt="not supported" /> |   <img alt="supported" />   | <img alt="not supported" /> |
  | `claude-sonnet-4-5`<sup><a href="#footnote2">2</a></sup>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |                  <img alt="supported" />                  |                  <img alt="supported" />                 |                   <img alt="supported" />                  |   <img alt="supported" />   |   <img alt="supported" />   |   <img alt="supported" />   | <img alt="not supported" /> |   <img alt="supported" />   | <img alt="not supported" /> |
  | `claude-haiku-4-5`<sup><a href="#footnote2">2</a></sup>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |                  <img alt="supported" />                  |                  <img alt="supported" />                 |                   <img alt="supported" />                  |   <img alt="supported" />   |   <img alt="supported" />   |   <img alt="supported" />   | <img alt="not supported" /> |   <img alt="supported" />   | <img alt="not supported" /> |
  | <img alt="Google Gemini" /> <strong><a href="/ai/agents/develop/gen-ai-and-llms/providers/google-gemini">Google Gemini</a></strong>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |                                                           |                                                          |                                                            |                             |                             |                             |                             |                             |                             |
  | `gemini-2.5-pro` ([deprecated](https://cloud.google.com/vertex-ai/generative-ai/docs/learn/model-versions#legacy-stable))<sup><a href="#footnote2">2</a></sup><br />`gemini-2.5-flash`<sup><a href="#footnote2">2</a></sup><br />`gemini-2.5-flash-lite`<sup><a href="#footnote2">2</a>,<a href="#footnote10">10</a></sup> ([deprecated](https://cloud.google.com/vertex-ai/generative-ai/docs/learn/model-versions#legacy-stable))<br />`gemini-2.0-flash` ([deprecated](https://cloud.google.com/vertex-ai/generative-ai/docs/learn/model-versions#legacy-stable))                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |                <img alt="not supported" />                |                  <img alt="supported" />                 |                   <img alt="supported" />                  | <img alt="not supported" /> |   <img alt="supported" />   |   <img alt="supported" />   | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> |
  | `gemini-embedding-001`                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |                <img alt="not supported" />                |                <img alt="not supported" />               |                 <img alt="not supported" />                | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> |   <img alt="supported" />   | <img alt="not supported" /> | <img alt="not supported" /> |
  | <img alt="Google GenAI" /> <strong><a href="/ai/agents/develop/gen-ai-and-llms/providers/google-genai">Google GenAI</a></strong>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         |                                                           |                                                          |                                                            |                             |                             |                             |                             |                             |                             |
  | `gemini-3.1-pro-preview`<sup><a href="#footnote2">2</a></sup><br />`gemini-3-flash-preview`<sup><a href="#footnote2">2</a></sup><br />`gemini-3.5-flash`<sup><a href="#footnote2">2</a></sup><br />`gemini-3.6-flash`<sup><a href="#footnote2">2</a></sup><br />`gemini-3.7-flash`<sup><a href="#footnote2">2</a></sup><br />`gemini-3.1-flash-lite`<sup><a href="#footnote2">2</a></sup>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |                <img alt="not supported" />                |                  <img alt="supported" />                 |                 <img alt="not supported" />                | <img alt="not supported" /> |   <img alt="supported" />   |   <img alt="supported" />   | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> |
  | `gemini-3.5-flash-lite`<sup><a href="#footnote2">2</a>,<a href="#footnote10">10</a></sup>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |                <img alt="not supported" />                |                  <img alt="supported" />                 |                   <img alt="supported" />                  | <img alt="not supported" /> |   <img alt="supported" />   |   <img alt="supported" />   | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> |
  | `gemini-embedding-2`                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |                <img alt="not supported" />                |                <img alt="not supported" />               |                 <img alt="not supported" />                | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> |   <img alt="supported" />   | <img alt="not supported" /> | <img alt="not supported" /> |
  | <img alt="Aleph Alpha" /> **[Aleph Alpha](/ai/agents/develop/gen-ai-and-llms/providers/aleph-alpha)**                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |                                                           |                                                          |                                                            |                             |                             |                             |                             |                             |                             |
  | `luminous` ([deprecated](https://docs.aleph-alpha.com))                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |                <img alt="not supported" />                |                <img alt="not supported" />               |                 <img alt="not supported" />                | <img alt="not supported" /> |   <img alt="supported" />   |   <img alt="supported" />   | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> |
  | <img alt="Amazon Bedrock" /> **[Amazon Bedrock](/ai/agents/develop/gen-ai-and-llms/providers/amazon-bedrock)**                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |                                                           |                                                          |                                                            |                             |                             |                             |                             |                             |                             |
  | `anthropic.claude-3-5-sonnet-20240620-v1:0` ([deprecated](https://docs.aws.amazon.com/bedrock/latest/userguide/model-lifecycle.html))<br />`amazon.nova-lite-v1:0`<br />`amazon.nova-pro-v1:0`<br />`amazon.nova-premier-v1:0`<sup><a href="#footnote2">2</a></sup> ([deprecated](https://docs.aws.amazon.com/bedrock/latest/userguide/model-lifecycle.html))<br />`amazon.nova-2-lite-v1:0`                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |                <img alt="not supported" />                |                  <img alt="supported" />                 |                   <img alt="supported" />                  | <img alt="not supported" /> |   <img alt="supported" />   |   <img alt="supported" />   | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> |
  | `amazon.titan-embed-text-v2:0`<sup><a href="#footnote5">5</a></sup>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |                <img alt="not supported" />                |                <img alt="not supported" />               |                 <img alt="not supported" />                | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> |   <img alt="supported" />   | <img alt="not supported" /> | <img alt="not supported" /> |
  | `amazon.nova-micro-v1:0`                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |                <img alt="not supported" />                |                <img alt="not supported" />               |                 <img alt="not supported" />                | <img alt="not supported" /> |   <img alt="supported" />   |   <img alt="supported" />   | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> |
  | [Converse API](https://docs.aws.amazon.com/bedrock/latest/userguide/models-features.html)-compatible models                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |                <img alt="not supported" />                | Partially supported<sup><a href="#footnote6">6</a></sup> |                   <img alt="supported" />                  | <img alt="not supported" /> |   <img alt="supported" />   |   <img alt="supported" />   | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> |
  | <img alt="Mistral AI" /> **[Mistral AI](/ai/agents/develop/gen-ai-and-llms/providers/mistral)**                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |                                                           |                                                          |                                                            |                             |                             |                             |                             |                             |                             |
  | `mistral-large-latest`<sup><a href="#footnote4">4</a></sup><br />`mistral-medium-latest` ([deprecated](https://docs.mistral.ai/models/overview#legacy-models))<sup><a href="#footnote4">4</a></sup><br />`mistral-small-latest`<sup><a href="#footnote4">4</a></sup>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |                <img alt="not supported" />                |                  <img alt="supported" />                 |                   <img alt="supported" />                  | <img alt="not supported" /> |   <img alt="supported" />   |   <img alt="supported" />   | <img alt="not supported" /> | <img alt="not supported" /> | <img alt="not supported" /> |
</div>

## More Information

* [Overview](/ai/agents/develop/gen-ai-and-llms/llms)
* [All LLM Providers](/ai/agents/develop/gen-ai-and-llms/providers/all-providers)
* [Other LLM Operations](/ai/agents/develop/gen-ai-and-llms/other-operations)

***

<span>
  <sup>1</sup> [Design-time
  features](/ai/agents/develop/gen-ai-and-llms/generative-ai#design-time-generative-ai-features)
  include Intent Sentence Generation, Flow Generation, Adaptive Card Generation,
  Lexicon Generation.
</span>

<br />

<span>
  <sup>2</sup> Reasoning models consume more tokens and may incur
  higher costs. The models are optimized for tasks that require complex problem-solving
  and logical reasoning. Before using these models in production, test token consumption in [debug mode](/ai/agents/test/interaction-panel/chat#debug-mode) and use them with caution. To reduce costs, consider using a non-reasoning model. For more information about reasoning models, refer to the [Microsoft Azure OpenAI](https://learn.microsoft.com/en-us/azure/ai-foundry/openai/how-to/reasoning), [OpenAI](https://platform.openai.com/docs/guides/reasoning/advice-on-prompting), and [Google](https://cloud.google.com/vertex-ai/generative-ai/docs/models) documentation.

  <br />

  Anthropic's `claude-sonnet-4-6` model doesn't support [assistant message prefilling](https://platform.claude.com/docs/en/about-claude/models/migration-guide#breaking-changes-3).
  To keep your Flows compatible, ensure that the last transcript step before the LLM call is a user message. For example, add an [Add Transcript Step Node](/ai/agents/develop/node-reference/service/add-transcript-step) before the LLM call and set the role to `user`.
</span>

<br />

<span>
  <sup>3</sup> For Knowledge AI, we recommend using
  `text-embedding-ada-002`. However, if you want to use `text-embedding-3-small`
  and `text-embedding-3-large`, make sure that you familiarize yourself with the
  restrictions of these models in [Which Model to
  Choose?](/ai/agents/develop/knowledge-ai/overview).
</span>

<br />

<span>
  <sup>4</sup> The `*-latest` suffix indicates that the model you
  select in Cognigy points to the latest version of the model. For more
  information, read
  [Anthropic's](https://docs.anthropic.com/en/docs/about-claude/models/overview)
  or [Mistral
  AI's](https://docs.mistral.ai/getting-started/models/models_overview/) models
  documentation.
</span>

<br />

<span>
  <sup>5</sup> For Cognigy 2025.10 and earlier versions, the
  option to select this model is hidden behind the
  `FEATURE_ENABLE_AWS_BEDROCK_EMBEDDING_LLM_WHITELIST` feature flag.
</span>

<br />

<span>
  <sup>6</sup> Note that some models from the [Converse
  API](https://docs.aws.amazon.com/bedrock/latest/userguide/models-features.html)
  might not support the AI Agent Node feature.
</span>

<br />

<span>
  <sup>7</sup> This model supports only the Responses API. When you select this model, the Responses API is selected by default, and the **API Type** field is locked.
</span>

<span>
  <sup>8</sup> The Conversation Analyzer uses the selected model to analyze session transcripts in [Cognigy Insights](/insights/overview). If no model is explicitly selected in **Manage > Settings > Generative AI Settings**, the Project's default LLM is used. Any chat model that supports the AI Agent Node also supports the Conversation Analyzer.
</span>

<span>
  <sup>9</sup> These models don't support token-sampling parameters such as `temperature`, `top_p`, and `top_k`. If you configure these parameters at the Node level or in **Custom Model Options**, they are automatically removed before the request is sent. To control model behavior, use the available thinking and budget settings instead.
</span>

<span>
  <sup>10</sup> The `gemini-2.5-flash-lite` and `gemini-3.5-flash-lite` models can return unstable results in the [LLM Prompt](/ai/agents/develop/node-reference/service/llm-prompt) Node. Test your Flows with these models before you use them in production.
</span>


# Other LLM Operations
Source: https://docs.cognigy.com/ai/agents/develop/gen-ai-and-llms/other-operations



Once you've added an LLM to Cognigy, you can perform the following operations with your model:

<AccordionGroup>
  <Accordion title="Clone a Model">
    <Tabs>
      <Tab title="GUI">
        To create a copy of the existing model, follow these steps:

        1. Go to **Build > LLM**.
        2. Hover over the existing model and click <img alt="menu" />.
        3. Select **Clone** from the list.

        The model will contain the same settings as the initial one.
      </Tab>

      <Tab title="API">
        To create a copy of the existing model, use the [Cognigy API POST /v2.0/largelanguagemodels//clone](https://api-trial.cognigy.ai/openapi#post-/v2.0/largelanguagemodels/-largeLanguageModelId-/clone) request.
      </Tab>
    </Tabs>
  </Accordion>

  <Accordion title="Set a Model as Default">
    Setting a default model ensures a smooth transition when a specific model is removed. It guarantees that there is always a model available to handle compatible use cases, even if the assigned model is removed.

    <Tabs>
      <Tab title="GUI">
        To set a model as the default, follow these steps:

        1. Go to **Build > LLM**.
        2. Hover over the existing model and click <img alt="menu" />.
        3. Select **Make Default** from the list.

        The setting will be applied for the selected model.
      </Tab>

      <Tab title="API">
        To set a model as the default, use the `"isDefault": true` parameter in the [Cognigy API PATCH /v2.0/largelanguagemodels/](https://api-trial.cognigy.ai/openapi#patch-/v2.0/largelanguagemodels/-largeLanguageModelId-) request.
      </Tab>
    </Tabs>
  </Accordion>

  <Accordion title="Export a Model as a Package">
    To reuse a model in other Projects, package the model.

    To package a model, follow these steps:

    1. Go to **Build > LLM**.
    2. Hover over the existing model and click <img alt="menu" />.
    3. Select **Create a package**.
    4. Once the package has created, a new [task](/ai/agents/develop/projects-and-flows/projects), titled **Create package**, will be run. To view the task, click <img alt="task-menu" /> in the upper-right corner.

    When the task is completed, you can download the Package in the Task Panel.
    To ensure the model works correctly in a different Project, export its [connection](/ai/platform-features/connections) as a Package.
  </Accordion>

  <Accordion title="Delete a Model">
    Note that a default model cannot be deleted. Before deletion, you need to remove the default tag.

    <Tabs>
      <Tab title="GUI">
        To delete a model, follow these steps:

        1. Go to **Build > LLM**.
        2. Hover over the existing model and click <img alt="menu" />.
        3. Select **Delete**.
        4. Confirm the deletion. Features relying on this model will stop working if no default model is configured to support those features.

        The model will be deleted from the list.
      </Tab>

      <Tab title="API">
        To delete a model, use the [Cognigy API DELETE /v2.0/largelanguagemodels/](https://api-trial.cognigy.ai/openapi#delete-/v2.0/largelanguagemodels/-largeLanguageModelId-).
      </Tab>
    </Tabs>
  </Accordion>
</AccordionGroup>

## More Information

* [Overview](/ai/agents/develop/gen-ai-and-llms/llms)
* [LLM Providers](/ai/agents/develop/gen-ai-and-llms/providers/all-providers)
* [Model Support by Feature](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature)


# Aleph Alpha
Source: https://docs.cognigy.com/ai/agents/develop/gen-ai-and-llms/providers/aleph-alpha



<a href="/release-notes/2026.11"><Badge>Updated in 2026.11</Badge></a>

<Warning>
  As of Cognigy 2026.6.0, the support for the Completions API for custom LLMs is deprecated. The full removal is planned for Cognigy 2026.24.0.&#x20;
</Warning>

To start using an Aleph Alpha model with Cognigy features, follow these steps:

1. [Add a Model](#add-models)
2. [Apply the Model](#apply-the-model)

## Add Models

You can add a model using one of the following interfaces:

* [GUI](#add-models-via-gui)
* [API](#add-models-via-the-api)

### Add Models via GUI

You can add a model provided by Aleph Alpha to Cognigy in **Build > LLM**. To add the model, you will need the following parameters:

<Tabs>
  <Tab title="Standard Model">
    | Parameter  | Description                                                                                                                                                                                                                                                                                  |
    | ---------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Token      | Enter the key that you created in your [Aleph Alpha account](https://docs.aleph-alpha.com/).                                                                                                                                                                                                 |
    | Custom URL | This parameter is optional. To control the connection between your clusters and the Aleph Alpha provider, you can route connections through dedicated proxy servers, creating an additional layer of security. To do this, specify the base URL. For example, `https://api.aleph-alpha.com`. |
  </Tab>

  <Tab title="Custom Model">
    | Parameter  | Description                                                                                                                                                                                                                                                                                  |
    | ---------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Model Type | Select the **Completion** type.                                                                                                                                                                                                                                                              |
    | Model Name | Enter the name of the model that you want to use as a custom model. To find model names, refer to the [Aleph Alpha](https://docs.aleph-alpha.com/docs/Deprecated%20Luminous/Deprecated-Luminous/model-card/) documentation.<br />                                                            |
    | Token      | Enter the key that you created in your [Aleph Alpha account](https://docs.aleph-alpha.com/).                                                                                                                                                                                                 |
    | Custom URL | This parameter is optional. To control the connection between your clusters and the Aleph Alpha provider, you can route connections through dedicated proxy servers, creating an additional layer of security. To do this, specify the base URL. For example, `https://api.aleph-alpha.com`. |
  </Tab>
</Tabs>

Apply changes. Check if the connection was set up by clicking **Test**. You can switch generative models at any time using the **Model** list without setting up a new connection. For embedding models, the selection is locked because different embedding models can generate incompatible embeddings for [Knowledge AI](/ai/agents/develop/knowledge-ai/overview#embedding-model). To switch an embedding, delete the connection and create a new one.

### Add Models via the API

You can add either a standard or custom model using the [Cognigy API POST /v2.0/largelanguagemodels](https://api-trial.cognigy.ai/openapi#post-/v2.0/largelanguagemodels) request.
Then, test your connection for the created model via the [Cognigy API POST /v2.0/largelanguagemodels//test](https://api-trial.cognigy.ai/openapi#post-/v2.0/largelanguagemodels/-largeLanguageModelId-/test).

## Apply the Model

To apply a model, follow these steps:

1. In the left-side menu of the Project, go to **Manage > Settings**.
2. Go to the section based on your use case for using a model:
   * **Generative AI Settings**. In the **Generative AI Settings** section, activate **Enable Generative AI Features**. This setting is toggled on by default if you have previously set up the Generative AI credentials.
   * **Knowledge AI Settings**. Use this section if you need to add a model for Knowledge AI. Select a model for the **Knowledge Search** and **Answer Extraction** features. Refer to the [list of standard models](https://docs.cognigy.com/ai/empower/llms/model-support-by-feature/) and find the models that support these features.
3. Navigate to the desired feature and select a model from the list. If there are no models available for the selected feature, the system will automatically select **None**. Save changes.

## More Information

* [Other LLM Operations](/ai/agents/develop/gen-ai-and-llms/other-operations)
* [Overview](/ai/agents/develop/gen-ai-and-llms/llms)
* [All LLM Providers](/ai/agents/develop/gen-ai-and-llms/providers/all-providers)
* [Model support by feature](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature)


# All LLM Providers
Source: https://docs.cognigy.com/ai/agents/develop/gen-ai-and-llms/providers/all-providers



To add a standard or custom model, select a large language model (LLM) provider integrated with Cognigy:

<CardGroup>
  <Card title="Platform-Provided LLM" href="/ai/agents/develop/gen-ai-and-llms/providers/cognigy-platform-provided">
    Cognigy provides a GPT-based LLM for setting up and testing [design-time features](/ai/agents/develop/gen-ai-and-llms/generative-ai#design-time-generative-ai-features).
  </Card>

  <Card title="Azure OpenAI" href="/ai/agents/develop/gen-ai-and-llms/providers/microsoft-azure-openai">
    Add credentials for standard models, such as `gpt-5-4`, `gpt-5-2`, `gpt-5-1`, `gpt-5`, and `text-embedding`, or add a custom model based on an Azure OpenAI model.
  </Card>

  <Card title="OpenAI" href="/ai/agents/develop/gen-ai-and-llms/providers/openai">
    Add credentials for standard models, such as `gpt-5-4`, `gpt-5-2`, `gpt-5-1`, `gpt-5`, and `text-embedding`, or add a custom model based on an OpenAI model.
  </Card>

  <Card title="OpenAI-Compatible LLMs" href="/ai/agents/develop/gen-ai-and-llms/providers/openai-compatible">
    Add credentials for models from any provider that adheres to the OpenAI API standards, including Hugging Face, DeepSeek, xAI, Fireworks AI, and others.
  </Card>

  <Card title="Anthropic" href="/ai/agents/develop/gen-ai-and-llms/providers/anthropic">
    Add credentials for standard models, such as `claude-sonnet-4-6`, `claude-sonnet-4-5`, and `claude-haiku-4-5`, or add a custom model based on an Anthropic model.
  </Card>

  <Card title="Google Vertex" href="/ai/agents/develop/gen-ai-and-llms/providers/google-vertex-ai">
    Add a custom model based on a model from the Vertex AI family.
  </Card>

  <Card title="Google Gemini" href="/ai/agents/develop/gen-ai-and-llms/providers/google-gemini">
    Add credentials for standard models, such as `gemini-2.5-pro` and `gemini-2.5-flash`, or add a custom model based on a model from the Gemini family.
  </Card>

  <Card title="Google GenAI" href="/ai/agents/develop/gen-ai-and-llms/providers/google-genai">
    Add credentials for standard models, such as `gemini-3.1-pro-preview`, `gemini-3.1-flash-lite`, `gemini-3-flash-preview`, `gemini-3.5-flash`, `gemini-3.6-flash`, `gemini-3.7-flash`, and `gemini-3.5-flash-lite`.
  </Card>

  <Card title="Aleph Alpha" href="/ai/agents/develop/gen-ai-and-llms/providers/aleph-alpha">
    Add credentials for the standard `luminous` model, or add a custom model based on an Aleph Alpha model.
  </Card>

  <Card title="Amazon Bedrock" href="/ai/agents/develop/gen-ai-and-llms/providers/amazon-bedrock">
    Add credentials for standard models or add a custom model based on any Amazon Bedrock model that supports the [Converse API](https://docs.aws.amazon.com/bedrock/latest/userguide/models-features.html).
  </Card>

  <Card title="Mistral AI" href="/ai/agents/develop/gen-ai-and-llms/providers/mistral">
    Add credentials for standard models, such as `mistral-large-latest` and `mistral-small-latest`, or add a custom model based on a Mistral AI model.
  </Card>

  <Card title="Cognigy Nexus LLM" href="/ai/agents/develop/gen-ai-and-llms/providers/cognigy-nexus-llm">
    Cognigy Nexus LLM is a proprietary LLM, provided and hosted by Cognigy, optimized for AI-driven customer service.
  </Card>
</CardGroup>

***

<sup>1</sup>: The platform-provided LLM is available only to customers with SaaS installations.


# Amazon Bedrock
Source: https://docs.cognigy.com/ai/agents/develop/gen-ai-and-llms/providers/amazon-bedrock



<a href="/release-notes/2026.11"><Badge>Updated in 2026.11</Badge></a>

To start using an Amazon Bedrock model with Cognigy features, follow these steps:

1. [Add a Model](#add-models)
2. [Apply the Model](#apply-the-model)

## Restrictions

* AWS Bedrock doesn't accept requests with an empty or whitespace-only system prompt and returns an HTTP 400 error. This behavior applies to the following Nodes where system prompts are configured:
  * [LLM Prompt Node](/ai/agents/develop/node-reference/service/llm-prompt) — **System Prompt** field.
  * [Question Node](/ai/agents/develop/node-reference/basic/question) — **LLM Instructions** field.
  * [AI Agent Node](/ai/agents/develop/node-reference/ai/ai-agent) — the system prompt is built from **Persona**, **Job Description**, and **Instructions and Context**. The restriction applies only when all three fields are empty.

## Add Models

You can add a model using one of the following interfaces:

* [GUI](#add-models-via-gui)
* [API](#add-models-via-the-api)

### Add Models via GUI

You can add a model provided by Amazon Bedrock to Cognigy in **Build > LLM**. To add the model, you will need the following parameters:

<Tabs>
  <Tab title="Standard Model">
    | Parameter         | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
    | ----------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Access Key ID     | Enter the Access Key ID. Log in to the AWS Management Console, go to the IAM dashboard, select **Users**, and choose the IAM user. Navigate to the **Security credentials** tab, and under **Access keys**, create a new access key if one hasn't been created. Copy the Access Key ID provided after creation.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
    | Secret Access Key | Enter the Secret Access Key. After creating the access key, you'll be prompted to download a file containing the Access Key ID and the Secret Access Key. Alternatively, you can retrieve the Secret Access Key by navigating to the IAM dashboard, selecting the user, going to the **Security credentials** tab, and clicking **Show** next to the Access Key ID to reveal and copy the Secret Access Key.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
    | Region            | Enter the AWS region where your model is located, for example, `us-east-1` for the US East (N. Virginia) region.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
    | Location          | Select the routing behavior of your inference calls. Select one of the following options:<ul><li>**Region** — the data sent in the request to the LLM stays within the AWS region you provided in **Region**. This is the default behavior.</li><li>**Geo** — the data sent in the request can be routed within a geographic boundary. This option allows for higher request throughput while keeping the benefits of a geographic boundary, for example, data residency for compliance with regulations. Enter the geographic boundary in the **Geo** field, for example, `us`, `eu`, or `apac`. The AWS region must be located within the geographic boundary entered in this parameter.</li><li>**Global** — the data sent in the request can be routed worldwide. This option allows for the highest request throughput and best performance for cases with no data residency constraints.</li></ul>For more information, read the [Regional availability](https://docs.aws.amazon.com/bedrock/latest/userguide/models-region-compatibility.html) article. |
  </Tab>

  <Tab title="Custom Model">
    | Parameter         | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
    | ----------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Model Type        | Select the **Chat** type for models that support the [Converse API](https://docs.aws.amazon.com/bedrock/latest/userguide/conversation-inference-supported-models-features.html). Note that the model will only work if your AWS administrator gives you [access to this model](https://docs.aws.amazon.com/bedrock/latest/userguide/model-access.html).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
    | Model Name        | Enter the ID of the model that you want to use as a custom model. To find model IDs, refer to the [Amazon Bedrock](https://docs.aws.amazon.com/bedrock/latest/userguide/model-ids.html) documentation.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
    | Access Key ID     | Enter the Access Key ID. Log in to the AWS Management Console, go to the IAM dashboard, select **Users**, and choose the IAM user. Navigate to the **Security credentials** tab, and under **Access keys**, create a new access key if one hasn't been created. Copy the Access Key ID provided after creation.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |
    | Secret Access Key | Enter the Secret Access Key. After creating the access key, you'll be prompted to download a file containing the Access Key ID and the Secret Access Key. Alternatively, you can retrieve the Secret Access Key by navigating to the IAM dashboard, selecting the user, going to the **Security credentials** tab, and clicking **Show** next to the Access Key ID to reveal and copy the Secret Access Key.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
    | Region            | Enter the AWS region where your model is located, for example, `us-east-1` for the US East (N. Virginia) region.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | Location          | Select one of the following routing behaviors for your inference calls:<ul><li>**Region** — the data sent in the request to the LLM stays within the AWS region you provided in **Region**. This is the default behavior.</li><li>**Geo** — the data sent in the request can be routed within a geographic boundary. This option allows for higher request throughput while keeping the benefits of a geographic boundary, for example, data residency for compliance with regulations. Enter the geographic boundary in the **Geo** field, for example, `us`, `eu`, or `apac`. The AWS region must be located within the geographic boundary entered in this parameter.</li><li>**Global** — the data sent in the request can be routed worldwide. This option allows for the highest request throughput and best performance for cases with no data residency constraints.</li></ul>For more information, read the [Regional availability](https://docs.aws.amazon.com/bedrock/latest/userguide/models-region-compatibility.html) article. |
  </Tab>
</Tabs>

Apply changes. Check if the connection was set up by clicking **Test**. You can switch generative models at any time using the **Model** list without setting up a new connection. For embedding models, the selection is locked because different embedding models can generate incompatible embeddings for [Knowledge AI](/ai/agents/develop/knowledge-ai/overview#embedding-model). To switch an embedding, delete the connection and create a new one.

### Add Models via the API

You can add either a standard or custom model using the [Cognigy API POST /v2.0/largelanguagemodels](https://api-trial.cognigy.ai/openapi#post-/v2.0/largelanguagemodels) request.
Then, test your connection for the created model via the [Cognigy API POST /v2.0/largelanguagemodels//test](https://api-trial.cognigy.ai/openapi#post-/v2.0/largelanguagemodels/-largeLanguageModelId-/test).

## Apply the Model

To apply a model, follow these steps:

1. In the left-side menu of the Project, go to **Manage > Settings**.
2. Go to the section based on your use case for using a model:
   * **Generative AI Settings**. In the **Generative AI Settings** section, activate **Enable Generative AI Features**. This setting is toggled on by default if you have previously set up the Generative AI credentials.
   * **Knowledge AI Settings**. Use this section if you need to add a model for Knowledge AI. Select a model for the **Knowledge Search** and **Answer Extraction** features. Refer to the [list of standard models](https://docs.cognigy.com/ai/empower/llms/model-support-by-feature/) and find the models that support these features.
3. Navigate to the desired feature and select a model from the list. If there are no models available for the selected feature, the system will automatically select **None**. Save changes.

## More Information

* [Other LLM Operations](/ai/agents/develop/gen-ai-and-llms/other-operations)
* [Overview](/ai/agents/develop/gen-ai-and-llms/llms)
* [All LLM Providers](/ai/agents/develop/gen-ai-and-llms/providers/all-providers)
* [Model support by feature](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature)


# Anthropic
Source: https://docs.cognigy.com/ai/agents/develop/gen-ai-and-llms/providers/anthropic



<a href="/release-notes/2026.17"><Badge>Updated in 2026.17</Badge></a>

<Warning>
  As of Cognigy 2026.6.0, the support for the Completions API for custom LLMs is deprecated. The full removal is planned for Cognigy 2026.24.0.&#x20;
</Warning>

<Warning>
  Anthropic's `claude-sonnet-4-6` model doesn't support [assistant message
  prefilling](https://platform.claude.com/docs/en/about-claude/models/migration-guide#breaking-changes-3). If your AI Agent uses this model and a conversation ends with an assistant message, for example,
  after a [Say Node](/ai/agents/develop/node-reference/basic/say), requests can
  fail with the error `The conversation must end with a user message`.

  To keep your Flows compatible, ensure that the last transcript step before the
  LLM call is a user message. For example, add an [Add Transcript Step
  Node](/ai/agents/develop/node-reference/service/add-transcript-step) before
  the LLM call and set the role to `user`.
</Warning>

<Note>
  The `claude-sonnet-5`, `claude-opus-4-7`, and `claude-opus-4-8` models don't support token-sampling parameters such as `temperature`, `top_p`, or `top_k`. If you set these parameters at the Node level or in **Custom Model Options**, Cognigy automatically removes them before sending the request. Instead, set Anthropic's `thinking.type` and `thinking.effort` parameters in **Custom Model Options** to control model behavior.
</Note>

To start using an Anthropic model with Cognigy features, follow these steps:

1. [Add a Model](#add-models)
2. [Apply the Model](#apply-the-model)

## Add Models

You can add a model using one of the following interfaces:

* [GUI](#add-models-via-gui)
* [API](#add-models-via-the-api)

### Add Models via GUI

You can add a model provided by Anthropic to Cognigy in **Build > LLM**. To add the model, you will need the following parameters:

<Tabs>
  <Tab title="Standard Model">
    <Note>
      The `*-latest` suffix indicates that the model you select points to the latest version of the model. For more information, read the [Anthropic's models](https://docs.anthropic.com/en/docs/about-claude/models/overview) documentation.
    </Note>

    | Parameter | Description                                                                                                             |
    | --------- | ----------------------------------------------------------------------------------------------------------------------- |
    | apiKey    | Add an API Key that you generated via [Account Settings](https://platform.claude.com/docs/en/get-started) in Anthropic. |
  </Tab>

  <Tab title="Custom Model">
    | Parameter  | Description                                                                                                                                                                                                                                              |
    | ---------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Model Type | Select **Chat** for models that support the Messages API and **Completion** for the Completions API. For more information, refer to the [Anthropic Model Comparison (API format)](https://docs.anthropic.com/claude/docs/models-overview) documentation. |
    | Model Name | Enter the name of the model that you want to use as a custom model. To find model names, refer to the [Anthropic](https://docs.anthropic.com/claude/docs/models-overview) documentation.                                                                 |
    | apiKey     | Add an API Key that you generated via [Account Settings](https://platform.claude.com/docs/en/get-started) in Anthropic.                                                                                                                                  |
  </Tab>
</Tabs>

Apply changes. Check if the connection was set up by clicking **Test**. You can switch generative models at any time using the **Model** list without setting up a new connection. For embedding models, the selection is locked because different embedding models can generate incompatible embeddings for [Knowledge AI](/ai/agents/develop/knowledge-ai/overview#embedding-model). To switch an embedding, delete the connection and create a new one.

### Add Models via the API

You can add either a standard or custom model using the [Cognigy API POST /v2.0/largelanguagemodels](https://api-trial.cognigy.ai/openapi#post-/v2.0/largelanguagemodels) request.
Then, test your connection for the created model via the [Cognigy API POST /v2.0/largelanguagemodels//test](https://api-trial.cognigy.ai/openapi#post-/v2.0/largelanguagemodels/-largeLanguageModelId-/test).

## Apply the Model

To apply a model, follow these steps:

1. In the left-side menu of the Project, go to **Manage > Settings**.
2. Go to the section based on your use case for using a model:
   * **Generative AI Settings**. In the **Generative AI Settings** section, activate **Enable Generative AI Features**. This setting is toggled on by default if you have previously set up the Generative AI credentials.
   * **Knowledge AI Settings**. Use this section if you need to add a model for Knowledge AI. Select a model for the **Knowledge Search** and **Answer Extraction** features. Refer to the [list of standard models](https://docs.cognigy.com/ai/empower/llms/model-support-by-feature/) and find the models that support these features.
3. Navigate to the desired feature and select a model from the list. If there are no models available for the selected feature, the system will automatically select **None**. Save changes.

## More Information

* [Other LLM Operations](/ai/agents/develop/gen-ai-and-llms/other-operations)
* [Overview](/ai/agents/develop/gen-ai-and-llms/llms)
* [All LLM Providers](/ai/agents/develop/gen-ai-and-llms/providers/all-providers)
* [Model support by feature](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature)


# Cognigy Platform-Provided LLM
Source: https://docs.cognigy.com/ai/agents/develop/gen-ai-and-llms/providers/cognigy-platform-provided



<a href="/release-notes/2025.14"><Badge>Updated in 2025.14</Badge></a>

Cognigy offers a platform-provided LLM based on the Azure OpenAI GPT-4o family of models. This model is fully integrated into the platform and requires no additional setup or configuration.
Use this model to set up and test [design-time features](/ai/agents/develop/gen-ai-and-llms/generative-ai#design-time-generative-ai-features).

## Prerequisites

* The platform-provided LLM is available only for SaaS installations.

## Apply the Model

Starting with version 2025.14, this model is selected by default for design-time features in new Projects.

To apply another model, go to **Manage > Settings > Generative AI Settings** and select a model in the **Design-Time LLM Features** section.

For Projects created before 2025.14, the model you previously selected will remain in use.


# Google Gemini
Source: https://docs.cognigy.com/ai/agents/develop/gen-ai-and-llms/providers/google-gemini



<a href="/release-notes/2026.11"><Badge>Updated in 2026.11</Badge></a>

To start using a Google Gemini model with Cognigy features, follow these steps:

1. [Add a Model](#add-models)
2. [Apply the Model](#apply-the-model)

## Restriction

* Only users with the [admin role](/ai/administer/access/members) in Cognigy can upload JSON service account files to configure Google models.

## Add Models

You can add a model using one of the following interfaces:

* [GUI](#add-models-via-gui)
* [API](#add-models-via-the-api)

### Add Models via GUI

You can add a model provided by Google Gemini to Cognigy in **Build > LLM**. To add the model, you will need the following parameters:

<Tabs>
  <Tab title="Standard Model">
    | Parameter | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
    | --------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | JSON      | To upload the JSON file with a key for your model, you need to obtain this key. If you have previously used a key for the Google Vertex AI connection, you can also use this key for Google Gemini. If you're setting up the connection for the first time, go to the Google Cloud console and find Vertex AI via the search bar. On the **Vertex AI** page, click the **Enable All Recommended APIs** button to activate an API connection, if it is not activated. Ensure that the Vertex AI API is enabled. In the left-side menu, go to the **IAM & Admin > Service Accounts**. Select **Actions** and click **Manage Keys**. On the **Keys** page, select **Add Key** and click **Create new Key**. In the appeared window, select the **JSON** key type and click **Create**. The file will be downloaded. |
    | Location  | Add a [region](https://cloud.google.com/vertex-ai/docs/general/locations) for the model. For example, `us-central1`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
  </Tab>

  <Tab title="Custom Model">
    | Parameter  | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
    | ---------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Model Type | Select the **Chat** type.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
    | Model Name | Enter the name of the model that you want to use as a custom model. To find model names, refer to the [Google Gemini](https://ai.google.dev/gemini-api/docs/models/gemini) documentation. Note that, within this connection, Cognigy supports only the `gemini` models.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
    | JSON       | To upload the JSON file with a key for your model, you need to obtain this key. If you have previously used a key for the Google Vertex AI connection, you can also use this key for Google Gemini. If you're setting up the connection for the first time, go to the Google Cloud console and find Vertex AI via the search bar. On the **Vertex AI** page, click the **Enable All Recommended APIs** button to activate an API connection, if it is not activated. Ensure that the Vertex AI API is enabled. In the left-side menu, go to the **IAM & Admin > Service Accounts**. Select **Actions** and click **Manage Keys**. On the **Keys** page, select **Add Key** and click **Create new Key**. In the appeared window, select the **JSON** key type and click **Create**. The file will be downloaded. |
    | Location   | Add a [region](https://cloud.google.com/vertex-ai/docs/general/locations) for the model. For example, `us-central1`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
  </Tab>
</Tabs>

Apply changes. Check if the connection was set up by clicking **Test**. You can switch generative models at any time using the **Model** list without setting up a new connection. For embedding models, the selection is locked because different embedding models can generate incompatible embeddings for [Knowledge AI](/ai/agents/develop/knowledge-ai/overview#embedding-model). To switch an embedding, delete the connection and create a new one.

### Add Models via the API

You can add either a standard or custom model using the [Cognigy API POST /v2.0/largelanguagemodels](https://api-trial.cognigy.ai/openapi#post-/v2.0/largelanguagemodels) request.
Then, test your connection for the created model via the [Cognigy API POST /v2.0/largelanguagemodels//test](https://api-trial.cognigy.ai/openapi#post-/v2.0/largelanguagemodels/-largeLanguageModelId-/test).

## Apply the Model

To apply a model, follow these steps:

1. In the left-side menu of the Project, go to **Manage > Settings**.
2. Go to the section based on your use case for using a model:
   * **Generative AI Settings**. In the **Generative AI Settings** section, activate **Enable Generative AI Features**. This setting is toggled on by default if you have previously set up the Generative AI credentials.
   * **Knowledge AI Settings**. Use this section if you need to add a model for Knowledge AI. Select a model for the **Knowledge Search** and **Answer Extraction** features. Refer to the [list of standard models](https://docs.cognigy.com/ai/empower/llms/model-support-by-feature/) and find the models that support these features.
3. Navigate to the desired feature and select a model from the list. If there are no models available for the selected feature, the system will automatically select **None**. Save changes.

## More Information

* [Other LLM Operations](/ai/agents/develop/gen-ai-and-llms/other-operations)
* [Overview](/ai/agents/develop/gen-ai-and-llms/llms)
* [All LLM Providers](/ai/agents/develop/gen-ai-and-llms/providers/all-providers)
* [Model support by feature](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature)


# Google GenAI
Source: https://docs.cognigy.com/ai/agents/develop/gen-ai-and-llms/providers/google-genai



<a href="/release-notes/2026.13"><Badge>Updated in 2026.13</Badge></a>

To start using a Google GenAI model with Cognigy features, follow these steps:

1. [Add a Model](#add-models)
2. [Apply the Model](#apply-the-model)

## Restriction

* Only users with the [admin role](/ai/administer/access/members) in Cognigy can upload JSON service account files to configure Google models.

## Add Models

You can add a model using one of the following interfaces:

* [GUI](#add-models-via-gui)
* [API](#add-models-via-the-api)

### Add Models via GUI

You can add a model provided by Google GenAI to Cognigy in **Build > LLM**. To add the model, you will need the following parameters:

<Tabs>
  <Tab title="Standard Model">
    | Parameter | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
    | --------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | JSON      | To upload the JSON file with a key for your model, you need to obtain this key. If you have previously used a key for the Google Vertex AI or Google Gemini connection, you can reuse this key for Google GenAI. If you're setting up the connection for the first time, go to the Google Cloud console and find Vertex AI via the search bar. On the **Vertex AI** page, click the **Enable All Recommended APIs** button to activate an API connection, if it is not activated. Ensure that the Vertex AI API is enabled. In the left-side menu, go to the **IAM & Admin > Service Accounts**. Select **Actions** and click **Manage Keys**. On the **Keys** page, select **Add Key** and click **Create new Key**. In the appeared window, select the **JSON** key type and click **Create**. The file will be downloaded. |
    | Location  | Add a [region](https://cloud.google.com/vertex-ai/docs/general/locations) for the model. For example, `us-central1`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
  </Tab>

  <Tab title="Custom Model">
    <Note>
      The Google GenAI provider supports custom models based on any model from the [Gemini families](https://ai.google.dev/gemini-api/docs/models).
    </Note>

    | Parameter  | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
    | ---------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Model Type | Select the model type: **Chat** or **Embedding**.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | Model Name | Enter the name of the model that you want to use as a custom model. To find model names, refer to the [Google GenAI](https://ai.google.dev/gemini-api/docs/models/gemini) documentation.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
    | JSON       | To upload the JSON file with a key for your model, you need to obtain this key. If you have previously used a key for the Google Vertex AI or Google Gemini connection, you can reuse this key for Google GenAI. If you're setting up the connection for the first time, go to the Google Cloud console and find Vertex AI via the search bar. On the **Vertex AI** page, click the **Enable All Recommended APIs** button to activate an API connection, if it is not activated. Ensure that the Vertex AI API is enabled. In the left-side menu, go to the **IAM & Admin > Service Accounts**. Select **Actions** and click **Manage Keys**. On the **Keys** page, select **Add Key** and click **Create new Key**. In the appeared window, select the **JSON** key type and click **Create**. The file will be downloaded. |
    | Location   | Add a [region](https://cloud.google.com/vertex-ai/docs/general/locations) for the model. For example, `us-central1`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
  </Tab>
</Tabs>

Apply changes. Check if the connection was set up by clicking **Test**. You can switch generative models at any time using the **Model** list without setting up a new connection. For embedding models, the selection is locked because different embedding models can generate incompatible embeddings for [Knowledge AI](/ai/agents/develop/knowledge-ai/overview#embedding-model). To switch an embedding, delete the connection and create a new one.

### Add Models via the API

You can add either a standard or custom model using the [Cognigy API POST /v2.0/largelanguagemodels](https://api-trial.cognigy.ai/openapi#post-/v2.0/largelanguagemodels) request.
Then, test your connection for the created model via the [Cognigy API POST /v2.0/largelanguagemodels//test](https://api-trial.cognigy.ai/openapi#post-/v2.0/largelanguagemodels/-largeLanguageModelId-/test).

## Apply the Model

To apply a model, follow these steps:

1. In the left-side menu of the Project, go to **Manage > Settings**.
2. Go to the section based on your use case for using a model:
   * **Generative AI Settings**. In the **Generative AI Settings** section, activate **Enable Generative AI Features**. This setting is toggled on by default if you have previously set up the Generative AI credentials.
   * **Knowledge AI Settings**. Use this section if you need to add a model for Knowledge AI. Select a model for the **Knowledge Search** and **Answer Extraction** features. Refer to the [list of standard models](https://docs.cognigy.com/ai/empower/llms/model-support-by-feature/) and find the models that support these features.
3. Navigate to the desired feature and select a model from the list. If there are no models available for the selected feature, the system will automatically select **None**. Save changes.

## More Information

* [Other LLM Operations](/ai/agents/develop/gen-ai-and-llms/other-operations)
* [Overview](/ai/agents/develop/gen-ai-and-llms/llms)
* [All LLM Providers](/ai/agents/develop/gen-ai-and-llms/providers/all-providers)
* [Model support by feature](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature)


# Google Vertex AI
Source: https://docs.cognigy.com/ai/agents/develop/gen-ai-and-llms/providers/google-vertex-ai



<a href="/release-notes/2026.11"><Badge>Updated in 2026.11</Badge></a>

<Warning>
  As of Cognigy 2026.6.0, the support for the Completions API for custom LLMs is deprecated. The full removal is planned for Cognigy 2026.24.0.&#x20;
</Warning>

To start using a Google Vertex AI or Gemini model with Cognigy features, follow these steps:

1. [Add a Model](#add-models)
2. [Apply the Model](#apply-the-model)

## Restriction

* Only users with the [admin role](/ai/administer/access/members) in Cognigy can upload JSON service account files to configure Google models.

## Add Models

You can add a model using one of the following interfaces:

* [GUI](#add-models-via-gui)
* [API](#add-models-via-the-api)

### Add Models via GUI

To add a model provided by Google Vertex AI to Cognigy, follow these steps:

<Tabs>
  <Tab title="Standard Model (Cognigy 2025.15 and earlier versions)">
    | Parameter    | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
    | ------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | JSON         | To upload the JSON file with a key for your model, you need to obtain this key. Go to the Google Cloud console and find Vertex AI via the search bar. On the **Vertex AI** page, click the **Enable All Recommended APIs** button to activate an API connection, if it is not activated. Ensure that the Vertex AI API is enabled. In the left-side menu, go to the **IAM & Admin > Service Accounts**. Select **Actions** and click **Manage Keys**. On the **Keys** page, select **Add Key** and click **Create new Key**. In the appeared window, select the **JSON** key type and click **Create**. The file will be downloaded. You can also use this key for [Google Gemini](/ai/agents/develop/gen-ai-and-llms/providers/google-gemini). |
    | Location     | Add a [region](https://cloud.google.com/vertex-ai/docs/general/locations) for the model. For example, `us-central1`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | API Endpoint | Add a [service endpoint](https://cloud.google.com/vertex-ai/docs/reference/rest#service-endpoint) for the model. For example, `us-central1-aiplatform.googleapis.com`. Note that the endpoint should be specified without `https://` or `http://`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |
    | Publisher    | Add an owner's name of the model. If not specified, `Google` will be used by default.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
  </Tab>

  <Tab title="Custom Model">
    | Parameter    | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
    | ------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Model Type   | Select the **Completion** type.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
    | Model Name   | Enter the name of the model that you want to use as a custom model. To find model names, refer to the [Google Vertex AI](https://cloud.google.com/vertex-ai/generative-ai/docs/learn/model-versioning) documentation. Note that, within this connection, Cognigy supports only the `text-bison` models.                                                                                                                                                                                                                                                                                                                                                                                                                                         |
    | JSON         | To upload the JSON file with a key for your model, you need to obtain this key. Go to the Google Cloud console and find Vertex AI via the search bar. On the **Vertex AI** page, click the **Enable All Recommended APIs** button to activate an API connection, if it is not activated. Ensure that the Vertex AI API is enabled. In the left-side menu, go to the **IAM & Admin > Service Accounts**. Select **Actions** and click **Manage Keys**. On the **Keys** page, select **Add Key** and click **Create new Key**. In the appeared window, select the **JSON** key type and click **Create**. The file will be downloaded. You can also use this key for [Google Gemini](/ai/agents/develop/gen-ai-and-llms/providers/google-gemini). |
    | Location     | Add a [region](https://cloud.google.com/vertex-ai/docs/general/locations) for the model. For example, `us-central1`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | API Endpoint | Add a [service endpoint](https://cloud.google.com/vertex-ai/docs/reference/rest#service-endpoint) for the model. For example, `us-central1-aiplatform.googleapis.com`. Note that the endpoint should be specified without `https://` or `http://`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |
    | Publisher    | Add an owner's name of the model. If not specified, `Google` will be used by default.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
  </Tab>
</Tabs>

Apply changes. Check if the connection was set up by clicking **Test**. You can switch generative models at any time using the **Model** list without setting up a new connection. For embedding models, the selection is locked because different embedding models can generate incompatible embeddings for [Knowledge AI](/ai/agents/develop/knowledge-ai/overview#embedding-model). To switch an embedding, delete the connection and create a new one.

### Add Models via the API

You can add either a standard or custom model using the [Cognigy API POST /v2.0/largelanguagemodels](https://api-trial.cognigy.ai/openapi#post-/v2.0/largelanguagemodels) request.
Then, test your connection for the created model via the [Cognigy API POST /v2.0/largelanguagemodels//test](https://api-trial.cognigy.ai/openapi#post-/v2.0/largelanguagemodels/-largeLanguageModelId-/test).

## Apply the Model

To apply a model, follow these steps:

1. In the left-side menu of the Project, go to **Manage > Settings**.
2. Go to the section based on your use case for using a model:
   * **Generative AI Settings**. In the **Generative AI Settings** section, activate **Enable Generative AI Features**. This setting is toggled on by default if you have previously set up the Generative AI credentials.
   * **Knowledge AI Settings**. Use this section if you need to add a model for Knowledge AI. Select a model for the **Knowledge Search** and **Answer Extraction** features. Refer to the [list of standard models](https://docs.cognigy.com/ai/empower/llms/model-support-by-feature/) and find the models that support these features.
3. Navigate to the desired feature and select a model from the list. If there are no models available for the selected feature, the system will automatically select **None**. Save changes.

## More Information

* [Other LLM Operations](/ai/agents/develop/gen-ai-and-llms/other-operations)
* [Overview](/ai/agents/develop/gen-ai-and-llms/llms)
* [All LLM Providers](/ai/agents/develop/gen-ai-and-llms/providers/all-providers)
* [Model support by feature](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature)


# Microsoft Azure OpenAI
Source: https://docs.cognigy.com/ai/agents/develop/gen-ai-and-llms/providers/microsoft-azure-openai



<a href="/release-notes/2026.17"><Badge>Updated in 2026.17</Badge></a>

<Warning>
  Use of the **Model** field for deployment selection in Responses API requests has been deprecated. The removal is planned for Cognigy 2027.3.0. Use the **Deployment Name** field instead by activating the **Use deployment name as model** parameter.
</Warning>

<Warning>
  As of Cognigy 2026.6.0, the support for the Completions API for custom LLMs is deprecated. The full removal is planned for Cognigy 2026.24.0.&#x20;
</Warning>

To start using a Microsoft Azure OpenAI model with Cognigy features, follow these steps:

1. [Add a Model](#add-models)
2. [Apply the Model](#apply-the-model)

## Add Models

You can add a model using one of the following interfaces:

* [GUI](#add-models-via-gui)
* [API](#add-models-via-the-api)

### Add Models via GUI

You can add a model provided by Azure OpenAI to Cognigy in **Build > LLM**. To add the model, you will need the following parameters:

<Tabs>
  <Tab title="Standard Model">
    | Parameter                    | Option  | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         |
    | ---------------------------- | ------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Connection Type              |         | Select one of the following authorization methods: API Key or OAuth2.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
    |                              | API Key | Enter the [Azure API Key](https://learn.microsoft.com/en-us/azure/cognitive-services/openai/quickstart?tabs=command-line\&pivots=rest-api#retrieve-key-and-endpoint). This value can be found in the **Keys & Endpoint** section when examining your resource from the Azure portal. You can use either `KEY1` or `KEY2`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    |                              | OAuth2  | Enter the OAuth2 key. You can use any OAuth2-compliant provider. OAuth 2.0 offers more control and security than API keys by allowing specific permissions, expiring tokens, and reducing exposure through short-lived tokens instead of constant client secret use. To use this type of connection, fill in the following fields:<ul><li>**Client ID** — enter the Application (client) ID assigned to your app. This ID can be found in the Microsoft Entra admin center: go to **App Registrations**, select an app, then click **Overview**.</li><li>**Client Secret** — enter the application secret created in the **Certificates & secrets** section in the Microsoft Entra admin center: go to **App Registrations**, select an app, then click **Overview**.</li><li>**OAuth URL** — enter the URL to retrieve the access token. This URL points to the `/token` endpoint on the authorization server. The URL can have any format that returns an access token.</li><li>**Scope** — add a list of scopes for user permissions, for example, `urn:grp:chatgpt`.</li><li>**Additional Header Name** — enter the key for an additional authentication header. This key can have any format, for example, of the [OCP APIM Subscription Key](https://learn.microsoft.com/en-us/azure/api-management/api-management-subscriptions).</li><li>**Additional Header Value** — enter the value for the additional authentication header.</li></ul>Cognigy uses the following steps to make requests with OAuth 2.0 to an Azure OpenAI model: <ol><li>**Requesting a token.** Cognigy sends an HTTP POST request to <code>OAuth URL</code> with the following parameters: <code>Client ID</code>, <code>Client Secret</code>, and <code>Scope</code>.</li><li>**Receiving the token.** The authorization server validates the request and returns an access token to Cognigy.</li><li>**Using the token.** When making requests to the Azure OpenAI model, Cognigy includes the received token in the request instead of using an API key.</li></ol> |
    | Resource Name                |         | Enter the [resource name](https://learn.microsoft.com/en-us/azure/cognitive-services/openai/how-to/create-resource?pivots=web-portal#create-a-resource). To find this value, go to the **Microsoft Azure** home page. Under **Azure services**, click **Azure OpenAI**. In the left-side menu, under the **Azure AI Services** section, select **Azure OpenAI**. Copy the desired resource name from the **Name** column.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | Deployment Name              |         | Enter the [deployment name](https://learn.microsoft.com/en-us/azure/cognitive-services/openai/how-to/create-resource?pivots=web-portal#deploy-a-model). To find this value, go to the **Microsoft Azure** home page. Under **Azure services**, click **Azure OpenAI**. In the left-side menu, under **Azure AI Services**, select **Azure OpenAI**. Select a resource from the **Name** column. On the resource page, go to **Resource Management > Model deployments**. On the **Model deployments** page, click **Manage Deployments**. On the **Deployments** page, copy the desired deployment name from the **Deployment name** column.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
    | API Version                  |         | Enter the [API version](https://learn.microsoft.com/en-us/azure/cognitive-services/openai/reference#rest-api-versioning). The API version to use for this operation in the `YYYY-MM-DD` format. Note that the version may have an extended format, for example, `YYYY-MM-DD-preview`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
    | Custom URL                   |         | This parameter is optional. You can use it to route connections between your clusters and the Azure OpenAI provider through dedicated proxy servers for enhanced security. The custom URL can be any valid endpoint URL and must be complete, including any required query parameters. When you specify a custom URL, the **Resource Name**, **Deployment Name**, and **API Version** fields are ignored.<br />For versions 2025.20 and earlier, when selecting the **API key** connection type, you must use the following pattern: `https://<resource-name>.openai.azure.com/openai/deployments/<deployment-name>/<model-type>?api-version=<api-version>`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
    | API Type                     |         | Select the API format for chat models. Select one of the following API options for chat models:<ul><li>**Chat Completion** — uses the [Chat Completions API](https://learn.microsoft.com/en-us/azure/ai-foundry/openai/how-to/chatgpt). This option is selected by default.</li><li>**Responses** — uses the [Responses API](https://learn.microsoft.com/en-us/azure/ai-foundry/openai/how-to/responses).</li></ul> For more information, refer to the [Microsoft Azure OpenAI](https://learn.microsoft.com/en-us/azure/foundry/foundry-models/concepts/models-sold-directly-by-azure?tabs=americas%2Caz-global-standard%2Cglobal-standard\&pivots=azure-openai) documentation.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
    | Use deployment name as model |         | Appears only when **Responses** is selected as the API type. When this parameter is activated, Cognigy sends the value of **Deployment Name** as the model selector in Responses API requests. Microsoft Azure OpenAI deployments require this behavior. When this parameter is deactivated, Cognigy uses the legacy behavior and sends the value of **Model** instead of **Deployment Name**. New connections have this parameter active by default. Existing connections have it inactive to preserve backward compatibility.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
  </Tab>

  <Tab title="Custom Model">
    | Parameter                    | Option  | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
    | ---------------------------- | ------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Model Type                   |         | Select one of the following options:<ul><li>**Chat** — lets you use either the [Chat Completions API](https://learn.microsoft.com/en-us/azure/ai-foundry/openai/how-to/chatgpt) or [Responses API](https://learn.microsoft.com/en-us/azure/ai-foundry/openai/how-to/responses) for chat models.</li><li>**Completion** — uses the Completions API.</li><li>**Embedding** — uses the Embedding API.</li></ul>For more information, refer to the [Azure OpenAI](https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/models) documentation.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
    | Model Name                   |         | Enter the ID of the model that you want to use as a custom model. To find model IDs, refer to the [Azure OpenAI](https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/models) documentation.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
    | Connection Type              |         | Select one of the following authorization methods: API Key or OAuth2.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
    |                              | API Key | Enter the [Azure API Key](https://learn.microsoft.com/en-us/azure/cognitive-services/openai/quickstart?tabs=command-line\&pivots=rest-api#retrieve-key-and-endpoint). This value can be found in the **Keys & Endpoint** section when examining your resource from the Azure portal. You can use either `KEY1` or `KEY2`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
    |                              | OAuth2  | Enter the OAuth2 key. You can use any OAuth2-compliant provider. OAuth 2.0 offers more control and security than API keys by allowing specific permissions, expiring tokens, and reducing exposure through short-lived tokens instead of constant client secret use. To use this type of connection, fill in the following fields:<ul><li>**Client ID** — add the Application (client) ID assigned to your app. This ID can be found in the Microsoft Entra admin center: go to **App Registrations**, select an app, then click **Overview**.</li><li>**Client Secret** — add the application secret created in the **Certificates & secrets** section in the Microsoft Entra admin center: go to **App Registrations**, select an app, then click **Overview**.</li><li>**OAuth URL** — add the URL to retrieve the access token. This URL points to the `/token` endpoint on the authorization server. The URL can have any format that returns an access token.</li><li>**Scope** — add a list of scopes for user permissions, for example, `urn:grp:chatgpt`.</li><li>**Additional Header Name** — enter the key for an additional authentication header. This key can have any format, for example, of the [OCP APIM Subscription Key](https://learn.microsoft.com/en-us/azure/api-management/api-management-subscriptions).</li><li>**Additional Header Value** — enter the value for the additional authentication header.</li></ul>Cognigy uses the following steps to make requests with OAuth 2.0 to an Azure OpenAI model:<ol><li>**Requesting a token.** Cognigy sends an HTTP POST request to <code>OAuth URL</code> with the following parameters: <code>Client ID</code>, <code>Client Secret</code>, and <code>Scope</code>.</li><li>**Receiving the token.** The authorization server validates the request and returns an access token to Cognigy.</li><li>**Using the token.** When making requests to the Azure OpenAI model, Cognigy includes the received token in the request instead of using an API key.</li></ol> |
    | Resource Name                |         | Enter the [resource name](https://learn.microsoft.com/en-us/azure/cognitive-services/openai/how-to/create-resource?pivots=web-portal#create-a-resource). To find this value, go to the **Microsoft Azure** home page. Under **Azure services**, click **Azure OpenAI**. In the left-side menu, under the **Azure AI Services** section, select **Azure Open AI**. Copy the desired resource name from the **Name** column.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
    | Deployment Name              |         | Enter the [deployment name](https://learn.microsoft.com/en-us/azure/cognitive-services/openai/how-to/create-resource?pivots=web-portal#deploy-a-model). To find this value, go to the **Microsoft Azure** home page. Under **Azure services**, click **Azure OpenAI**. In the left-side menu, under **Azure AI Services**, select **Azure OpenAI**. Select a resource from the **Name** column. On the resource page, go to **Resource Management > Model deployments**. On the **Model deployments** page, click **Manage Deployments**. On the **Deployments** page, copy the desired deployment name from the **Deployment name** column.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
    | API Version                  |         | Enter the [API version](https://learn.microsoft.com/en-us/azure/cognitive-services/openai/reference#rest-api-versioning). The API version to use for this operation in the `YYYY-MM-DD` format. Note that the version may have an extended format, for example, `YYYY-MM-DD-preview`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
    | Custom URL                   |         | This parameter is optional. You can use it to route connections between your clusters and the Azure OpenAI provider through dedicated proxy servers for enhanced security. The custom URL can be any valid endpoint URL and must be complete, including any required query parameters. When you specify a custom URL, the **Resource Name**, **Deployment Name**, and **API Version** fields are ignored.<br />For versions 2025.20 and earlier, when selecting the **API key** connection type, you must use the following pattern: `https://<resource-name>.openai.azure.com/openai/deployments/<deployment-name>/<model-type>?api-version=<api-version>`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
    | API Type                     |         | Appears only when **Chat** is selected as model type. Select one of the following API options for chat models:<ul><li>**Chat Completion** — uses the [Chat Completions API](https://learn.microsoft.com/en-us/azure/ai-foundry/openai/how-to/chatgpt). This option is selected by default.</li><li>**Responses** — uses the [Responses API](https://learn.microsoft.com/en-us/azure/ai-foundry/openai/how-to/responses).</li></ul>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | Use deployment name as model |         | Appears only when **Responses** is selected as the API type. When this parameter is activated, Cognigy sends the value of **Deployment Name** as the model selector in Responses API requests. Microsoft Azure OpenAI deployments require this behavior. When this parameter is deactivated, Cognigy uses the legacy behavior and sends the value of **Model** instead of **Deployment Name**. New connections have this parameter active by default. Existing connections have it inactive to preserve backward compatibility.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |
  </Tab>
</Tabs>

Apply changes. Check if the connection was set up by clicking **Test**. You can switch generative models at any time using the **Model** list without setting up a new connection. For embedding models, the selection is locked because different embedding models can generate incompatible embeddings for [Knowledge AI](/ai/agents/develop/knowledge-ai/overview#embedding-model). To switch an embedding, delete the connection and create a new one.

### Add Models via the API

You can add either a standard or custom model using the [Cognigy API POST /v2.0/largelanguagemodels](https://api-trial.cognigy.ai/openapi#post-/v2.0/largelanguagemodels) request.
Then, test your connection for the created model via the [Cognigy API POST /v2.0/largelanguagemodels//test](https://api-trial.cognigy.ai/openapi#post-/v2.0/largelanguagemodels/-largeLanguageModelId-/test).

### Deprecation of Old Connections for Microsoft Azure OpenAI

As of Cognigy 4.82.0, Azure OpenAI connections that use the `AzureOpenAIProvider` connection type are deprecated. Cognigy supports the `AzureOpenAIProviderV2` connection type instead.

If you have old connections (for example, created in the 4.53 release) to Azure OpenAI,
these connections have an `AzureOpenAIProvider` type and are marked with a **Deprecated** label.
Although these connections can still be active,
we strongly recommend creating a model with the new `AzureOpenAIProviderV2` type,
as old connection types will no longer be available in the future.

Note that for some Microsoft Azure OpenAI models, such as `text-embedding-ada-002` for knowledge search features, you might encounter the following error when an LLM is triggered:
`Error while performing knowledge search. Remote returned error: Search failed: Could not fetch embeddings due to missing API resource name for Azure OpenAI`.
To resolve the issue, recreate the model and the connection so that both are updated to the latest format.

## Apply the Model

To apply a model, follow these steps:

1. In the left-side menu of the Project, go to **Manage > Settings**.
2. Go to the section based on your use case for using a model:
   * **Generative AI Settings**. In the **Generative AI Settings** section, activate **Enable Generative AI Features**. This setting is toggled on by default if you have previously set up the Generative AI credentials.
   * **Knowledge AI Settings**. Use this section if you need to add a model for Knowledge AI. Select a model for the **Knowledge Search** and **Answer Extraction** features. Refer to the [list of standard models](https://docs.cognigy.com/ai/empower/llms/model-support-by-feature/) and find the models that support these features.
3. Navigate to the desired feature and select a model from the list. If there are no models available for the selected feature, the system will automatically select **None**. Save changes.

## More Information

* [Other LLM Operations](/ai/agents/develop/gen-ai-and-llms/other-operations)
* [Overview](/ai/agents/develop/gen-ai-and-llms/llms)
* [All LLM Providers](/ai/agents/develop/gen-ai-and-llms/providers/all-providers)
* [Model support by feature](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature)
* [OpenAI](/ai/agents/develop/gen-ai-and-llms/providers/openai)


# Mistral AI
Source: https://docs.cognigy.com/ai/agents/develop/gen-ai-and-llms/providers/mistral



<a href="/release-notes/2026.11"><Badge>Updated in 2026.11</Badge></a>

To start using a Mistral AI model with Cognigy features, follow these steps:

1. [Add a Model](#add-models)
2. [Apply the Model](#apply-the-model)

## Add Models

You can add a model using one of the following interfaces:

* [GUI](#add-models-via-gui)
* [API](#add-models-via-the-api)

### Add Models via GUI

You can add a model provided by Mistral AI to Cognigy in **Build > LLM**. To add the model, you will need the following parameters:

<Tabs>
  <Tab title="Standard Model">
    <Note>
      The `*-latest` suffix indicates that the model you select points to the latest version of the model. For more information, read the [Mistral AI's models](https://docs.mistral.ai/getting-started/models/compare) documentation.
    </Note>

    | Parameter | Description                                                                                                                                                           |
    | --------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | API Key   | Add the API Key that you [generated](https://docs.mistral.ai/getting-started/quickstart/) under **API > API Keys** in your [Mistral AI](https://mistral.ai/) account. |
  </Tab>

  <Tab title="Custom Model">
    | Parameter  | Description                                                                                                                                                                                      |
    | ---------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Model Type | Select **Chat**.                                                                                                                                                                                 |
    | Model Name | Enter the name of the model that you want to use as a custom model. To find model names, refer to the [Mistral AI models](https://docs.mistral.ai/getting-started/models/compare) documentation. |
    | API Key    | Add the API Key that you [generated](https://docs.mistral.ai/getting-started/quickstart/) under **API > API Keys** in your [Mistral AI](https://mistral.ai/) account.                            |
  </Tab>
</Tabs>

Apply changes. Check if the connection was set up by clicking **Test**. You can switch generative models at any time using the **Model** list without setting up a new connection. For embedding models, the selection is locked because different embedding models can generate incompatible embeddings for [Knowledge AI](/ai/agents/develop/knowledge-ai/overview#embedding-model). To switch an embedding, delete the connection and create a new one.

### Add Models via the API

You can add either a standard or custom model using the [Cognigy API POST /v2.0/largelanguagemodels](https://api-trial.cognigy.ai/openapi#post-/v2.0/largelanguagemodels) request.
Then, test your connection for the created model via the [Cognigy API POST /v2.0/largelanguagemodels//test](https://api-trial.cognigy.ai/openapi#post-/v2.0/largelanguagemodels/-largeLanguageModelId-/test).

## Apply the Model

To apply a model, follow these steps:

1. In the left-side menu of the Project, go to **Manage > Settings**.
2. Go to the section based on your use case for using a model:
   * **Generative AI Settings**. In the **Generative AI Settings** section, activate **Enable Generative AI Features**. This setting is toggled on by default if you have previously set up the Generative AI credentials.
   * **Knowledge AI Settings**. Use this section if you need to add a model for Knowledge AI. Select a model for the **Knowledge Search** and **Answer Extraction** features. Refer to the [list of standard models](https://docs.cognigy.com/ai/empower/llms/model-support-by-feature/) and find the models that support these features.
3. Navigate to the desired feature and select a model from the list. If there are no models available for the selected feature, the system will automatically select **None**. Save changes.

## More Information

* [Other LLM Operations](/ai/agents/develop/gen-ai-and-llms/other-operations)
* [Overview](/ai/agents/develop/gen-ai-and-llms/llms)
* [All LLM Providers](/ai/agents/develop/gen-ai-and-llms/providers/all-providers)
* [Model support by feature](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature)


# OpenAI
Source: https://docs.cognigy.com/ai/agents/develop/gen-ai-and-llms/providers/openai



<a href="/release-notes/2026.11"><Badge>Updated in 2026.11</Badge></a>

<Warning>
  As of Cognigy 2026.6.0, the support for the Completions API for custom LLMs is deprecated. The full removal is planned for Cognigy 2026.24.0.&#x20;
</Warning>

To start using an OpenAI model with Cognigy features, follow these steps:

1. [Add Models](#add-models)
2. [Apply the Model](#apply-the-model)

## Add Models

You can add a model using one of the following interfaces:

* [GUI](#add-models-via-gui)
* [API](#add-models-via-the-api)

### Add Models via GUI

You can add a model provided by OpenAI to Cognigy in **Build > LLM**. To add the model, you will need the following parameters:

<Tabs>
  <Tab title="Standard Model">
    | Parameter | Description                                                                                                                                                                                                                                                                                                                                                       |
    | --------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | apiKey    | Add an API Key from your OpenAI account. You can find this key in the [User settings](https://help.openai.com/en/articles/4936850-where-do-i-find-my-secret-api-key) of your OpenAI account.                                                                                                                                                                      |
    | API Type  | Select the API format for chat models. **Chat Completion** uses the [Chat Completions API](https://platform.openai.com/docs/api-reference/chat). **Responses** uses the [Responses API](https://platform.openai.com/docs/api-reference/responses). The default is **Chat Completion**. This option is only available for chat models that support both API types. |
  </Tab>

  <Tab title="Custom Model">
    | Parameter  | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
    | ---------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Model Type | Select one of the following options:<ul><li>**Chat** — lets you use either the [Chat Completions](https://platform.openai.com/docs/api-reference/chat) or [Responses](https://platform.openai.com/docs/api-reference/responses) API for chat models.</li><li>**Completion** — uses the `https://api.openai.com/v1/completions` API</li><li>**Embedding** — uses the `https://api.openai.com/v1/embeddings` API.</li></ul> For more information, refer to the [OpenAI Text Generation Models](https://platform.openai.com/docs/guides/text-generation) documentation. |
    | Model Name | Enter the name of the model that you want to use as a custom model. To find model names, refer to the [OpenAI models](https://developers.openai.com/api/docs/models) documentation.<br />                                                                                                                                                                                                                                                                                                                                                                            |
    | apiKey     | Add an API Key from your OpenAI account. You can find this key in the [User settings](https://help.openai.com/en/articles/4936850-where-do-i-find-my-secret-api-key) of your OpenAI account.                                                                                                                                                                                                                                                                                                                                                                         |
    | API Type   | Appears only when **Chat** is selected as the model type. Select one of the following API options for chat models:<ul><li>**Chat Completion** — uses the `https://api.openai.com/v1/chat/completions` API. This option is selected by default.</li><li>**Responses** — uses the `https://api.openai.com/v1/responses` API.</li></ul> Make sure the model supports the selected API type.                                                                                                                                                                             |
  </Tab>
</Tabs>

Apply changes. Check if the connection was set up by clicking **Test**. You can switch generative models at any time using the **Model** list without setting up a new connection. For embedding models, the selection is locked because different embedding models can generate incompatible embeddings for [Knowledge AI](/ai/agents/develop/knowledge-ai/overview#embedding-model). To switch an embedding, delete the connection and create a new one.

### Add Models via the API

You can add either a standard or custom model using the [Cognigy API POST /v2.0/largelanguagemodels](https://api-trial.cognigy.ai/openapi#post-/v2.0/largelanguagemodels) request.
Then, test your connection for the created model via the [Cognigy API POST /v2.0/largelanguagemodels//test](https://api-trial.cognigy.ai/openapi#post-/v2.0/largelanguagemodels/-largeLanguageModelId-/test).

## Apply the Model

To apply a model, follow these steps:

1. In the left-side menu of the Project, go to **Manage > Settings**.
2. Go to the section based on your use case for using a model:
   * **Generative AI Settings**. In the **Generative AI Settings** section, activate **Enable Generative AI Features**. This setting is toggled on by default if you have previously set up the Generative AI credentials.
   * **Knowledge AI Settings**. Use this section if you need to add a model for Knowledge AI. Select a model for the **Knowledge Search** and **Answer Extraction** features. Refer to the [list of standard models](https://docs.cognigy.com/ai/empower/llms/model-support-by-feature/) and find the models that support these features.
3. Navigate to the desired feature and select a model from the list. If there are no models available for the selected feature, the system will automatically select **None**. Save changes.

## More Information

* [Other LLM Operations](/ai/agents/develop/gen-ai-and-llms/other-operations)
* [Overview](/ai/agents/develop/gen-ai-and-llms/llms)
* [All LLM Providers](/ai/agents/develop/gen-ai-and-llms/providers/all-providers)
* [Model support by feature](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature)
* [Azure OpenAI](/ai/agents/develop/gen-ai-and-llms/providers/microsoft-azure-openai)


# OpenAI-Compatible LLMs
Source: https://docs.cognigy.com/ai/agents/develop/gen-ai-and-llms/providers/openai-compatible



<a href="/release-notes/2026.11"><Badge>Updated in 2026.11</Badge></a>

<Warning>
  As of Cognigy 2026.6.0, the support for the Completions API for custom LLMs is deprecated. The full removal is planned for Cognigy 2026.24.0.&#x20;
</Warning>

<Warning>
  You are subject to the terms and conditions of the third-party LLM providers
  you are using.
</Warning>

To start using an OpenAI-compatible model with Cognigy features, follow these steps:

1. [Add a Model](#add-models)
2. [Apply the Model](#apply-the-model)

## Add Models

You can add a model using one of the following interfaces:

* [GUI](#add-models-via-gui)
* [API](#add-models-via-the-api)

### Add Models via GUI

You can add a model compatible with OpenAI's API standards to Cognigy in **Build > LLM**. To add the model, you will need the following parameters:

| Parameter                         | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
| --------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| apiKey                            | Add an API Key from the respective provider.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         |
| Model                             | Select **Custom Model**.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
| Model Type                        | Select one of the following options:<ul><li>**Chat** — lets you use either the [Chat Completions](https://platform.openai.com/docs/api-reference/chat) or [Responses](https://platform.openai.com/docs/api-reference/responses) API for chat models.</li><li>**Completion** — uses the `https://api.openai.com/v1/completions` API.</li><li>**Embedding** — uses the `https://api.openai.com/v1/embeddings` API.</li></ul>For more information, refer to the [OpenAI Text Generation Models](https://platform.openai.com/docs/guides/text-generation) documentation. |
| Model Name                        | Enter the name of the model supported by the respective provider. To find model names, refer to the LLM provider's documentation.                                                                                                                                                                                                                                                                                                                                                                                                                                    |
| Base URL                          | Enter the base URL of the LLM provider. For example, `https://api-inference.huggingface.co/v1/` for Hugging Face.                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
| Custom Authentication Header Name | Enter the name of a custom HTTP header for authentication, such as `X-Custom-Auth`. This parameter is helpful when integrating with OpenAI-compatible models or third-party APIs that don't accept the standard `Authorization` header. Instead of sending the API key as `Authorization: Bearer <your-api-key>`, the system will send it using your custom header name. For example, `X-Custom-Auth: <your-api-key>`.                                                                                                                                               |
| API Type                          | Appears only when **Chat** is selected as the model type. Select one of the following API options for chat models:<ul><li>**Chat Completion** — uses the `https://api.openai.com/v1/chat/completions` API. This option is selected by default.</li><li>**Responses** — uses the `https://api.openai.com/v1/responses` API.</li></ul>Make sure that your LLM provider supports the selected API type.                                                                                                                                                                 |

Apply changes. Check if the connection was set up by clicking **Test**. You can switch generative models at any time using the **Model** list without setting up a new connection. For embedding models, the selection is locked because different embedding models can generate incompatible embeddings for [Knowledge AI](/ai/agents/develop/knowledge-ai/overview#embedding-model). To switch an embedding, delete the connection and create a new one.

### Add Models via the API

You can add either a standard or custom model using the [Cognigy API POST /v2.0/largelanguagemodels](https://api-trial.cognigy.ai/openapi#post-/v2.0/largelanguagemodels) request.
Then, test your connection for the created model via the [Cognigy API POST /v2.0/largelanguagemodels//test](https://api-trial.cognigy.ai/openapi#post-/v2.0/largelanguagemodels/-largeLanguageModelId-/test).

## Apply the Model

To apply a model, follow these steps:

1. In the left-side menu of the Project, go to **Manage > Settings**.
2. Go to the section based on your use case for using a model:
   * **Generative AI Settings**. In the **Generative AI Settings** section, activate **Enable Generative AI Features**. This setting is toggled on by default if you have previously set up the Generative AI credentials.
   * **Knowledge AI Settings**. Use this section if you need to add a model for Knowledge AI. Select a model for the **Knowledge Search** and **Answer Extraction** features. Refer to the [list of standard models](https://docs.cognigy.com/ai/empower/llms/model-support-by-feature/) and find the models that support these features.
3. Navigate to the desired feature and select a model from the list. If there are no models available for the selected feature, the system will automatically select **None**. Save changes.

## More Information

* [Other LLM Operations](/ai/agents/develop/gen-ai-and-llms/other-operations)
* [Overview](/ai/agents/develop/gen-ai-and-llms/llms)
* [All LLM Providers](/ai/agents/develop/gen-ai-and-llms/providers/all-providers)
* [Model support by feature](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature)
* [Azure OpenAI](/ai/agents/develop/gen-ai-and-llms/providers/microsoft-azure-openai)


# Activate Knowledge AI
Source: https://docs.cognigy.com/ai/agents/develop/knowledge-ai/activate



<a href="/release-notes/4.80"><Badge>Updated in 4.80</Badge></a>

To activate Knowledge AI, first complete the following steps:

1. [Apply for a license](#apply-for-a-license)
2. [Set quotas](#set-quotas)
3. [Activate the feature flag](#activate-the-feature-flag)
4. [Familiarize yourself with Knowledge AI metrics](#view-metrics)

## Apply for a License

You need to obtain a license from Cognigy to use Knowledge AI. The license allows you to use Knowledge AI with your installation and allocate Knowledge Chunks quotas to your organizations. If you exceed your quotas, you can't create new Knowledge AI resources or edit existing ones.

A user with `admin` rights needs to add the license to your installation. If you are using a SaaS installation with version 4.78 or later, contact [Cognigy technical support](/help/get-help) to add a quota for your organization. For on-premises SaaS installations, contact Customer Success Management or Account Executives at Cognigy.

## Set Quotas

After adding the license, you need to set your Knowledge Chunk quotas based on the number of quotas you purchased within your license. You can set quotas to only one organization or distribute them among multiple organizations.

Only a user with `admin` rights for the Management UI can set quotas via the Management UI or the Cognigy API. If you use a shared SaaS environment, contact [Cognigy technical support](/help/get-help) to configure quotas. For on-premises installations, contact Customer Success Management or Account Executives at Cognigy.

### Set Quotas in Management UI

To set quotas in [Management UI](https://management-ui-v4.cognigy.ai/), select an organization on the **Organisations** page. Then, click <img alt="menu" /> **> Change Organisation Quotas** in the upper-right corner and enter the value in the **Max knowledge chunks** field.

### Set Quotas via API

To set quotas via the [Cognigy API](https://api-trial.cognigy.ai/openapi), follow these steps:

1. Under **AUTHENTICATION > HTTP Basic**, enter your Management UI credentials, then click **SET**.
2. Send the following HTTP request for a new or an existing organization:

<Tabs>
  <Tab title="New Organization">
    Use the [POST /management/v2.0/organisations](https://api-trial.cognigy.ai/openapi#post-/management/v2.0/organisations) request.
    Send the request with a JSON body. In the JSON body, specify a unique name for the new organization and the maximum number of Knowledge Chunks in the `quotaMaxKnowledgeChunks` parameter.

    ```bash theme={null}
    POST /management/v2.0/organisations
    Content-Type: application/json

    {
      "name": "My Organization",
      "quotaMaxKnowledgeChunks": 1000
    }
    ```

    If the response is successful, the server returns a `200` status code with the `organisationId` and set quota.
  </Tab>

  <Tab title="Existing Organization">
    Use the [PATCH /management/v2.0/organisations/](https://api-trial.cognigy.ai/openapi#patch-/management/v2.0/organisations/-organisationId-) request.
    Include the `organisationId` parameter in the path and a JSON body. In the JSON body, specify the maximum number of Knowledge Chunks in the `quotaMaxKnowledgeChunks` parameter. Find `organisationId` in the **ID** column under **Management UI > Organisations**.

    ```bash theme={null}
    PATCH /management/v2.0/organisations/{organisationId}
    Content-Type: application/json

    {
      "quotaMaxKnowledgeChunks": 1000
    }
    ```

    If the response is successful, the server returns a `200` status code with an empty body.
  </Tab>
</Tabs>

### Billable Units

Knowledge Chunks and Knowledge Queries are billable units:

* **Number of Knowledge Chunks** — Knowledge AI calculates the number of Knowledge Chunks you consume when creating or editing Knowledge Sources or Knowledge Chunks. If your quota is exceeded, you can't edit any Knowledge Sources or create new ones.
* **Number of Knowledge Queries** — Knowledge AI counts the number of Queries performed to retrieve relevant Knowledge Chunks. The number of Knowledge Queries is unlimited. However, you will be billed based on the Queries you have performed. Each time the following Nodes are executed they increase the number of performed Knowledge Queries by one:
  * [AI Agent Node](/ai/agents/develop/node-reference/ai/ai-agent)
  * [Search Extract Output Node](/ai/agents/develop/node-reference/other-nodes/search-extract-output)
  * [Copilot: Knowledge Tile Node](/ai/agents/develop/node-reference/agent-copilot/knowledge-tile)

## Activate the Feature Flag

To activate the Knowledge AI feature flag, contact [Cognigy technical support](/help/get-help).

## View Metrics

Familiarize yourself with the [Admin Center](/ai/administer/access/admin-center/overview) page. On this page, you can monitor Knowledge AI metrics to manage the consumption of Knowledge AI resources effectively.

<Frame>
  <img alt="Knowledge AI metrics in the Admin Center showing queries and chunks" />
</Frame>

The Knowledge AI metrics include the following:

1. The number of Knowledge Queries performed within the selected period.
2. The total number of Knowledge Queries performed.
3. The number of Knowledge Chunks currently available in the organization.
4. The maximum number of Knowledge Chunks allocated to the organization.


# Migration from Weaviate to Qdrant: Distance Value Change
Source: https://docs.cognigy.com/ai/agents/develop/knowledge-ai/distance-value-change-after-migration



After the release of Cognigy 4.74.0, all Cognigy installations have been updated to use Qdrant as the vector database for Knowledge AI instead of Weaviate. This migration changes the Knowledge AI calculates the `distance` value in the `topK` object of the Search Extract Output Node.

Weaviate and Qdrant are tools that calculate [similarity](https://en.wikipedia.org/wiki/Cosine_similarity) in different ways:

* Weaviate calculates similarity values within the range of 0 to 2, where 0 represents maximum similarity and 2 represents minimum similarity.
* Qdrant calculates similarity values within the range of -1 to 1, where 1 represents maximum similarity and -1 represents minimum similarity.

These differences lead to a new interpretation of the `distance` value. If you use `distance` to make decisions in the [Flow](/ai/agents/develop/projects-and-flows/overview), consider the new similarity range of -1 to 1.

<Warning>
  The [LLM embedding model](/ai/agents/develop/gen-ai-and-llms/llms) you use defines how precisely Knowledge AI calculates `distance`. The different LLM embedding models can provide considerably different `distance` values. Use the `distance` value for decision-making in Flows with caution.
</Warning>

## Adapting Distance Values: Before and After 4.74.0 Version

If your Flows used `distance` values within the range of 0 to 2 before the release of Cognigy 4.74.0 and you want to keep this approach, change your Flows as follows:

* Replace all `distance` instances with `1 - distance`.

If you are using `distance` for the first time in Cognigy 4.74.0 or later, use the value within the range of -1 to 1.

## More Information

* [Flows](/ai/agents/develop/projects-and-flows/overview)


# Getting started with Knowledge AI
Source: https://docs.cognigy.com/ai/agents/develop/knowledge-ai/getting-started



This guide walks you through the process of creating an AI Agent that uses Knowledge AI to search information in a knowledge database, extract information, and output answers generated by a large language model (LLM). You'll learn how to [create your first Knowledge Store](#configure-knowledge-store), [upload Knowledge Sources](#upload-knowledge), and [test your AI Agent](#test-your-agent) using the Interaction Panel in the Flow editor.

## Prerequisites

* You have [activated Knowledge AI](/ai/agents/develop/knowledge-ai/activate).
* You have created a [Project](/ai/agents/develop/projects-and-flows/projects) in Cognigy.
* You have added a [Large Language Model (LLM) to your Project](/ai/agents/develop/gen-ai-and-llms/providers/all-providers) that [supports Knowledge Search and Answer Extraction](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature). For more details on how to add a model, read the [Generative AI](/ai/agents/develop/gen-ai-and-llms/generative-ai) documentation.

## Set up Knowledge Store

Set up [Knowledge Store](/ai/agents/develop/knowledge-ai/knowledge-store) involves three steps in the Knowledge AI wizard:

1. **Configure Store** — enter the details of your Knowledge Store.
2. **Upload Knowledge** — upload files to create your first [Knowledge Source](/ai/agents/develop/knowledge-ai/knowledge-source/overview).
3. **Configure Flow** — configure an Answer Extraction model and create a simple Knowledge AI Flow.

In your new Project, select Build > Knowledge to open the Knowledge AI wizard. Then, follow the sections below.

### Configure Knowledge Store

Select your embedding model. You will need the following information to configure the Knowledge Store as follows according to your [LLM Provider](/ai/agents/develop/gen-ai-and-llms/providers/all-providers):

<Tabs>
  <Tab title="Microsoft Azure OpenAI">
    | Information                  | Details                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
    | ---------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | **Generative AI Connection** | Your LLM connection.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
    | **Resource Name**            | To find the [resource name](https://learn.microsoft.com/en-us/azure/cognitive-services/openai/how-to/create-resource?pivots=web-portal#create-a-resource) value, go to the **Microsoft Azure** homepage. Under **Azure services**, click **Azure OpenAI**. In the left-side menu, under the **Azure AI Services** section, select **Azure Open AI**. Copy the desired resource name from the **Name** column.                                                                                                                                                                                                                           |
    | **Deployment Name**          | To find the [deployment name](https://learn.microsoft.com/en-us/azure/cognitive-services/openai/how-to/create-resource?pivots=web-portal#deploy-a-model), go to the **Microsoft Azure** home page. Under **Azure services**, click **Azure OpenAI**. In the left-side menu, under the **Azure AI Services** section, select **Azure Open AI**. Select a resource from the **Name** column. On the resource page, go to **Resource Management > Model deployments**. On the **Model deployments** page, click **Manage Deployments**. On the **Deployments** page, copy the desired deployment name from the **Deployment name** column. |
    | **API Version**              | Use the `YYYY-MM-DD` format. Note that the version may have an extended format, for example, `YYYY-MM-DD-preview`. For more details, read [Microsoft Azure documentation](https://learn.microsoft.com/en-us/azure/cognitive-services/openai/reference#rest-api-versioning).                                                                                                                                                                                                                                                                                                                                                             |
    | **Custom URL**               | This parameter is optional. To control the connection between your clusters and the Azure OpenAI provider, you can route connections through dedicated proxy servers, creating an additional layer of security. To do this, enter the URL in the following pattern: `https://<resource-name>.openai.azure.com/openai/deployments/<deployment-name>/completions?api-version=<api-version>`. When a Custom URL is added, the **Resource Name**, **Deployment Name**, and **API Version** fields will be ignored.                                                                                                                          |
  </Tab>

  <Tab title="OpenAI">
    | Information                  | Details              |
    | ---------------------------- | -------------------- |
    | **Generative AI Connection** | Your LLM connection. |
  </Tab>

  <Tab title="Aleph Alpha">
    | Information                  | Details              |
    | ---------------------------- | -------------------- |
    | **Generative AI Connection** | Your LLM connection. |
  </Tab>
</Tabs>

### Upload Knowledge

For this step, download [this Cognigy text (CTXT) file](https://docscognigyassets.blob.core.windows.net/assets/cognigy-sample.ctxt) about xApps.

To use a [CTXT file](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/ctxt) as a Knowledge Source, select **Cognigy CTXT** from the **Type** list and upload the downloaded CTXT file.

### Configure Flow

In this step, proceed as follows:

1. Configure an Answer Extraction model in the **Configure Answer Extraction Model** section. The Answer Extraction model generates the AI Agent's answer based on the Knowledge Search results.
2. Create a simple Flow with a Search Extract Output Node for testing Knowledge AI by clicking **Create Flow**. Cognigy creates a Flow named **Knowledge AI** followed by a number.

## Test your Agent

In **Build > Flows**, select the Flow you created in the Knowledge Store Wizard.

<Frame>
  <img alt="Knowledge AI Flow selected for testing in the Flow list" />
</Frame>

To test Knowledge AI, follow these steps:

1. Open the Interaction Panel by clicking <img alt="interaction-panel" /> in the upper-right corner of the page.
2. Ask the AI Agent `What are xApps?`. The AI Agent will search the Knowledge Store and generate an answer in the chat.

## What's Next?

1. Try creating a [Knowledge Store](/ai/agents/develop/knowledge-ai/knowledge-store) with [Knowledge Sources](/ai/agents/develop/knowledge-ai/knowledge-source/overview) relevant to your business needs.
2. Add Knowledge AI to your Flow with a [Search Extract Output Node](/ai/agents/develop/knowledge-ai/knowledge-search-and-extraction/search-extract-output) and refine the [Node settings](/ai/agents/develop/node-reference/other-nodes/search-extract-output) to cater to your needs.


# Knowledge Chunk
Source: https://docs.cognigy.com/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk



<a href="/release-notes/2025.13"><Badge>Updated in 2025.13</Badge></a>

A *Knowledge Chunk* is a knowledge unit extracted from a [Knowledge Source](/ai/agents/develop/knowledge-ai/knowledge-source/overview). Knowledge Chunks are smaller, self-contained pieces of information that Knowledge AI can process and send to AI Agents.

A Knowledge Chunk can be a single paragraph, a sentence, or even a smaller unit of text from a Knowledge Source. By dividing the content into Knowledge Chunks, Knowledge AI gains better granularity to process information. This way, Knowledge AI analyzes and responds to user queries more efficiently and accurately. Use the [Chunk editor](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk-editor) to edit and manage Knowledge Chunks.

Knowledge Chunks include [Knowledge Chunk text](#knowledge-chunk-text) and [Chunk metadata](#chunk-metadata).

## Restrictions

* The Knowledge Chunk text field can't be empty or contain only whitespace. The same rule applies to the `text` parameter in the [Cognigy KnowledgeChunks API](https://api-trial.cognigy.ai/openapi#tag--KnowledgeChunks).

## Limitations

* You can add up to 1000 Knowledge Chunks per Knowledge Source.
* Knowledge Chunk body and text can have up to 2000 characters together.
* Knowledge Chunk texts can contain bulleted and numbered lists in Markdown format. Other elements, such as links, may not display correctly in the chat conversation.
* Chunk metadata can have up to 1000 characters per Knowledge Chunk.
* You can add up to 20 Knowledge Chunk metadata pairs per Knowledge Chunk.

## Knowledge Chunk Text

[Knowledge AI extracts information from files](/ai/agents/develop/knowledge-ai/knowledge-source/text-extraction/overview) and processes their content into Knowledge Chunk text. You can also create Knowledge Chunks directly in [CTXT files](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/ctxt).

## Chunk Metadata

Chunk metadata contains information related to the Knowledge Chunk. Chunk metadata is optional and can supplement the content of your AI Agent's answers, for example, with links and images.

You can add metadata as JSON key-value pairs to Knowledge Chunks in the Chunk editor or directly to a [CTXT file](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/ctxt). Knowledge Chunks support only numbers, strings, and boolean as Chunk metadata.

You can store Chunk metadata in the [Input object](/ai/agents/develop/ai-agent-memory/input) or [Context object](/ai/agents/develop/ai-agent-memory/context) and give your AI Agents access to Chunk metadata through:

* `input.knowledgeSearch.topK[0].chunkMetaData.{key}` in the [Input object](/ai/agents/develop/ai-agent-memory/input)
* `context.knowledgeSearch.topK[0].chunkMetaData.{key}` in the [Context object](/ai/agents/develop/ai-agent-memory/context)

Where `{key}` is your Chunk metadata key.

You must format Chunk metadata as JSON key-value pairs wrapped with backticks. The keys refer to the information the metadata holds, and the value is the information.

### Chunk Metadata Examples

You can find examples of common Chunk metadata by key in the following table:

| Key     | Description                                                        | Example                                                                                    |
| ------- | ------------------------------------------------------------------ | ------------------------------------------------------------------------------------------ |
| `url`   | Links to external sources or additional reading material.          | `url: https://docs.cognigy.com/ai/build/node-reference/other-nodes/search-extract-output/` |
| `image` | Graphics that illustrate the information in the Knowledge Chunk.   | `image: https://docs.cognigy.com/ai/images/xApp/architecture.png`                          |
| `video` | Multimedia content that helps comprehend the Knowledge Chunk text. | `video: https://www.youtube.com/watch?v=U87psHIQF3w`                                       |

For example, for the `url: https://docs.cognigy.com/ai/build/node-reference/other-nodes/search-extract-output/` Chunk metadata stored in the Input object, where `url` is the key, retrieve this information by using `input.knowledgeSearch.topK[0].chunkMetaData.url`.

## Working with Knowledge Chunks

<Tabs>
  <Tab title="GUI">
    You can view, create, edit, and delete Knowledge Chunks using the [Chunk editor](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk-editor).
  </Tab>

  <Tab title="API">
    You can view, create, edit, and delete Knowledge Chunks using the [Cognigy API](https://api-trial.cognigy.ai/openapi#tag--KnowledgeChunks).
  </Tab>
</Tabs>

## More information

* [Knowledge Source](/ai/agents/develop/knowledge-ai/knowledge-source/overview)
* [Chunk editor](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk-editor)
* [CTXT file](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/ctxt)
* [Input](/ai/agents/develop/ai-agent-memory/input)


# Chunk Editor
Source: https://docs.cognigy.com/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk-editor



<a href="/release-notes/2025.13"><Badge>Updated in 2025.13</Badge></a>

The *Chunk editor* is an interface for managing [Knowledge Chunks](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk) and editing their content.

You can add, edit, rearrange, disable, and delete Knowledge Chunks, as well as add metadata to them. This way, you can ensure the accuracy and relevance of the knowledge you provide to your AI Agents. You can't save empty Knowledge Chunks.

To access the Chunk editor, select a Knowledge Store in **Build > Knowledge**, then a [Knowledge Source](/ai/agents/develop/knowledge-ai/knowledge-source/overview).

## Explore the Knowledge Editor

The Chunk editor includes the Knowledge Chunks in the Knowledge Source as numbered text fields. Clicking the Knowledge Chunks opens the Knowledge Chunk in a new window and displays two areas where you can edit [Knowledge Chunk texts](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk#knowledge-chunk-text) and [Chunk metadata](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk#chunk-metadata).

<Frame>
  <img alt="Knowledge Chunk editor showing numbered text fields for editing Knowledge Chunks" />
</Frame>

<Accordion title="Controls">
  | Button                                       | Description                                                                                                                 |
  | -------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------- |
  | <img alt="download" />                       | You can download the Knowledge Source as a [CTXT file](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/ctxt). |
  | <img alt="more options" />                   | Menu with options to edit, delete, or disable the Knowledge Chunk.                                                          |
  | <img alt="collapse knowledge chunk field" /> | Collapses Knowledge Chunk field.                                                                                            |
  | <img alt="add knowledge chunk" />            | Adds a Knowledge Chunk.                                                                                                     |
</Accordion>

## More Information

* [Knowledge Chunk](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk)
* [Knowledge Source](/ai/agents/develop/knowledge-ai/knowledge-source/overview)


# Knowledge Connectors
Source: https://docs.cognigy.com/ai/agents/develop/knowledge-ai/knowledge-connectors



<a href="/release-notes/2026.7"><Badge>Updated in 2026.7</Badge></a>

*Knowledge Connectors* integrate external services to import content directly into your Flows. This feature provides greater flexibility and control compared to other methods of uploading knowledge, such as uploading multiple PDFs or connecting to a single web page. You can schedule synchronizations to automatically import knowledge, or manually trigger knowledge imports. Knowledge Connectors are developed and delivered as part of [Extension](/ai/for-developers/extensions) packages.

## Key Benefits

* **Flexible Sources**. Ingest data from databases, APIs, document libraries, or Confluence, or crawl websites with Diffbot.
* **Customizable**. Modify prebuilt Knowledge Connectors or create your own from scratch.
* **Centralized Management**. Keep all external knowledge structured and accessible in your Flows.
* **Automated Synchronization**. Schedule regular imports from external sources to keep the AI Agents' knowledge up to date.

## Restrictions

<Accordion title="Knowledge Synchronization Restrictions">
  * Extensions that support knowledge synchronization aren't compatible with Cognigy 2026.4 and earlier versions. Using an Extension with knowledge synchronization in Cognigy 2026.4 and earlier versions can lead to unexpected behavior. If you update an Extension from the Marketplace to a version that supports knowledge synchronization, you can't revert the Extension to an earlier version. If needed, use [Snapshots](/ai/agents/deploy/snapshots) to back up your AI Agent with an Extension compatible with earlier versions of Cognigy.
  * Knowledge Sources imported with Knowledge Connectors that don't support knowledge synchronization can't be synchronized. To use knowledge synchronization, make sure you delete the previously imported Knowledge Sources and re-import them with a Knowledge Connector that supports knowledge synchronization.
  * If you import a Package that includes a Knowledge Connector, the knowledge synchronization is deactivated by default. To activate it, go to the **Connector Configuration** tab in **Build > Knowledge**, click <img alt="edit-assign" /> next to the Knowledge Connector and set the synchronization schedule.
</Accordion>

## Working with Knowledge Connectors

### Install a Knowledge Connector

Select how you want to install a Knowledge Connector:

<Tabs>
  <Tab title="Install from the Marketplace">
    Install a prebuilt Extension that includes a Knowledge Connector from the Marketplace. To do so, follow these steps:

    1. Go to **Manage > Extensions**.
    2. In the **Marketplace** section, select an Extension that includes a Knowledge Connector, for example, the Confluence Extension 6.0.0 or Diffbot Extension 2.0.0, and click **Install**.
  </Tab>

  <Tab title="Modify Existing Knowledge Connectors">
    Clone the Extensions repository and modify the Knowledge Connector of an Extension, for example, the Confluence or Diffbot Extension. To do so, follow these steps:

    1. Clone the [Cognigy Extensions](https://github.com/Cognigy/Extensions) repository from GitHub to your local machine and open it in your code editor.
    2. Go to the [`extensions/confluence`](https://github.com/Cognigy/Extensions/tree/master/extensions/confluence) or [`extensions/diffbot`](https://github.com/Cognigy/Extensions/tree/master/extensions/diffbot) folder. Modify the code as needed. Run `npm install` to install all required dependencies for the Extension.
    3. Run `npx tsc` to compile the TypeScript code. This command converts TypeScript files into JavaScript in the build folder. If the build process doesn't create a tar.gz package automatically, create one manually:
       ```bash theme={null}
       tar -czvf extension-name.tar.gz build/* icon.png README.md package.json package-lock.json
       ```
    4. Go to **Manage > Extensions** in Cognigy and click **+ Upload Extension**.
    5. Select the created package and click **Open**. The installation of the Extension begins.
  </Tab>

  <Tab title="Create from Scratch">
    Create a custom Knowledge Connector from scratch. To do so, follow these steps:

    1. Add the [createKnowledgeConnector()](https://www.npmjs.com/package/@cognigy/extension-tools#knowledge-import) and [upsertKnowledgeConnector()](https://www.npmjs.com/package/@cognigy/extension-tools#knowledge-import) functions to your custom Extension. Use the [Extension template](https://github.com/Cognigy/Extensions/tree/master/docs/example) for structural reference and the [Knowledge Connector example](https://github.com/Cognigy/Extensions/tree/master/docs/example/src/knowledge-connectors). Run `npm install` to install all required dependencies for the Extension.
    2. Run `npx tsc` to compile the TypeScript code. This command converts TypeScript files into JavaScript in the build folder. If the build process doesn't create a tar.gz package automatically, create one manually:
       ```bash theme={null}
       tar -czvf extension-name.tar.gz build/* icon.png README.md package.json package-lock.json
       ```
    3. Go to **Manage > Extensions** in Cognigy and click **+ Upload Extension**.
    4. Select the created package and click **Open**. The installation of the Extension begins.
  </Tab>
</Tabs>

### Set up the Knowledge Connector

1. After the installation, go to **Build > Knowledge** and select the Knowledge Store to which you want to import knowledge.
2. On the **Knowledge Sources** page, click **+ Add Knowledge**.
3. In the **Add Knowledge** window, select a Knowledge Connector from the **Type** list and fill in the required fields for the connection.
4. (Optional) In the **Schedule Sync** section, select one or more weekdays and set the sync time to schedule automatic knowledge synchronization. Click **Save and Sync**.

Wait for the Knowledge Connector to ingest the knowledge from the external source and make it available for your AI Agents. Then, the results are displayed on the following tabs:

* **Knowledge Sources** — displays the imported knowledge files.
* **Connector Configuration** — displays the Knowledge Connector configuration. On this tab, you can edit the configuration by clicking <img alt="edit-assign" />, or trigger a manual synchronization by clicking <img alt="refresh" />.

### Scheduled Synchronizations

When you set up a schedule, you select the weekdays and the time of day for automatic knowledge synchronization (sync). The sync time is displayed in the time zone configured in your [Project settings](/ai/administer/access/project-settings).

If you configure sync, any changes to the external source, including adding and deleting files, are imported into Cognigy on the scheduled day and at the scheduled time.

<Note>
  Imports might be delayed depending on system resources.
</Note>

Changes made directly to Knowledge Sources and Knowledge Chunks in Cognigy aren't exported to the external source. Instead, the Knowledge Connector overwrites these changes, including the metadata, during synchronization. Edit and update files in the external source to keep your knowledge up to date.

### Use Ingested Knowledge in your Flow

You can now use the ingested knowledge in your Flow in the following ways:

* Add a [Search Extract Output Node](/ai/agents/develop/knowledge-ai/knowledge-search-and-extraction/search-extract-output).
* Add an [AI Agent Node with Grounding Knowledge selected](/ai/agents/develop/node-reference/ai/ai-agent).
* Select a Knowledge Store when [creating an AI Agent persona](/ai/agents/develop/persona).
* Add a [Copilot: Knowledge Tile Node](/ai/agents/develop/knowledge-ai/knowledge-search-and-extraction/agent-copilot-knowledge#set-up-the-copilot-knowledge-tile) to use the knowledge in the Agent Copilot dashboard for human agents.

In the Node Editor, select the Knowledge Store for which you set up the Knowledge Connector. Then, test your Flow in the Interaction Panel to see how the AI Agent uses the ingested knowledge to answer user queries.

<Frame>
  <img alt="Knowledge extraction Flow with a Knowledge Store selected in the Node editor" />
</Frame>

## More Information

* [Knowledge Sources](/ai/agents/develop/knowledge-ai/knowledge-source/overview)
* [Overview](/ai/agents/develop/knowledge-ai/overview)


# Copilot: Knowledge Tile Node
Source: https://docs.cognigy.com/ai/agents/develop/knowledge-ai/knowledge-search-and-extraction/agent-copilot-knowledge



<a href="/release-notes/4.72"><Badge>Updated in 4.72</Badge></a>

The *Copilot: Knowledge Tile Node* is an [Agent Copilot Node](/ai/agents/develop/node-reference/agent-copilot/overview) that uses [Knowledge AI](/ai/agents/develop/knowledge-ai/overview) to provide human agents with accurate information about customer questions.

## Key Features

* The Copilot: Knowledge Tile adds an [Agent Copilot widget](/agent-copilot/configure/nodes) in which human agents can use a search bar to access [Knowledge Sources](/ai/agents/develop/knowledge-ai/knowledge-source/overview) and find information to answer questions from customers.
* With the [Copilot: Knowledge Tile Node](/ai/agents/develop/node-reference/agent-copilot/knowledge-tile), you can use the [Search Extract Output Node](/ai/agents/develop/knowledge-ai/knowledge-search-and-extraction/search-extract-output) to retrieve context-aware information, along with additional materials such as references and videos, with [Source metadata](/ai/agents/develop/knowledge-ai/knowledge-source/overview#source-metadata) and [Chunk metadata](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk).

## Prerequisites

* You have configured [Agent Copilot](/agent-copilot/getting-started/chat).
* You have added a [Large Language Model (LLM) to your Project](/ai/agents/develop/gen-ai-and-llms/providers/all-providers) that supports [Knowledge Search and Answer Extracting](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature).
* You have created [Knowledge Sources](/ai/agents/develop/knowledge-ai/knowledge-source/overview).

## Set Up the Copilot: Knowledge Tile

To set up the Copilot: Knowledge Tile, select the Flow where you configured Agent Copilot in **Build > Flows** and add the following Nodes in the Flow editor:

* [If Node](#if-node) — separates the Copilot: Knowledge Tile Node from other Agent Copilot Nodes. This separation prevents the content of other widgets from updating while the AI Agent uses Copilot: Knowledge Tile Node.
* [First Copilot: Knowledge Tile Node](#first-copilot-knowledge-tile-node) — starts the knowledge retrieval.
* [Once Node](#once-node) — guarantees that the Flow executes the Copilot: Knowledge Tile Node once.
* [Second Copilot: Knowledge Tile Node](#second-copilot-knowledge-tile-node) — a duplicate of the first Copilot: Knowledge Tile Node that executes the knowledge retrieval.

### If Node

1. Click **+** under **Start** and add an [If Node](/ai/agents/develop/node-reference/logic/if).

2. In the **Condition** field, enter `input.data._cognigy._agentAssist.payload.tileId === TILE_ID`, where `TILE_ID` is the Tile ID you created in the [Agent Copilot configuration](/agent-copilot/configure/overview).

3. After the **Then** Child Node of the If Node, add the first Copilot: Knowledge Tile Node.
   * **Tile ID** — enter the Tile ID you created in the [Agent Copilot configuration](/agent-copilot/getting-started/chat).
   * **Knowledge Source** — select the Knowledge Source from which you want to retrieve information.

### Once Node

1. Add a [Once Node](/ai/agents/develop/node-reference/logic/once) after the If Node.

### Second Copilot: Knowledge Tile Node

To copy the first Copilot: Knowledge Tile Node and paste it after the **On First Time** Child Node, do the following:

1. Click the first Copilot: Knowledge Tile Node and press **Cmd + C** or **Ctrl + C**.
2. Click the **On First Time** Child Node and press **Cmd + V** or **Ctrl + V**.

For more detailed information about the Copilot: Knowledge Tile Node, read the [Copilot: Knowledge Tile Node reference](/ai/agents/develop/node-reference/agent-copilot/knowledge-tile).

## More Information

* [Getting Started with Agent Copilot](/agent-copilot/getting-started/overview)
* [Knowledge Sources](/ai/agents/develop/knowledge-ai/knowledge-source/overview)
* [Agent Copilot Nodes](/ai/agents/develop/node-reference/agent-copilot/overview)
* [If Node](/ai/agents/develop/node-reference/logic/if)
* [Copilot: Knowledge Tile Node](/ai/agents/develop/node-reference/agent-copilot/knowledge-tile)


# Search Extract Output Node
Source: https://docs.cognigy.com/ai/agents/develop/knowledge-ai/knowledge-search-and-extraction/search-extract-output



<a href="/release-notes/4.84"><Badge>Updated in 4.84</Badge></a>

The Search Extract Output Node uses an embedding model to search [Knowledge Stores](/ai/agents/develop/knowledge-ai/knowledge-store) and a generative AI model to extract relevant information and generate outputs. For example, you can upload manuals and technical documents as [Knowledge Sources](/ai/agents/develop/knowledge-ai/knowledge-source/overview) to create an AI Agent that helps customers with technical issues.

## Prerequisites

* For the Search only mode, you need access to an embedding model that [supports Knowledge Search](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature).
* For the Search & Extract & Output and Search & Extract modes, you need access to both an embedding model and a generative AI model to support [Knowledge Search and Answer Extraction](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature).

## Modes

The Search Extract Output Node has the following modes:

* **Search & Extract & Output** — With this mode, the AI Agent searches Knowledge Sources, extracts answers, and outputs the results automatically as text or adaptive card.
* **Search & Extract** — With this mode, the AI Agent searches Knowledge Sources and extracts answers. You have the flexibility to use these answers in other Nodes or Extension Nodes.
* **Search only** — With this mode, the AI Agent searches Knowledge Sources. You have the flexibility to use the information found.

For more detailed information about the Search Extract Output Node, read the [Node reference](/ai/agents/develop/node-reference/other-nodes/search-extract-output).


# Knowledge Source Tags
Source: https://docs.cognigy.com/ai/agents/develop/knowledge-ai/knowledge-source/knowledge-source-tags



<a href="/release-notes/2026.3"><Badge>Updated in 2026.3</Badge></a>

*Knowledge Source Tags*, also known as Source Tags, are labels that you can add to your [Knowledge Sources](/ai/agents/develop/knowledge-ai/knowledge-source/overview) to improve how your AI Agents reply to your customers. With Source Tags, Knowledge AI's search engine can focus on specific Knowledge Sources and provide AI Agents targeted results.

Source Tags have the following benefits:

* **Increased search accuracy**. Source Tags direct the search engine to specific Knowledge Sources, ensuring that only the most relevant results are included, which is useful with a large number of Knowledge Sources.
* **Flexible tag filtering**. Using `AND` and `OR` operators for Source Tags refine searches and improve results.
* **Faster search performance**. Limiting the queried Knowledge Sources speeds up Knowledge Queries.
* **Enhanced content organization**. Categorizing Knowledge Sources improves information management and retrieval.

## Prerequisites

* You have created a [Knowledge Store](/ai/agents/develop/knowledge-ai/knowledge-store).
* You have created a [Knowledge Source](/ai/agents/develop/knowledge-ai/knowledge-source/overview).

## Limitations

* You can add up to 50 Source Tags per Knowledge Source.
* You can add up to 5 Source Tags per Search Extract Output Node.

## Naming Guidelines

* Include only lowercase and alpha-numerical (a-z, 0-9) characters, underscores, and hyphens.
* You can include [CognigyScript](/ai/platform-features/cognigyscript) in Source Tag names. In this case, make sure the CognigyScript input is lowercase with the `.toLowerCase()` method, for example, `{{input.text.toLowerCase()}}`. Using CognigyScript as a Source Tag name allows your AI Agent to filter Knowledge Sources dynamically according to the user input.

## Working with Source Tags

To configure Source Tags, you need to first add them to a Knowledge Source in **Build > Knowledge** or directly to the [CTXT file](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/ctxt). Then, you need to add the Source Tag to the [Search Extract Output Node](/ai/agents/develop/knowledge-ai/knowledge-search-and-extraction/search-extract-output) in the [Flow editor](/ai/agents/develop/projects-and-flows/editor#exploring-the-flow-editor). You can test the knowledge query results in the [Interaction Panel](/ai/agents/test/interaction-panel/overview).

## Examples

To reproduce the results presented later in this section, follow these steps:

1. Create a new Knowledge Store.

2. Upload the following files as Knowledge Sources:

   * [CTXT](https://docscognigyassets.blob.core.windows.net/assets/cognigy-voice-copilot-endpoint-sample.ctxt) — the Source Tags are already in the CTXT file.
   * [DOCX](https://docscognigyassets.blob.core.windows.net/assets/cognigy-voice-gateway-endpoint-sample.docx) — add the following Source Tags: `endpoint`, `webchat`, and `text-based`.
   * [PDF](https://docscognigyassets.blob.core.windows.net/assets/cognigy-webchat-v3-endpoint-sample.pdf) — add the following Source Tags: `endpoint`, `voice`, and `voice-gateway`.

   After uploading the files, the Knowledge Sources have the following Source Tags:

   | Knowledge Sources | Source Tags in                 |
   | ----------------- | ------------------------------ |
   | CTXT              | endpoint, voice, copilot       |
   | DOCX              | endpoint, voice, voice-gateway |
   | PDF               | endpoint, webchat, text-based  |

3. Navigate to the Flow and add the Search Extract Output Node. Test the following examples to retrieve information based on the Source Tags you add to the Search Extract Output Node. To track how the output changes with different Source Tag combinations, use the input `Could you share a list of native endpoints for Cognigy?` for each Knowledge Source in the Interaction Panel. To check the output results in the Interaction Panel, use the `input.knowledgeSearch.topK` Input object in **INFO > Input**.

### Example 1

Source Tags in the Search Extract Output Node: `endpoint`

| Operator | Expected Result |
| -------- | --------------- |
| AND      | All files       |
| OR       | All files       |

Top-3 Key Results for:

<Accordion title="Both Operators">
  ```json theme={null}
  "knowledgeSearch": {
    "topK": [
      {
        "text": "Cognigy Voice Copilot Endpoint Cognigy Voice Copilot Endpoint is a native Endpoint designed for voice-based applications within Cognigy. The Endpoint manages audio stream transcriptions and integrates seamlessly with Agent Copilot Flows. This Endpoint splits the audio into user and agent inputs, ensuring accurate updates to the Agent Copilot workspace.",
        "distance": 0.85143244,
        "sourceMetaData": {
          "sourceName": "cognigy-voice-copilot-endpoint.ctxt",
          "sourceType": "ctxt",
          "title": "Cognigy Voice Copilot Endpoint",
          "url": "https://www.cognigy.com"
        },
        "chunkMetaData": {
          "image": "https://docs.cognigy.com/_assets/ai/deploy/endpoint-reference/voice-copilot.png",
          "url": "https://docs.cognigy.com/ai/deploy/endpoint-reference/voice-copilot/"
        }
      },
      {
        "text": "Cognigy Webchat v3 Endpoint is a native Endpoint within the Cognigy platform designed to integrate AI Agents directly into your website. This connection allows for seamless interaction between users and AI Agents via the Webchat v3 interface.",
        "distance": 0.85023737,
        "sourceMetaData": {
          "sourceName": "cognigy-webchat-v3-endpoint.pdf",
          "sourceType": "pdf"
        },
        "chunkMetaData": {
          "firstPage": "1",
          "lastPage": "1"
        }
      },
      {
        "text": " Cognigy Voice Gateway Endpoint Cognigy Voice Gateway Endpoint is a native Endpoint within the Cognigy platform that enables AI Agents to interact with users through voice communication. This native or first-party integration connects AI Agents to contact centers or phone numbers, allowing users to speak with the agents instead of writing.",
        "distance": 0.84896475,
        "sourceMetaData": {
          "sourceName": "cognigy-voice-gateway-endpoint.docx",
          "sourceType": "docx"
        },
        "chunkMetaData": {
          "firstPage": "1",
          "lastPage": "1"
        }
      }
    ]
  }
  ```
</Accordion>

### Example 2

Source Tags in the Search Extract Output Node: `endpoint`, `voice`

| Operator | Expected Result |
| -------- | --------------- |
| AND      | CTXT and PDF    |
| OR       | All files       |

Top-3 Key Results for:

<AccordionGroup>
  <Accordion title="AND">
    ```json theme={null}
      "knowledgeSearch": {
        "topK": [
          {
            "text": "Cognigy Voice Copilot Endpoint Cognigy Voice Copilot Endpoint is a native Endpoint designed for voice-based applications within Cognigy. The Endpoint manages audio stream transcriptions and integrates seamlessly with Agent Copilot Flows. This Endpoint splits the audio into user and agent inputs, ensuring accurate updates to the Agent Copilot workspace.",
            "distance": 0.85143244,
            "sourceMetaData": {
              "sourceName": "cognigy-voice-copilot-endpoint.ctxt",
              "sourceType": "ctxt",
              "title": "Cognigy Voice Copilot Endpoint",
              "url": "https://www.cognigy.com"
            },
            "chunkMetaData": {
              "image": "https://docs.cognigy.com/_assets/ai/deploy/endpoint-reference/voice-copilot.png",
              "url": "https://docs.cognigy.com/ai/deploy/endpoint-reference/voice-copilot/"
            }
          },
          {
            "text": " Cognigy Voice Gateway Endpoint Cognigy Voice Gateway Endpoint is a native Endpoint within the Cognigy platform that enables AI Agents to interact with users through voice communication. This native or first-party integration connects AI Agents to contact centers or phone numbers, allowing users to speak with the agents instead of writing.",
            "distance": 0.84896475,
            "sourceMetaData": {
              "sourceName": "cognigy-voice-gateway-endpoint.docx",
              "sourceType": "docx"
            },
            "chunkMetaData": {
              "firstPage": "1",
              "lastPage": "1"
            }
          },
          {
            "text": "Key Features - Automatic Transcription Handling: Receives and processes transcriptions, identifying user and session IDs for precise session management. - Native Integration: As a native Endpoint, it is built directly into the Cognigy platform, offering smooth functionality and compatibility with supported handover providers.",
            "distance": 0.8258752,
            "sourceMetaData": {
              "sourceName": "cognigy-voice-copilot-endpoint.ctxt",
              "sourceType": "ctxt",
              "title": "Cognigy Voice Copilot Endpoint",
              "url": "https://www.cognigy.com"
            },
            "chunkMetaData": {
              "url": "https://docs.cognigy.com/ai/deploy/endpoint-reference/voice-copilot/"
            }
          }
        ]
      }
    ```
  </Accordion>

  <Accordion title="OR">
    ```json theme={null}
      "knowledgeSearch": {
        "topK": [
          {
            "text": "Cognigy Voice Copilot Endpoint Cognigy Voice Copilot Endpoint is a native Endpoint designed for voice-based applications within Cognigy. The Endpoint manages audio stream transcriptions and integrates seamlessly with Agent Copilot Flows. This Endpoint splits the audio into user and agent inputs, ensuring accurate updates to the Agent Copilot workspace.",
            "distance": 0.85143244,
            "sourceMetaData": {
              "sourceName": "cognigy-voice-copilot-endpoint.ctxt",
              "sourceType": "ctxt",
              "title": "Cognigy Voice Copilot Endpoint",
              "url": "https://www.cognigy.com"
            },
            "chunkMetaData": {
              "image": "https://docs.cognigy.com/_assets/ai/deploy/endpoint-reference/voice-copilot.png",
              "url": "https://docs.cognigy.com/ai/deploy/endpoint-reference/voice-copilot/"
            }
          },
          {
            "text": "Cognigy Webchat v3 Endpoint is a native Endpoint within the Cognigy platform designed to integrate AI Agents directly into your website. This connection allows for seamless interaction between users and AI Agents via the Webchat v3 interface.",
            "distance": 0.85023737,
            "sourceMetaData": {
              "sourceName": "cognigy-webchat-v3-endpoint.pdf",
              "sourceType": "pdf"
            },
            "chunkMetaData": {
              "firstPage": "1",
              "lastPage": "1"
            }
          },
          {
            "text": " Cognigy Voice Gateway Endpoint Cognigy Voice Gateway Endpoint is a native Endpoint within the Cognigy platform that enables AI Agents to interact with users through voice communication. This native or first-party integration connects AI Agents to contact centers or phone numbers, allowing users to speak with the agents instead of writing.",
            "distance": 0.84896475,
            "sourceMetaData": {
              "sourceName": "cognigy-voice-gateway-endpoint.docx",
              "sourceType": "docx"
            },
            "chunkMetaData": {
              "firstPage": "1",
              "lastPage": "1"
            }
          }
        ]
      }
    ```
  </Accordion>
</AccordionGroup>

### Example 3

Source Tags in the Search Extract Output Node: `voice`, `text-base`

| Operator | Expected Result           |
| -------- | ------------------------- |
| AND      | No Knowledge Chunks found |
| OR       | All Knowledge Chunks      |

Top-3 Key Results for:

<AccordionGroup>
  <Accordion title="AND">
    ```json theme={null}
    "knowledgeSearch": {
      "topK": []
    }
    ```
  </Accordion>

  <Accordion title="OR">
    ```json theme={null}
      "knowledgeSearch": {
        "topK": [
          {
            "text": "Cognigy Webchat v3 Endpoint is a native Endpoint within the Cognigy platform designed to integrate AI Agents directly into your website. This connection allows for seamless interaction between users and AI Agents via the Webchat v3 interface.",
            "distance": 0.8642316,
            "sourceMetaData": {
              "sourceName": "cognigy-webchat-v3-endpoint.pdf",
              "sourceType": "pdf"
            },
            "chunkMetaData": {
              "firstPage": "1",
              "lastPage": "1"
            }
          },
          {
            "text": " Cognigy Voice Gateway Endpoint Cognigy Voice Gateway Endpoint is a native Endpoint within the Cognigy platform that enables AI Agents to interact with users through voice communication. This native or first-party integration connects AI Agents to contact centers or phone numbers, allowing users to speak with the agents instead of writing.",
            "distance": 0.85569,
            "sourceMetaData": {
              "sourceName": "cognigy-voice-gateway-endpoint.docx",
              "sourceType": "docx"
            },
            "chunkMetaData": {
              "firstPage": "1",
              "lastPage": "1"
            }
          },
          {
            "text": "Cognigy Voice Copilot Endpoint Cognigy Voice Copilot Endpoint is a native Endpoint designed for voice-based applications within Cognigy. The Endpoint manages audio stream transcriptions and integrates seamlessly with Agent Copilot Flows. This Endpoint splits the audio into user and agent inputs, ensuring accurate updates to the Agent Copilot workspace.",
            "distance": 0.8550185,
            "sourceMetaData": {
              "sourceName": "cognigy-voice-copilot-endpoint.ctxt",
              "sourceType": "ctxt",
              "title": "Cognigy Voice Copilot Endpoint",
              "url": "https://www.cognigy.com"
            },
            "chunkMetaData": {
              "image": "https://docs.cognigy.com/_assets/ai/deploy/endpoint-reference/voice-copilot.png",
              "url": "https://docs.cognigy.com/ai/deploy/endpoint-reference/voice-copilot/"
            }
          }
        ]
      }
    ```
  </Accordion>
</AccordionGroup>

## More Information

* [Knowledge Source](/ai/agents/develop/knowledge-ai/knowledge-source/overview)
* [Knowledge Store](/ai/agents/develop/knowledge-ai/knowledge-store)
* [CTXT file](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/ctxt)
* [Search Extract Output Node](/ai/agents/develop/node-reference/other-nodes/search-extract-output)
* [Interaction Panel](/ai/agents/test/interaction-panel/overview)


# Knowledge Source
Source: https://docs.cognigy.com/ai/agents/develop/knowledge-ai/knowledge-source/overview



<a href="/release-notes/2025.21"><Badge>Updated in 2025.21</Badge></a>

A *Knowledge Source* is structured, easily accessible information that Knowledge AI extracts from documents or external systems and provides to AI Agents.
You can create Knowledge Sources from content types, such as PDF, CTXT, web pages, or via [*Knowledge Connectors*](/ai/agents/develop/knowledge-ai/knowledge-connectors) that link to external knowledge bases. Knowledge AI breaks content into smaller units called [*Knowledge Chunks*](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk). You can also add [*Source metadata*](#source-metadata) and [*Knowledge Source Tags*](/ai/agents/develop/knowledge-ai/knowledge-source/knowledge-source-tags) to improve how AI Agents use the content. You can create any number of Knowledge Sources per Knowledge Store.

## Limitations

* You can upload files with up to 10 MB as Knowledge Sources.
* You can add up to 20 Source metadata pairs per Knowledge Source.
* Source metadata pairs can have up to 1000 characters.

## Knowledge Types

Knowledge AI organizes content using content types and Knowledge Connectors.
While both provide information for your AI Agents, they are independent and use separate parsers.

<AccordionGroup>
  <Accordion title="Content Types">
    Content types define the formats of content that can be used as Knowledge Sources. Supported formats include:

    * [Cognigy Text Format](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/ctxt) (CTXT, recommended)
    * Files in the following formats:
      * [PDF](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/pdf)
      * TXT
      * DOCX
      * PPTX
      * JPEG
      * JPG
      * PNG
      * BMP
      * HEIF
      * TIFF
    * [Web Pages](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/web-page)

    Depending on the format of the file you use, you need to configure a specific [parser](/ai/agents/develop/knowledge-ai/knowledge-source/text-extraction/overview) to convert the content into Knowledge Sources.
  </Accordion>

  <Accordion title="Knowledge Connectors">
    [Knowledge Connectors](/ai/agents/develop/knowledge-ai/knowledge-connectors) link your Knowledge Store to external systems, such as CRMs, help centers, or databases. They operate independently of content types and include their own parsers to fetch and transform data into usable knowledge for AI Agents.
    You can use prebuilt Knowledge Connectors from the Marketplace, such as Confluence and Diffbot, or create custom ones using the [Extension framework](/ai/for-developers/extensions).
  </Accordion>
</AccordionGroup>

## Source Metadata

*Source metadata* contains information about the Knowledge Source and is formatted as JSON key-value pairs. The keys refer to the information the Source metadata holds and the value is the information, for example, `publication: Elsevier`.

You can store Source metadata in the [Input object](/ai/agents/develop/ai-agent-memory/input) or [Context object](/ai/agents/develop/ai-agent-memory/context) and give your AI Agents access to the Source metadata through:

* `input.knowledgeSearch.topK[0].sourceMetadata` in the Input object
* `context.knowledgeSearch.topK[0].sourceMetadata` in the Context object

This way, AI Agents can provide relevant information about the source of their replies to users.

### Examples

You can find examples of Source metadata in the following table:

| Key     | Description                                                                                                                                                                                                        | Mandatory | Example                                     |
| ------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | --------- | ------------------------------------------- |
| `title` | This key defines the document's name. If a user requests the information source, the AI Agent can provide the document's title in the response message.                                                            | no        | `title: Cognigy xApps FAQs`                 |
| `url`   | This key defines a URL as the document's source. If a user requests the information source, the AI Agent can provide the link to the document's source in the response message.                                    | no        | <span>`url: https://www.cognigy.com`</span> |
| `tags`  | This key defines [Knowledge Source Tags](/ai/agents/develop/knowledge-ai/knowledge-source/knowledge-source-tags). To add more than Knowledge Source Tag, use an array. You can add up to 10 Knowledge Source Tags. | no        | `[cognigy, xapps, architecture]`            |

You can add any Source metadata to a Knowledge Source and give AI Agents access to the Source metadata through `input.knowledgeSearch.topK[0].sourceMetaData.{key}`. For example, if you add the Source metadata `author: Cognigy`, where `author` is the key, and store it in the Input object, retrieve this information by using `input.knowledgeSearch.topK[0].sourceMetaData.author`.

## Working with Knowledge Sources

<Tabs>
  <Tab title="GUI">
    You can view, create, edit, and delete Knowledge Sources in **Build >
    Knowledge**. Also, you can copy Knowledge Sources' Reference ID, disable and
    export Knowledge Sources, and [download them as CTXT
    file](#download-knowledge-sources-as-ctxt).
  </Tab>

  <Tab title="API">
    You can view, create, edit, and delete Knowledge Sources using the
    [Cognigy API](https://api-trial.cognigy.ai/openapi#tag--KnowledgeSources).
    You can
    also use the API to upload files as Knowledge Source.
  </Tab>

  <Tab title="CLI">
    You can ingest documents to create Knowledge Sources and view the resulting
    number of tokens using the [Cognigy CLI](https://github.com/Cognigy/Cognigy-CLI/blob/main/KNOWLEDGE-AI-README.md)
  </Tab>
</Tabs>

## Download Knowledge Sources as CTXT

You can download Knowledge Sources as a CTXT file, including Knowledge Sources created from web pages and other file formats. Downloading Knowledge Sources as CTXT files offers more flexibility in managing Knowledge Sources. With a CTXT file, you can:

* Back up Knowledge Sources more easily after you have edited the Knowledge Chunks.
* Use Knowledge Sources across Knowledge Stores with different embedding vector dimensions.
* Edit downloaded Knowledge Sources more efficiently with external tools, such as text editors, with features like search and replace, bulk editing, and integration of AI tools.

## More information

* [Knowledge Store](/ai/agents/develop/knowledge-ai/knowledge-store)
* [Knowledge Chunk](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk)
* [Knowledge Source Tags](/ai/agents/develop/knowledge-ai/knowledge-source/knowledge-source-tags)
* [Chunk editor](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk-editor)
* [Input](/ai/agents/develop/ai-agent-memory/input)


# Advanced Parser
Source: https://docs.cognigy.com/ai/agents/develop/knowledge-ai/knowledge-source/text-extraction/advanced-parser



<a href="/release-notes/4.80"><Badge>Updated in 4.80</Badge></a>

<Note>
  By using the Advanced Parser, you are accepting the [Cognigy Additional Data Privacy Terms](https://www.cognigy.com/additional-privacy-terms).
</Note>

The *Advanced Parser* is Knowledge AI's default content parser. When you upload a file as a [Knowledge Source](/ai/agents/develop/knowledge-ai/knowledge-source/overview), the Advanced Parser splits the file content semantically into Knowledge Chunks and converts them into Markdown format. This parsing method results in more accurate answers compared to the [Basic Parser](/ai/agents/develop/knowledge-ai/knowledge-source/text-extraction/basic-parser).

## Key Features

* **Improved quality of responses**. Breaking down the text into [Knowledge Chunks](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk) allows AI Agents to understand the context and provide more accurate responses.
* **Recognition of a wide range of text elements**. The Advanced Parser recognizes plain text and more complex text elements, such as tables, titles, and footers as well as images, by using [Optical character recognition (OCR)](https://en.wikipedia.org/wiki/Optical_character_recognition).
* **Effective use of Markdown**. The Advanced Parser converts the text into Markdown format and in Knowledge Chunks. AI Agents can process Markdown and recognize text elements, for example, tables and images, and distinguish them from plain text. This approach gives AI Agents more context and helps them better understand how information is organized in the text.
* **Improved ability to reference the source**. Adding page numbers of the source file to Knowledge Sources as [Chunk metadata](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk#chunk-metadata) helps track an information source in large documents.

## Prerequisites

* Cognigy 4.71 or later.

## Limitations

* The parser may not always preserve hyperlinks correctly when processing PDF files.

## Supported Formats

| Cognigy versions | Formats                                                                               |
| ---------------- | ------------------------------------------------------------------------------------- |
| 4.80 or later    | PDF, DOCX, PPTX, TXT<sup>[1](#footnote1)</sup>, JPEG, JPG, PNG, BMP, HEIF, TIFF files |
| 4.79             | PDF, DOCX, PPTX, and TXT<sup>[1](#footnote1)</sup> files                              |
| 4.71-4.78        | PDF and DOCX files                                                                    |
| 4.71 or earlier  | PDF files                                                                             |

## Availability

The Advanced Parser is available in all Cognigy installations. For the [on-premises installation](#activate-the-advanced-parser-for-on-premises), you need to activate the Advanced Parser.

### Activate the Advanced Parser for On-Premises

Activating the Advanced Parser varies depending on which version you use:

* [Cognigy 4.80 or later](#cognigyai-480-or-later)
* [Cognigy 4.79 or earlier](#cognigyai-479-or-earlier)

#### Cognigy 4.80 or Later

To activate the Advanced Parser, add the following parameters to the `knowledgeSearch` section in the `cognigy-ai-values.yaml` file:

```yaml theme={null}
knowledgeSearch:
  enabled: true

  globalAzureDocumentIntelligenceConfig:
    enabled: true
    apiKey: API_KEY
    endpointUrl: ENDPOINT_URL
```

Replace `API_KEY` with your Azure AI Document Intelligence API key and `ENDPOINT_URL` with your Azure AI Document Intelligence endpoint URL.

#### Cognigy 4.79 or Earlier

To activate the Advanced Parser for all or specific organizations in your Cognigy installation, set the following environment variables in the ` cognigy-ai-values.yaml` file:

* `FEATURE_ENABLE_AZURE_DOCUMENT_INTELLIGENCE_ORG_WHITELIST`:
  * For all organizations, set to `*`.
  * For specific organizations, enter the organization IDs separated by a comma, for example, `63bda3d4tlp3c95977bb8604,63babf6e92asd791923e17b7,670k1273fx492448dc288b36`.
* `AZURE_DOCUMENT_INTELLIGENCE_ENDPOINT`

***

<sup>1</sup>: TXT files are internally handled by the Basic Parser.


# Azure AI Document Intelligence
Source: https://docs.cognigy.com/ai/agents/develop/knowledge-ai/knowledge-source/text-extraction/azure-ai



<a href="/release-notes/4.80"><Badge>Updated in 4.80</Badge></a>

<Note>
  * By using Azure AI Document Intelligence as a content parser, you agree to the [Cognigy Additional Data Privacy Terms](https://www.cognigy.com/additional-privacy-terms).
  * Azure AI Document Intelligence stores data and results in Azure Storage for 24 hours after submitting an analyze request. To learn more about data, privacy, and security for Document Intelligence, read the [Azure AI Document Intelligence](https://learn.microsoft.com/en-us/legal/cognitive-services/document-intelligence/data-privacy-security) documentation.
</Note>

*[Azure AI Document Intelligence](https://learn.microsoft.com/en-us/azure/ai-services/document-intelligence/)* is an external content parser that you can use in Knowledge AI for text extraction. Azure AI Document Intelligence splits texts semantically into Knowledge Chunks, resulting in more accurate answers compared to the [Basic Parser](/ai/agents/develop/knowledge-ai/knowledge-source/text-extraction/basic-parser).

When you upload a file as a [Knowledge Source](/ai/agents/develop/knowledge-ai/knowledge-source/overview), Knowledge AI sends a request to Azure AI Document Intelligence to extract text from the file. After the text extraction, Knowledge AI retrieves the text, converts it into Markdown format, and splits it into Knowledge Chunk.

## Key Features

* **Improved quality of responses**. Breaking down the text into [Knowledge Chunks](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk) allows AI Agents to understand the context and provide more accurate responses.
* **Recognition of a wide range of text elements**. Azure AI Document Intelligence recognizes plain text and more complex text elements, such as tables, titles, and footers as well as images, by using [Optical character recognition (OCR)](https://en.wikipedia.org/wiki/Optical_character_recognition).
* **Effective use of Markdown**. Azure AI Document Intelligence converts the text into Markdown format. AI Agents can process Markdown and recognize text elements, for example, tables and images, and distinguish them from plain text. This approach gives AI Agents more context and helps them better understand how information is organized in the text.
* **Improved ability to reference the source**. Adding page numbers of the source file to Knowledge Sources as [Chunk metadata](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk#chunk-metadata) helps track an information source in large documents.

## Prerequisite

* You have access to an [Azure AI Document Intelligence account](https://azure.microsoft.com/en-us/products/ai-services/ai-document-intelligence), and the account must be hosted in one of the following [regions](https://datacenters.microsoft.com/globe/explore):

  * East US
  * West US2
  * West Europe

## Supported Formats

| Cognigy versions | Formats                                                         |
| ---------------- | --------------------------------------------------------------- |
| 4.80 or later    | PDF, DOCX, PPTX, TXT, JPEG, JPG, PNG, BMP, HEIF, and TIFF files |

### Chunk Metadata

<Warning>
  Supported formats: PDF, PPTX, and TIFF files.
</Warning>

Azure AI Document Intelligence recognizes page numbers and adds them to the Knowledge Chunks as Chunk metadata. In the Chunk editor, you can see the page numbers as Chunk metadata. If the page numbers appear in the Chunk editor, it indicates that the text extraction was successful. If they don't, it suggests an error, for example, a failure to access Azure AI Document Intelligence.

In the following example, the Chunk metadata indicates that the Knowledge Chunk text starts on page 1 of the source file and ends on page 2.

```txt theme={null}
{
  "firstPage": "1",
  "lastPage": "2"
}
```

## Set Up Azure AI Document Intelligence

By default, Knowledge AI uses the Advanced Parser for text extraction. Set up the Azure AI Document Intelligence in **Manage > Settings > Knowledge AI Settings**:

<Accordion title="Set Up Azure AI Document Intelligence">
  <Tabs>
    <Tab title="You don't have an Azure AI Document Intelligence Connection">
      1. In the **Document Processing** section, select **Azure AI Document Intelligence** from the **Content Parser** list.
      2. Click **+** next to the **Connection to Azure AI Document Intelligence** field.
      3. Fill in the following fields:
         * **Connection name** — enter a unique name for your connection.
         * **API Key** — enter an API key for the Azure AI Document Intelligence service. To copy the key, go to the Azure portal. On the left-side menu, select **Resource Management > Key and Endpoint**. Copy the key from the **Key** field. For more information, read the [Azure AI Document Intelligence](https://learn.microsoft.com/en-us/azure/ai-services/document-intelligence/create-document-intelligence-resource?view=doc-intel-4.0.0#get-endpoint-url-and-keys) documentation.
         * **Endpoint URL** — add an Endpoint URL from your Document Intelligence resource. To copy the URL, go to the Azure portal. On the left-side menu, select **Resource Management > Key and Endpoint**. Copy the endpoint URL from the **Endpoint** field. For more information, read the [Azure AI Document Intelligence](https://learn.microsoft.com/en-us/azure/ai-services/document-intelligence/create-document-intelligence-resource?view=doc-intel-4.0.0#get-endpoint-url-and-keys) documentation.
    </Tab>

    <Tab title="You have an Azure AI Document Intelligence Connection">
      1. In the **Document Processing** section, select **Azure AI Document Intelligence** from the **Content Parser** list.
      2. Select a connection from the **Connection to Azure AI Document Intelligence** field.
    </Tab>
  </Tabs>
</Accordion>

## More Information

* [Text Extraction with Content Parsers](/ai/agents/develop/knowledge-ai/knowledge-source/text-extraction/overview)
* [Advanced Parser](/ai/agents/develop/knowledge-ai/knowledge-source/text-extraction/advanced-parser)
* [Basic Parser](/ai/agents/develop/knowledge-ai/knowledge-source/text-extraction/basic-parser)
* [Knowledge Source](/ai/agents/develop/knowledge-ai/knowledge-source/overview)


# Basic Parser
Source: https://docs.cognigy.com/ai/agents/develop/knowledge-ai/knowledge-source/text-extraction/basic-parser



The *Basic Parser* is a content parser that Knowledge AI uses to extract content from files more effectively. The Basic Parser splits the extracted content from a file into [Knowledge Chunks](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk) based on a fixed token length during the creation of a [Knowledge Source](/ai/agents/develop/knowledge-ai/knowledge-source/overview). Based on our research, this parsing method works best with the **Top K** value set to 5 in the [Search Extract Output](/ai/agents/develop/node-reference/other-nodes/search-extract-output) Node.

## Limitations

* The parser may not always preserve hyperlinks correctly when processing PDF files.

## Supported Formats

| Cognigy versions | Formats                  |
| ---------------- | ------------------------ |
| 4.79 or later    | PDF, DOCX, PPTX, and TXT |
| 4.78 and earlier | PDF, DOCX, and TXT       |

## Availability

* Available in all installations.

## Set Up the Basic Parser

By default, Knowledge AI uses the [Advanced Parser](/ai/agents/develop/knowledge-ai/knowledge-source/text-extraction/advanced-parser) for text extraction. To activate the Basic Parser, select **Basic** from the **Content Parser** list in the **Document Processing** section in **Manage > Settings > Knowledge AI Settings**.

## More Information

* [Text Extraction with Content Parsers](/ai/agents/develop/knowledge-ai/knowledge-source/text-extraction/overview)
* [Text Extraction with the Advanced Parser](/ai/agents/develop/knowledge-ai/knowledge-source/text-extraction/advanced-parser)
* [Text Extraction with Azure AI Document Intelligence](/ai/agents/develop/knowledge-ai/knowledge-source/text-extraction/azure-ai)


# Text Extraction with Content Parsers
Source: https://docs.cognigy.com/ai/agents/develop/knowledge-ai/knowledge-source/text-extraction/overview



Knowledge AI uses content parsers to extract texts from uploaded files in different formats. The content parsers are categorized based on their complexity and the techniques they use.

You can use the following content parsers:

* [Basic Parser](/ai/agents/develop/knowledge-ai/knowledge-source/text-extraction/basic-parser)
* [Advanced Parser (Default)](/ai/agents/develop/knowledge-ai/knowledge-source/text-extraction/advanced-parser)
* [Azure AI Document Intelligence](/ai/agents/develop/knowledge-ai/knowledge-source/text-extraction/azure-ai)

## Content Parsers Overview

The table provides a comparative overview of the content parsers.

| Details       | Basic Parser                                                                                                                                                      | Advanced Parser (Default)                                                                                                                                                                                                              | Azure AI Document Intelligence                                                                                                                                      |
| ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Functionality | Fundamental text extraction.                                                                                                                                      | Advanced text extraction.                                                                                                                                                                                                              | Advanced, cloud-based text extraction.                                                                                                                              |
| Use Cases     | Extracting text from simple documents.                                                                                                                            | Extracting text from complex documents that include tables, images, footers, and more.                                                                                                                                                 | Extracting text from complex documents that include tables, images, footers, and more.                                                                              |
| Advantages    | <ul><li>Fast processing.</li><li>Low resource consumption.</li></ul>                                                                                              | <ul><li>Better accuracy.</li><li>Handles various formats.</li><li>Includes OCR.</li></ul>                                                                                                                                              | <ul><li>Better accuracy.</li><li>Handles various formats.</li><li>Includes OCR.</li><li>Customizable settings on the Azure AI Document Intelligence side.</li></ul> |
| Limitations   | <ul><li>Limited handling of complex documents. </li><li>Less accuracy. </li><li>Limited file format support. </li><li>Lack of customization capability.</li></ul> | Lack of customization capability.                                                                                                                                                                                                      | <ul><li>Dependency on a third-party service. </li><li>Payment for using the service.</li></ul>                                                                      |
| Price         | Free of charge for all environments. Included in the Knowledge AI license.                                                                                        | <ul><li>Free of charge for SaaS installations. Included in the Knowledge AI license.</li><li>For on-premises installations, payment proceeds through Azure AI Document Intelligence in accordance with their pricing policy.</li></ul> | For all installations, payment proceeds through Azure AI Document Intelligence in accordance with their pricing policy.                                             |

## Automatic Fallback to Basic Parsing

If the Advanced Parser or Azure AI Document Intelligence fails, Knowledge AI falls back automatically on the Basic Parser.

## More Information

* [Knowledge Source](/ai/agents/develop/knowledge-ai/knowledge-source/overview)
* [PDF](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/pdf)


# Cognigy Text Format
Source: https://docs.cognigy.com/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/ctxt



<a href="/release-notes/4.63"><Badge>Updated in 4.63</Badge></a>

The *Cognigy Text* format (CTXT) is a file format developed by Cognigy for extracting knowledge with Knowledge AI. You can segment the text in a CTXT file into [Knowledge Chunks](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk) before uploading the file as a [Knowledge Source](/ai/agents/develop/knowledge-ai/knowledge-source/overview).

You can edit CTXT files in text editors, such as [Visual Studio Code](https://code.visualstudio.com/) and [Notepad++](https://notepad-plus-plus.org/downloads/). Text editors improve the visual representation of Markdown syntax and make it easier to read a CTXT file. To make CTXT easier to read, you need to configure the text editor to highlight Markdown syntax:

* **Visual Studio Code** — Access the [Change Language Mode](https://code.visualstudio.com/docs/languages/overview#_change-the-language-for-the-selected-file) section and select Markdown.
* **Notepad++** — Select Markdown as a language using a [User Defined Language file](https://npp-user-manual.org/docs/user-defined-language-system/).

#### Limitations

* [Header](#header) limitations:
  * You can add up to 20 Source metadata pairs per Knowledge Source.
  * Source metadata can have up to 1000 characters.
* [Body](#body) limitations:
  * You can add up to 1000 Knowledge Chunks per CTXT file.
  * You can add up to 20 Chunk metadata pairs per CTXT file.
  * Knowledge Chunk text can have up to 2000 characters.
  * Chunk metadata can have up to 1000 characters.

## Restrictions

* Formatting restrictions:

  * The file extension must be `.ctxt`.
  * You must use the [CTXT structure](#ctxt-structure).
  * You must use Markdown syntax.

* Naming restrictions:

  * The file name can have up to 200 characters.
  * The file name can't start or end with a space.
  * The file name can't contain `/`, `\`, `:`, `*`, `?`, `"`, `<`, `>`, `|`, or `¥`.

* Knowledge Chunk text CTXT containing unclosed HTML tags for block-level elements in the [CTXT body](#body), such as `<div>`, `<table>`, and `<p>`, causes the Markdown parser to include the metadata of the following Knowledge Chunk in the previous Knowledge Chunk. To avoid unintended inclusion, of metadata in the Knowledge Chunk text, ensure you close HTML tags for block-level elements or remove them from the Knowledge Chunk text.

## Content Segmentation Basic Rules

To segment the content of a CTXT file into Knowledge Chunks, you need to meet the following rules:

* The first Knowledge Chunk starts immediately after the first double line break in the CTXT file.
* Double line breaks separate text into Knowledge Chunks.
* CTXT files must comply with [Knowledge Chunks limitations](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk#limitations).

## CTXT Structure

A CTXT file must include a [header](#header) and a [body](#body).

<Accordion title="CTXT file sample">
  ```markdown theme={null}
  `version: 1`
  `title: Cognigy xApps FAQs`
  `url: https://www.cognigy.com`
  `tags: [cognigy, xapps, architecture]`

  What is Cognigy?
  Cognigy is a leading enterprise software provider for Conversational AI automation. Our platform, Cognigy, automates customer and employee communications. Available in SaaS environments, Cognigy enables enterprises to have natural language conversations with their users on any channel - webchat, SMS, voice and mobile apps - and in any language. Cognigy powers intelligent voice and chatbots that communicate consistently and accurately beyond simple FAQ, resulting in reduced contact center costs and increased efficiency while improving user experiences. The worldwide client portfolio of Cognigy includes Toyota, E.ON, Lufthansa, and many more.
  `url: https://www.cognigy.com/`

  What are xApps?
  Cognigy xApps are infinitely flexible micro-web applications for customer self-service. xApps can be used as a stand-alone application or complementary to any channel to enhance virtual and human agents' capabilities. In multi-channel scenarios xApps close the experience gaps of existing voice and messaging channels to enable dynamic, fully connected customer engagements by helping to overcome limitations of communication channels like voice, Webchat Widget or Microsoft Teams compromising user experiences.
  `url: https://docs.cognigy.com/ai/xApp/overview/`

  What are use cases of xApps
  The use cases are endless as xApps are infinitely flexible due to the fact that they are essentially micro-web applications. Three potential use cases are listed here:

  - Authentication via Credentials: With xApps, you can allow users to authenticate using various methods such as email addresses, user numbers, or other identification details. Users can securely verify their identity and gain access to specific services or information.
  - Select a flight seat: With xApps, you can offer users an interface to choose their preferred seat for their flight. Users can view the available seats and select the seat of their choice, enhancing their travel experience.
  - Share location: With xApps, you can allow users to share their device's location information. This use case can be useful in various scenarios, such as providing location-based services, tracking deliveries, or finding nearby resources.
    `url: https://docs.cognigy.com/ai/xApp/use-cases/`

  How do xApps work?
  The behavior of xApps is controlled via a set of xApp Flow Nodes. An xApp session can be initialized via the "xApp: Init Session" Node. The xApp session can be updated with the actual xApp Pages in two ways, by providing HTML code via the "xApp: Show HTML" Node or by specifying an Adaptive Card in JSON via the "xApp: Show Adaptive Card" Node. Alternatively, for more customization, you can create your own Node to update the xApps state by implementing and uploading an Extension. To wait for the xApp user Input in a Flow, use the Question or the Optional Question Node. The xApp type of the Question Node is only considered to be answered if submitted results are available as part of the Input.
  `url: https://docs.cognigy.com/ai/xApp/build-an-xApp/`

  How is the xApps architecture designed?
  The xApp Shell Page loads its assets (HTML/CSS/JS) by requesting static files via HTTP/HTTPs 1.1 requests to service-static-files and establishes a WebSocket connection (via Socket.io) to serviceapp-session-manager to retrieve information about xApp sessions. This information includes an async event to initialize an xApp session, which creates a unique URL, and specifies which xApp will be displayed in a specific xApp session. The Elastic Load Balancer helps distribute incoming traffic, while Traefik dynamically routes traffic to backend servers based on specific requirements. The serviceapp-session-manager updates the xApp session state in the App Session DB, and the WebSocket transfers data to the Shell page, which loads xApps and provides data to the SDK.
  `url: https://docs.cognigy.com/ai/xApp/architecture/`
  `image: https://docs.cognigy.com/ai/images/xApp/architecture.png`
  ```

  You can download the [previous CTXT file sample](https://docscognigyassets.blob.core.windows.net/assets/cognigy-sample.ctxt) for testing Knowledge AI.
</Accordion>

### Header

The header of a CTXT file contains the [Source metadata](/ai/agents/develop/knowledge-ai/knowledge-source/overview#source-metadata) at the top of the CTXT file. In CTXT files, you must format Source metadata as key-value pairs wrapped by backticks. The keys refer to the information the Source metadata holds, and the value is the information. For example, in `title: Cognigy xApps FAQs`, `title` is the key and `Cognigy xApps FAQs` is the value.

#### Example

```markdown theme={null}
`version: 1`
`title: Cognigy xApps FAQs`
`url: https://www.cognigy.com`
`tags: [cognigy, xapps, architecture]`
```

### Body

The body of a CTXT file contains [Knowledge Chunk text](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk#knowledge-chunk-text) and [Chunk metadata](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk#chunk-metadata). Knowledge Chunks must be separated by double line breaks.

#### Examples

Example 1:

```markdown theme={null}
What are use cases of xApps?
The use cases are endless as xApps are infinitely flexible due to the fact that they are essentially micro-web applications. Three potential use cases are listed here:
- Authentication via Credentials: With xApps, you can allow users to authenticate using various methods such as email addresses, user numbers, or other identification details. Users can securely verify their identity and gain access to specific services or information.
- Select a flight seat: With xApps, you can offer users an interface to choose their preferred seat for their flight. Users can view the available seats and select the seat of their choice, enhancing their travel experience.
- Share location: With xApps, you can allow users to share their device's location information. This use case can be useful in various scenarios, such as providing location-based services, tracking deliveries, or finding nearby resources.
`url: https://docs.cognigy.com/ai/xApp/use-cases/`
```

Example 2:

```markdown theme={null}
How is the xApps architecture designed?
The xApp Shell Page loads its assets (HTML/CSS/JS) by requesting static files via HTTP/HTTPs 1.1 requests to service-static-files and establishes a WebSocket connection (via Socket.io) to serviceapp-session-manager to retrieve information about xApp sessions. This information includes an async event to initialize an xApp session, which creates a unique URL, and specifies which xApp will be displayed in a specific xApp session. The Elastic Load Balancer helps distribute incoming traffic, while Traefik dynamically routes traffic to backend servers based on specific requirements. The serviceapp-session-manager updates the xApp session state in the App Session DB, and the WebSocket transfers data to the Shell page, which loads xApps and provides data to the SDK.
`url: https://docs.cognigy.com/ai/xApp/architecture/`
`image: https://docs.cognigy.com/ai/images/xApp/architecture.png`
```

## More Information

* [Knowledge AI Overview](/ai/agents/develop/knowledge-ai/overview)
* [Knowledge Source](/ai/agents/develop/knowledge-ai/knowledge-source/overview)
* [Knowledge Chunk](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk)
* [Search Extract Output Node](/ai/agents/develop/node-reference/other-nodes/search-extract-output)
* [Web Page](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/web-page)
* [PDF](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/pdf)


# PDF
Source: https://docs.cognigy.com/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/pdf



When you use a PDF file as a Knowledge Source, you can split the PDF file into Knowledge Chunks with the following chunk strategies after uploading the file:

* [Default Chunk Strategy](#default-chunk-strategy)
* [Alternative Chunk Strategy](#alternative-chunk-strategy)

## Default Chunk Strategy

The Default Chunk Strategy extracts content from PDF and splits it into [Knowledge Chunks](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk) based on a fixed token length and recognizes tables. Based on our research, the Default Chunk Strategy returns better structured results and works best with the **Top K** value set to 5 in the [Search Extract Output](/ai/agents/develop/node-reference/other-nodes/search-extract-output) Node.

## Alternative Chunk Strategy

The Alternative Chunk Strategy is effective when your original PDF is divided into paragraphs in a logical order separated by double line breaks. This strategy attempts to distinguish the structure, such as sections, and then splits each of them into one or more Knowledge Chunks.

To apply the Alternative Chunk Strategy, append `.preset_uiolc_ls.pdf` to the file name when you upload a file. For example, if you have a PDF file named `cognigy.pdf`, rename it to `cognigy.preset_uiolc_ls.pdf` before the upload. The `.preset_uiolc_ls` appendix triggers the Alternative Chunk Strategy after you upload the PDF file.

While processing the PDF file, Knowledge AI may omit complex elements such as visually intricate headers or lists and not include these elements in the Knowledge Chunks.

## Examples

Assume you have the following text from the [Cognigy blog](https://www.cognigy.com/blog) in a PDF file:

<Accordion title="PDF sample">
  ```markdown theme={null}
  # Call Tracing: Unlock New Dimensions in VUX Optimization

  In the burgeoning field of voice technology, optimizing Voice User Experience (VUX) is a sophisticated endeavor fraught with unique challenges. 
  VUX designers confront intricate variables from timing nuances and intonation accuracy to the unpredictability of human speech and environmental interference.
  These factors often present hurdles that can transform an otherwise fluid dialogue into a disjointed exchange. 
  For enterprises aiming to scale and perfect their voice-enabled services, these are not mere technicalities but critical pivots on the customer experience journey.
  Recognizing this complex landscape, at Cognigy, we're excited to introduce the ultimate tool for developers and VUX designers in their quest for voice excellence: Call Tracing.

  ## Visualizing the Unseen: How Call Tracing Changes the Game 

  Traditionally, understanding and monitoring voice interaction requires painstaking analysis of transcriptions, user feedback, and basic audio playback. 
  Now, Cognigy's Call Tracing feature offers an actionable visualization of all activities taking place during the call, such as speech recognition and connections, together with the call recording waveforms. 
  This powerful tool grants VUX designers an unparalleled view of the dynamics of voice interactions.
  Event and soundwave visualization provides an analytical perspective that goes beyond the audio, giving professionals the ability to dissect and understand voice experiences at a granular level. 
  It brings critical elements to the forefront that might not be immediately evident in transcriptions or basic audio playback.
  ```
</Accordion>

You can download the [PDF sample](https://docscognigyassets.blob.core.windows.net/assets/call-tracing-sample.pdf).

### Example 1: Default Chunk Strategy

If you use the Default Chunk Strategy, Knowledge AI splits the PDF file into 3 equal Knowledge Chunks.

<Accordion title="Default Chunk Splitting">
  1.

  ```markdown theme={null}
  Call Tracing: Unlock New
  Dimensions in VUX
  Optimization
  In the burgeoning field of voice technology, optimizing Voice User Experience (VUX) is a
  sophisticated endeavor fraught with unique challenges. VUX designers confront intricate
  variables from timing nuances and intonation accuracy to the unpredictability of human
  speech and environmental interference.
  These factors often present hurdles that can transform an otherwise fluid dialogue into a
  disjointed exchange. For enterprises aiming to scale and perfect their voice-enabled
  services, these are not mere technicalities but critical pivots on the customer experience
  journey.
  Recognizing this complex landscape, at
  ```

  2.

  ```markdown theme={null}
    Cognigy, we're excited to introduce the ultimate
  tool for developers and VUX designers in their quest for voice excellence: Call Tracing.
  Visualizing the Unseen: How Call Tracing Changes the
  Game
  Traditionally, understanding and monitoring voice interaction requires painstaking
  analysis of transcriptions, user feedback, and basic audio playback. Now, Cognigy's
  Call Tracing feature offers an actionable visualization of all activities taking place during
  the call, such as speech recognition and connections, together with the call recording
  waveforms. This powerful tool grants VUX designers an unparalleled view of the
  dynamics of voice interactions
  ```

  3.

  ```markdown theme={null}
  .
  Event and soundwave visualization provides an analytical perspective that goes beyond
  the audio, giving professionals the ability to dissect and understand voice experiences
  at a granular level. It brings critical elements to the forefront that might not be
  immediately evident in transcriptions or basic audio playback.
  ```
</Accordion>

### Example 2: Alternative Chunk Strategy

If you use the Alternative Chunk Strategy, Knowledge AI splits this text into 5 Knowledge Chunks.

Note that in the first Knowledge Chunk, a title is missing because it is formatted as a complex element. In comparison, the [Default Chunk Strategy](#example-1-default-chunk-strategy) recognizes and includes the title in the Knowledge Chunks.

<Accordion title="Alternative Chunk Splitting">
  1.

  ```markdown theme={null}
  In the burgeoning field of voice technology, optimizing Voice User Experience (VUX)
  is a sophisticated endeavor fraught with unique challenges. 
  VUX designers confront intricate variables from timing nuances and intonation accuracy 
  to the unpredictability of human speech and environmental interference.
  ```

  2.

  ```markdown theme={null}
  These factors often present hurdles that can transform an otherwise fluid dialogue
  into a disjointed exchange. 
  For enterprises aiming to scale and perfect their voice-enabled services, 
  these are not mere technicalities but critical pivots on the customer experience journey.
  ```

  3.

  ```markdown theme={null}
  Recognizing this complex landscape, at Cognigy, 
  we're excited to introduce the ultimate tool for developers
  and VUX designers in their quest for voice excellence: Call Tracing.
  ```

  4.

  ```markdown theme={null}
  Traditionally, understanding and monitoring voice interaction requires painstaking analysis of 
  transcriptions, user feedback, and basic audio playback. 
  Now, Cognigy's Call Tracing feature offers an actionable visualization of all activities taking place during the call,
  such as speech recognition and connections, together with the call recording waveforms. 
  This powerful tool grants VUX designers an unparalleled view of the dynamics of voice interactions.
  ```

  5.

  ```markdown theme={null}
  Event and soundwave visualization provides an analytical perspective that goes beyond the audio, 
  giving professionals the ability to dissect and understand voice experiences at a granular level. 
  It brings critical elements to the forefront that might not be 
  immediately evident in transcriptions or basic audio playback.
  ```
</Accordion>

## More Information

* [Knowledge AI Overview](/ai/agents/develop/knowledge-ai/overview)
* [Cognigy Text Format](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/ctxt)
* [Web Page](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/web-page)


# Web Page
Source: https://docs.cognigy.com/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/web-page



<a href="/release-notes/4.73"><Badge>Updated in 4.73</Badge></a>

Knowledge AI supports using web pages to create [Knowledge Sources](/ai/agents/develop/knowledge-ai/knowledge-source/overview).

## Restrictions

* The web page must be hosted on a publicly accessible website.
* When you create a Knowledge Source from a web page, Knowledge AI processes all visible text on the web page, including potentially unwanted text, such as cookie notices.
* Knowledge AI doesn't process web page content with anti-crawling measures.
* Knowledge AI doesn't support images or Optical Character Recognition (OCR) capabilities.

## Chunking Process

When processing a web page, Knowledge AI:

1. Visits the URL as a page in a browser session.
2. Scrolls to the bottom of the web page.
3. Accesses lazy-loaded<sup>[1](#footnote1)</sup> content by checking for any text changes until the web page is stable and no longer loads additional text.
4. Generates Knowledge Source content based on the visible text result.

Knowledge AI imports the web page content into a Knowledge Source once. If the web page you used as a source is updated and you want to include the new content in the web page, create a new Knowledge Source based on the updated web page.

## More Information

* [Knowledge AI Overview](/ai/agents/develop/knowledge-ai/overview)
* [Create a Knowledge Store](/ai/agents/develop/knowledge-ai/getting-started)
* [Chunk editor](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk)
* [Cognigy Text Format](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/ctxt)
* [PDF](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/pdf)

***

<sup>1</sup>: Lazy loading is a web development technique that delays loading non-critical or non-visible content until it is necessary. This technique improves web page loading times and user experience.


# Knowledge Store
Source: https://docs.cognigy.com/ai/agents/develop/knowledge-ai/knowledge-store

Learn how to manage Knowledge Stores to apply RAG-based knowledge search in Cognigy.

A *Knowledge Store* provides a centralized and structured environment for managing and categorizing [Knowledge Sources](/ai/agents/develop/knowledge-ai/knowledge-source/overview). Knowledge Stores help streamline the knowledge management process by grouping related Knowledge Sources, making it easier to search and retrieve relevant information during run-time. You can select Knowledge Stores in [Search Extract Output Nodes](/ai/agents/develop/node-reference/other-nodes/search-extract-output) to retrieve information from Knowledge Sources for your AI Agents. You can create any number of Knowledge Stores per Project.

## Prerequisites

* You have added a [Large Language Model (LLM) to your Project](/ai/agents/develop/gen-ai-and-llms/providers/all-providers) that [supports Knowledge Search](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature).

## Restrictions

* All Knowledge Stores within a Project use the same embedding model. After you [configure the embedding model of your first Knowledge Store](/ai/agents/develop/knowledge-ai/overview), all other Knowledge Stores will use the same embedding model. If you want to switch to a different embedding model, delete all existing Knowledge Stores in the current Project and create new ones, or [create a new Project](/ai/agents/develop/projects-and-flows/projects).
* When you import a [Package](/ai/platform-features/packages) including a Knowledge Store, the source and target Projects must use the same embedding model. If they use different embedding models, recreate the [Knowledge Sources](/ai/agents/develop/knowledge-ai/knowledge-source/overview) to generate new embeddings with the target Project's model.

## Working with Knowledge Stores

<Tabs>
  <Tab title="GUI">
    You can view, create, edit, and delete Knowledge Stores in **Build > Knowledge**. Also, you can copy Knowledge Stores' Reference ID and export Knowledge Stores as Packages.
  </Tab>

  <Tab title="API">
    You can view, create, edit, and delete Knowledge Stores using the [Cognigy API](https://api-trial.cognigy.ai/openapi#tag--KnowledgeStores).
  </Tab>

  <Tab title="CLI">
    You can create Knowledge Stores using the [Cognigy CLI](https://github.com/Cognigy/Cognigy-CLI/blob/main/KNOWLEDGE-AI-README.md).
  </Tab>
</Tabs>

## More Information

* [Knowledge Source](/ai/agents/develop/knowledge-ai/knowledge-source/overview)
* [Cognigy Text Format](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/ctxt)
* [PDF](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/pdf)
* [Web Page](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/web-page)
* [Search Extract Output Node](/ai/agents/develop/node-reference/other-nodes/search-extract-output)
* [LLM](/ai/agents/develop/gen-ai-and-llms/llms)
* [Generative AI](/ai/agents/develop/gen-ai-and-llms/generative-ai)


# Knowledge AI
Source: https://docs.cognigy.com/ai/agents/develop/knowledge-ai/overview

Upload documents and use RAG to enable AI Agents to provide accurate, context-aware responses with Knowledge AI.

<a href="/release-notes/2026.13"><Badge>Updated in 2026.13</Badge></a>

<Note>
  * You are subject to the terms of the Generative AI model providers you are using. Cognigy doesn't take responsibility for your use of third-party services, systems, or materials.
  * If you are a Cognigy customer, the performance timelines (SLA) specified in your contract do not apply to the trial version of Knowledge AI.
</Note>

Knowledge AI is a knowledge search and management solution that uses Retrieval-Augmented Generation (RAG) to improve AI Agents' performance and assist human agents. Knowledge AI applies RAG to provide AI Agents with access to a large amount of structured information from different files in a knowledge base, such as articles, manuals, FAQs, and more. By accessing the knowledge base, AI Agents can retrieve and understand specific information, and provide more accurate, context-aware, and helpful responses to user queries.

With Knowledge AI, you have more options beyond relying on [Intents](/ai/platform-features/nlu/intents/overview) and [Default Replies](/ai/platform-features/nlu/intents/default-reply) to identify user questions and provide relevant content based on predefined questions and answers. Creating question-and-answer pairs can be time-consuming and require labor-intensive maintenance.

Knowledge AI lets you upload knowledge in various formats, such as [PDF](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/pdf), [Cognigy Text](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/ctxt) (CTXT) and [web pages](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/web-page), or connect to external systems via [Knowledge Connectors](/ai/agents/develop/knowledge-ai/knowledge-connectors), to create [Knowledge Sources](/ai/agents/develop/knowledge-ai/knowledge-source/overview). This technology extracts meaningful information from these documents and makes it accessible to conversation designers through the [Search Extract Output Node](/ai/agents/develop/knowledge-ai/knowledge-search-and-extraction/search-extract-output). With this Node, you can build knowledge-based AI Agents quickly, overcoming the limitations of traditional intent-based systems and simplifying the creation of sophisticated conversational experiences.

## Prerequisites

Before using Knowledge AI, follow these steps:

1. [Apply for a license and allocate quotas](/ai/agents/develop/knowledge-ai/activate).
2. Get access to an account from one of the following LLM Providers:

   * [OpenAI](https://platform.openai.com/)
   * [Microsoft Azure OpenAI](https://azure.microsoft.com/en-us/products/cognitive-services/openai-service)
   * [Aleph Alpha](https://aleph-alpha.com/)

### Which Models to Choose?

To use the full potential of Knowledge AI, you need the following models:

<AccordionGroup>
  <Accordion title="Embedding model">
    An embedding model converts information from documents into embeddings in [Knowledge Sources](/ai/agents/develop/knowledge-ai/knowledge-source/overview) and performs knowledge searches.

    <Note>
      Embedding models aren't interchangeable. Each [Project](/ai/agents/develop/projects-and-flows/projects) uses a single embedding model for all Knowledge Stores. You configure the embedding model when you create the first Knowledge Store in your Project. If you change the embedding model, the existing embeddings become incompatible with the new model. To use a different embedding model, you must either delete all [Knowledge Stores](/ai/agents/develop/knowledge-ai/knowledge-store) and recreate them, or create a new Project.

      When you import a [Package](/ai/platform-features/packages) including a Knowledge Store, the source and target Projects must use the same embedding model. If they use different embedding models, recreate the [Knowledge Sources](/ai/agents/develop/knowledge-ai/knowledge-source/overview) to generate new embeddings with the target Project's model.
    </Note>

    Knowledge AI supports the following embedding models:

    | Model                                       | Comment                                                                                                                                                                                                                                                                                                                                                                           |
    | ------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | <span>`text-embedding-ada-002`</span>       | –                                                                                                                                                                                                                                                                                                                                                                                 |
    | <span>`text-embedding-3-small`</span>       | The model may be available only in geographically remote locations, leading to high latency for the search operation.                                                                                                                                                                                                                                                             |
    | <span>`text-embedding-3-large`</span>       | The model may be available only in geographically remote locations, leading to high latency for the search operation. The model outputs embedding vectors twice as large as those from `text-embedding-ada-002` and `text-embedding-3-small`. Larger embedding vectors result in higher memory usage and larger [Package](/ai/platform-features/packages) sizes for Knowledge AI. |
    | <span>`amazon.titan-embed-text-v2:0`</span> | For Cognigy 2025.10 and earlier versions, use the `FEATURE_ENABLE_AWS_BEDROCK_EMBEDDING_LLM_WHITELIST` environment variable to make this model available to individual or all feature flags.                                                                                                                                                                                      |
    | <span>`gemini-embedding-2`</span>           | Available for Cognigy 2026.13 and later versions.                                                                                                                                                                                                                                                                                                                                 |
  </Accordion>

  <Accordion title="Generative AI model">
    A generative model must support [LLM Prompt Nodes or Answer Extraction](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) to use Knowledge Search results and generate context-aware response to user questions.
  </Accordion>
</AccordionGroup>

## Knowledge Management

Knowledge AI has the following knowledge hierarchy:

* [Knowledge Store](/ai/agents/develop/knowledge-ai/knowledge-store) — a container that includes Knowledge Sources.
* [Knowledge Source](/ai/agents/develop/knowledge-ai/knowledge-source/overview) — a collection of organized and structured knowledge extracted from files, for example, user manuals, articles, FAQs, and other relevant information.
* [Knowledge Chunk](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk) — a unit of knowledge inside a Knowledge Source.

<Frame>
  <img alt="Knowledge AI hierarchy from Knowledge Store to Source to Chunk" />
</Frame>

## Knowledge AI Workflow

Working with Knowledge AI involves two phases, similar to the Intents workflow. The first phase consists of collecting, preprocessing, and integrating data — a process known as ingestion — from your knowledge base, and the second phase is querying the knowledge base during [run-time](/ai/agents/develop/gen-ai-and-llms/generative-ai#run-time-generative-ai-features).

### First Phase

1. **Raw Information Upload** — you upload information as files to Knowledge AI.
2. **Knowledge Chunk Extraction** — Knowledge AI extracts text and metadata from the raw information through a collection of tools and processes the information into a Knowledge Source. The Knowledge Source is divided in Knowledge Chunks, which you can edit in the Chunk editor.
3. **Vectorization** — Knowledge AI encodes Knowledge Chunks into embeddings. Embeddings are high-dimensional vectors that encode word meaning and similarity into numeric representations. Knowledge AI stores the embeddings in a specialized internal database for quick access during run-time.

### Second Phase

1. **Knowledge Base Querying**. During run-time, Knowledge AI queries the Knowledge Sources to provide accurate and contextually appropriate responses to user questions.
2. **Knowledge-based AI Agents Building**. AI Agents use the information stored in the Knowledge Sources to engage in more sophisticated and intelligent conversations with users.

## Use cases

With Knowledge AI, both human agents and AI Agents can retrieve information from various Knowledge Sources and provide better responses to your customers:

* **Knowledge AI for AI Agents** — use the [Search Extract Output Node](/ai/agents/develop/node-reference/other-nodes/search-extract-output) so that AI Agents can retrieve product-specific knowledge from Knowledge Sources, for example, manuals and web pages. This way, AI Agent's output more accurate information to customers.
* **Use Knowledge AI for human agents** — by retrieving data from different Knowledge Sources, LLMs reduce the burden on human agents, making information access quicker and more efficient. To let human agents use the Knowledge Sources for solving tasks, add a widget to the Agent Copilot workspace using a [Copilot: Knowledge Tile Node](/ai/agents/develop/node-reference/agent-copilot/knowledge-tile). The widget serves as a search tool, enabling human agents to enter questions and retrieve information from the knowledge base.

To know more about the benefits of using Knowledge AI, watch this webinar:

<iframe title="Cognigy Sessions: Cognigy Slot Fillers" />

<Accordion title="FAQ">
  **Q1**: Is Knowledge AI free of charge?

  **A1**: Knowledge AI is not free of charge and requires a [separate license](/ai/agents/develop/knowledge-ai/activate#apply-for-a-license).

  **Q2**: I received `Request failed with status code 429` error message while attempting to upload a file. How can I solve this issue?

  **A2**: The `429` error occurs when your organization's rate limit is exceeded on the side of your LLM's provider. To learn more, refer to your provider's documentation. For instance, if you use the OpenAI API, check out the article [How can I solve 429: 'Too Many Requests' errors?](https://help.openai.com/en/articles/5955604-how-can-i-solve-429-too-many-requests-errors).

  **Q3**: I received an `Error while performing knowledge search. Remote returned error: Search failed: Could not fetch embeddings due to missing API resource name for Azure OpenAI` error message while using knowledge search features. How can I solve this issue?

  **A3**: In recent releases, we have updated the connection settings to Azure OpenAI LLMs and added new parameters, such as the Resource Name. If you have an older connection (for example, created in the 4.53 release) to Azure OpenAI LLMs, especially Azure OpenAI `text-embedding-ada-002` for Knowledge Search, you might receive this error message when you trigger an LLM. To resolve this issue, recreate the LLM and the connection so that both have the latest format.
</Accordion>

## More Information

* [Getting started with Knowledge AI](/ai/agents/develop/knowledge-ai/getting-started)
* [Cognigy Text Format](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/ctxt)
* [PDF](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/pdf)
* [Web Page](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/web-page)
* [Search Extract Output Node](/ai/agents/develop/node-reference/other-nodes/search-extract-output)
* [LLM](/ai/agents/develop/gen-ai-and-llms/llms)
* [Generative AI](/ai/agents/develop/gen-ai-and-llms/generative-ai)


# Manage AI Agents
Source: https://docs.cognigy.com/ai/agents/develop/manage-ai-agents



<a href="/release-notes/2026.16"><Badge>Updated in 2026.16.0</Badge></a>

To manage a new generation of AI Agents, Cognigy introduces the *AI Agent Management* interface,
where you can create, edit, and delete AI Agents.

## Create AI Agents

You can create an AI Agent from scratch or hire an AI Agent from the Job Market, designed for your industry, such as airlines, food services, or sales, and customize it to fit your business needs.

To create an AI Agent, select one of the following options:

* [Create from scratch](#create-from-scratch)
* [Hire from the Job Market](#hire-from-the-job-market)

### Create from Scratch

<Tabs>
  <Tab title="GUI">
    1. In the left-side menu of the **Project** page, select **Build > AI Agents**.
    2. Click **Create AI Agent** if you are creating an AI Agent for the first time, or click **+ New AI Agent** on the **AI Agents** page if you have created at least one AI Agent before. Configure the following settings:

    <AccordionGroup>
      <Accordion title="1. General Settings">
        1) In the **General Settings** step, fill in the following fields:
           * **Name** — enter a name for your AI Agent. This name will be visible to end users when the AI Agent greets them in the chat or when one AI Agent transfers the conversation to another AI Agent. For example, `Hello, my name is Sara. How can I help you?` or `Our support specialist Alex will help with this question, transferring the conversation now`.
           * **Description** — enter a description of the AI Agent that shapes its behavior and enhances its ability to understand the desired communication style. For example, describe the AI Agent as follows: `Anna works at ACME and is highly skilled at making customers feel comfortable.`
        2) In the **Avatar** section, select avatar of your choice or upload your custom one by clicking **+** next to avatars.
           You can use any image as an avatar for your AI Agent, or create a Cognigy-style avatar using the [Cognigy PSD template](https://docscognigyassets.blob.core.windows.net/assets/Template.psd). To open and customize the template, you will need Adobe Photoshop or [Paint.net](https://www.getpaint.net/) with the [PSD plugin](https://www.psdplugin.com/) installed. The requirements for the avatar are as follows:
           * Use an alpha channel for a transparent background.
           * Set the recommended width to 136px.
           * Set the recommended height to 184px.
           * Save the file as `.png`.
           * Include `_optimized` in the file name.
      </Accordion>

      <Accordion title="2. Speaking Style">
        1. In the **Speaking Style** step, configure the following settings:
           * **Set up Style** — configure how the AI Agent's choice of wording is influenced by the selected speaking style, which can affect the tone and clarity of its responses:
             * **Concise/Comprehensive** — adjust the slider to the left for brief responses and to the right for detailed responses.
             * **Formal/Informal** — adjust the slider to the left for more casual and conversational responses, and to the right for professional and structured responses.
           * **Voice Configuration** — assign a specific voice to your AI Agent if you want to use your AI Agent as a voice-based assistant. To configure this setting, ensure you have installed [Voice Gateway](/voice-gateway/index) and set up the [Voice Preview](/ai/agents/test/voice-preview) provider. Then, fill in the fields following the same process you used for [Set Session Config Node](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details):
             * **TTS Vendor** — select the text-to-speech provider for your AI Agent. This setting determines which service will convert the AI Agent's text responses into speech.
             * **TTS Language** — select the language that your AI Agent will use for speech output. Ensure this language aligns with the preferred language of the end user.
             * **TTS Voice** — select the voice from the selected TTS vendor. This setting allows you to customize the tone, gender, and style of your AI Agent's voice.
             * **TTS Label** — select the alternative name of the TTS vendor, as specified in the [Voice Gateway Self-Service Portal](/voice-gateway/webapp/applications#add-additional-tts-and-stt-vendor). If you have multiple speech services from the same vendor, use the label to specify which service to use.
             * **Disable TTS Audio Caching** — by default, this setting is deactivated. With caching enabled, previously requested TTS audio is stored in the AI Agent cache, and repeated requests for the same audio text will use the cached result instead of sending another request. With caching disabled, the AI Agent stores the TTS audio but doesn't use it; each request is sent directly to the speech provider. Note that disabling caching can increase TTS costs. For detailed information, contact your speech provider.
             * **Language Configuration** — by default, this setting is activated. Deactivate automatic language detection to ensure the AI Agent always responds in the language defined by your instructions. This setting provides predictable and consistent responses, preventing unintended language switching when users say or type words from more than one language. When active, this setting also simplifies the system prompt, giving you clearer control over language behavior.
      </Accordion>

      <Accordion title="3. Instructions">
        1. In the **Instructions** step, configure the following setting:
           * **Instructions** — provide special instructions to your AI Agent in bullet-point form. For example:
             ```txt theme={null}
              - Greet users warmly and professionally.
              - Keep responses concise; expand only if requested.
              - Start with a formal tone; adjust as needed.
              - Share troubleshooting links for technical issues.
              - Apologize if errors occur, then correct promptly.
             ```
      </Accordion>

      <Accordion title="4. Knowledge Store">
        1. In the **Knowledge Store** step, configure the following settings:
           * **Knowledge Store Type** — select or upload a knowledge source that the AI Agent will use to access information from the documents you provide. By accessing and understanding these knowledge bases, the AI Agent can deliver more accurate, context-aware, and helpful responses to user queries. You need to configure an [embedding model](/ai/agents/develop/knowledge-ai/overview) to use Knowledge AI. Select one of the following options:
             * **Choose existing Knowledge Store** — select the [Knowledge Store](/ai/agents/develop/knowledge-ai/knowledge-store) that the AI Agent will use to access information from the documents you provide.
             * **Upload Knowledge Source** — upload documents with supported formats, such as PDF, text, DOCX, PPTX, or CTXT file formats. The CTXT file format has restrictions. For more information, refer to the [CTXT](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/ctxt) article.
             * **Upload URL Knowledge Source** — enter the URL of the web page to be used as a Knowledge Source. This type of Knowledge Source has restrictions. For more information, refer to the [Web Page](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/web-page) article.
      </Accordion>

      <Accordion title="5. Data Privacy & Security">
        1. In the **Data Privacy & Security** step, configure the following fields:
           * **Contact Profile Information** — select which information the AI Agent should use from the [Cognigy Contact Profile](/ai/agents/analyze/contact-profiles):
             * **None** — no data will be used from the Contact Profile. This option is selected by default.
             * **Selected Profile Fields** — enter specific fields from the Contact Profile for targeted data use. Specify the field using the [Profile keys](/ai/agents/analyze/contact-profiles#profile-schema) format and press <kbd>Enter</kbd> to apply it.
             * **Complete Profile** — use all fields from the Contact Profile to provide comprehensive user details.
             * **Profile Memories** — use the [Memories](/ai/agents/analyze/contact-profiles#profile-schema) field from the Contact Profile.
           * **Safety Instructions** — adjust the AI Agent's safety settings to guide content generation, interactions, and responses, ensuring compliance with ethical, legal, and operational standards. Although these settings reduce risks, occasional unexpected outputs may still occur. The selected safety instructions are included in the prompt to enhance safety, which may increase token usage. Select the safety instructions you want to apply:
             * **Avoid harmful content** — prevent generating content that could be harmful, offensive, or abusive to end users.
             * **Avoid ungrounded content** — prevent generating content that is based on speculation or unsupported claims, ensuring it is reliable and verifiable.
             * **Avoid copyright infringements** — prevent generating content that violates intellectual property rights or uses copyrighted material without authorization.
             * **Prevent jailbreak and manipulations** — prevent attempts to bypass security measures or manipulate the AI Agent into producing unauthorized or unsafe content.
      </Accordion>

      <Accordion title="6. Job Selection">
        1. In the **Job Selection** step, select one of the following options:
           * **Default** — create an AI Agent with a predefined Flow, then click **Create & Configure LLM** to save changes and open the Flow with the created AI Agent. If you haven't added an LLM before, the system will prompt you to [configure a model](/ai/agents/overview#prerequisites) to ensure your AI-driven Flow works.
           * **Personality Only** — create an AI Agent without a predefined Flow, then click **Create** to save changes.
           * **Job** — select one of the available jobs you want to assign to the AI Agent, then click **Create** to save changes.
      </Accordion>
    </AccordionGroup>
  </Tab>

  <Tab title="API">
    You can create an Agent from scratch using the [Cognigy API POST beta/aiagents](https://api-trial.cognigy.ai/openapi#post-/beta/aiagents) request.
  </Tab>
</Tabs>

### Hire from the Job Market

<Tabs>
  <Tab title="GUI">
    1. In the left-side menu on the **Project** page, select **Build > AI Agents**.
    2. On the **AI Agents** page, click **Hire AI Agent**. The **Job Market** page displays a list of available AI Agents to hire.
    3. Hover your cursor over the desired AI Agent template and click **Hire**. This action will trigger the **Hire AI Agent** task in the Task Manager.
    4. Check the status of the **Hire AI Agent** task by clicking <img alt="task-menu" /> in the upper-right corner.
    5. Once the task is complete, go to **Build > Flows** and find the Flow with the AI Agent name from the template you installed.
  </Tab>

  <Tab title="API">
    You can hire an Agent from the Job Market using the [Cognigy API POST beta/aiagents/hire](https://api-trial.cognigy.ai/openapi#post-/beta/aiagents/hire) request.
  </Tab>
</Tabs>

## Combine Custom and Hired AI Agents

You can reassign your custom AI Agent to take on the responsibilities of the hired AI Agent.
To do this, hire an AI Agent from the Job Market
and configure their interaction by defining which functions and personality traits should be combined.
This approach lets you combine the strengths of both AI Agents: ready-made job functions from the hired AI Agent and your brand's AI Agent persona from your custom AI Agent.

To combine custom and hired AI Agents, follow these steps:

1. On the **AI Agents** page, select **+ Hire AI Agent**. The **Job Market** page displays a list of available AI Agents to hire.
2. Hover your cursor over the desired AI Agent template and click <img alt="expand" /> **> Hire & Configure**.
3. In the **Select an AI Agent to Combine** window, use the search field to find an AI Agent by name, then select the AI Agent from the list.
4. Click **Combine**. This action will trigger the **Hire AI Agent** and **Merge Package** tasks in the Task Manager.
5. In the left-side menu of the **Project** page, select **Build > Flows**.
6. Select the Flow that belongs to the hired AI Agent. In the Flow, you will see an AI Agent Node with your custom AI Agent persona.

## Configure Jobs and Tools for AI Agents

When you create an AI Agent from scratch or add one from the Job Market, you can add or change jobs and tools in the AI Agent configuration.

1. Open the created AI Agent and go to the **Jobs** tab.
2. Fill in the following fields:
   * **Job Name** — enter a title for the job you want to assign to the AI Agent. For example, `Customer Support Specialist`.
   * **Job Description** — enter a description of the job that outlines the responsibilities and tasks associated with the role. For example, `Handle customer inquiries, provide product information, and resolve issues efficiently`.
   * **Instructions and Context** — provide specific instructions or guidelines that the AI Agent should follow while performing its job. For example, `Always respond politely, escalate complex issues to a human agent, and ensure customer satisfaction`.

### Configure Tools for AI Agents

In the **Tools** section, click **Add Tool** to add tools that the AI Agent can use to perform its job. Select one of the following options:

<Tabs>
  <Tab title="Regular Tool">
    A tool is a specific task that an AI Agent can perform. The AI Agent can call tools to perform different tasks, such as retrieving information from a database or interacting with an external API. Fill in the following fields:

    <AccordionGroup>
      <Accordion title="Tool">
        | Parameter   | Type          | Description                                                                                                                                          |
        | ----------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------- |
        | Tool ID     | CognigyScript | Provide a meaningful name as a tool ID. This ID can contain only letters, numbers, underscores (`_`), or dashes (`-`). For example, `update_user-1`. |
        | Description | CognigyScript | Provide a detailed description of what the tool does, when it should be used, and its parameters.                                                    |
      </Accordion>

      <Accordion title="Parameters">
        Configure the parameters that will be collected by the AI Agent before the tool is called. You can switch between the Graphical and JSON editors. When editing the JSON, follow the [JSON Schema specification](https://json-schema.org).

        | Parameter       | Type     | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
        | --------------- | -------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
        | Use Parameters  | Toggle   | Activate this toggle to add parameters in addition to the tool name and description. The AI Agent will collect all data it needs and call a Tool with these parameters filled as arguments. These values can be accessed directly in the `input.aiAgent.toolArgs` object.                                                                                                                                                                                                                                                                                                                                                                            |
        | Name            | Text     | Specify the name of the parameter. The name should be clear and concise, and describe the purpose of the parameter.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
        | Type            | Selector | Select a type of the parameter:<ul><li>**String** — a sequence of characters. For example, `"hello"`, `"123"`.</li><li>**Number** — a numerical value, which can be either an integer (for example, `5`) or a floating point number (for example, `3.14`).</li><li>**Boolean** — a logical value representing `true` or `false`.</li><li>**Array** — a collection of elements, which can contain multiple values of any type. For example, `["apple", "banana", "cherry"]`.</li><li>**Object** — a collection of key-value pairs, where each key is a string and the value can be of any type. For example, `{"name": "John", "age": 30}`.</li></ul> |
        | Description     | Text     | Explain what the parameter means by providing a brief description of the parameter's usage.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
        | Enum (optional) | Enum     | Define a set of values that the parameter can accept. The enum restricts the input to one of the specified values, ensuring only valid options are chosen. The enum is only available for string-type parameters in the Graphical editor. For other types, use the JSON editor. May not be supported by all LLM providers.                                                                                                                                                                                                                                                                                                                           |
        | Add Parameter   | Button   | Add a new parameter.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
      </Accordion>

      <Accordion title="Debug Settings">
        | Parameter                 | Type   | Description                                                                                                       |
        | ------------------------- | ------ | ----------------------------------------------------------------------------------------------------------------- |
        | Debug Message when called | Toggle | Enable the output of a debug message when the tool is called to provide detailed information about the tool call. |
      </Accordion>

      <Accordion title="Advanced">
        | Parameter | Type          | Description                                                                                                                                                                                                                                                                                                                                                                        |
        | --------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
        | Condition | CognigyScript | The tool will be enabled only if the condition is evaluated as true. If false, the tool isn't part of the AI Agent's Tools within this execution. For example, when using the `unlock_account` tool, you can specify a condition like `context.accountStatus === "locked"`. This checks the value in the context, and if it is missing or different, the tool will not be enabled. |
      </Accordion>
    </AccordionGroup>
  </Tab>

  <Tab title="MCP">
    An MCP tool connects to a remote [MCP server](https://modelcontextprotocol.io/introduction) to load tools that the AI Agent can execute. Fill in the following fields:

    <AccordionGroup>
      <Accordion title="MCP Tool">
        | Parameter          | Type          | Description                                                                                                                                                                            |
        | ------------------ | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
        | Name               | CognigyScript | Provide a name for the MCP connection. This name helps you identify the source of the loaded tool.                                                                                     |
        | MCP Server SSE URL | CognigyScript | Provide the URL to an SSE (Server-Sent Events) endpoint from a remote [MCP server](https://modelcontextprotocol.io/introduction). Ensure that you connect only to trusted MCP servers. |
        | Timeout            | Slider        | Set the timeout time for the MCP connection in seconds.                                                                                                                                |
      </Accordion>

      <Accordion title="Authentication">
        | Parameter           | Type     | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
        | ------------------- | -------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
        | Authentication Type | Selector | The authentication type to use for the MCP server connection. Select one of the following options: <ul><li>**None** — no authentication. This option is used by default.</li><li>**OAuth2** — authenticate using OAuth 2.0, which provides improved security and control compared to API keys. OAuth 2.0 supports scoped permissions, expiring tokens, and reduced exposure through short-lived tokens.</li></ul> If you want to use a different authentication method, such as an API key, you can add it as a custom header in **Custom Headers** of the **Advanced** section.                                                                                                                                                                                                                                                             |
        | OAuth2 Parameters   | Selector | Opens the fields required to configure OAuth 2.0 authentication. These fields allow the AI Agent to securely request and use access tokens when communicating with the MCP server. Configure the following fields: <ul><li>**Connection Name** — enter a descriptive name for the connection, for example, `MCP Tool OAuth`.</li><li>**OAuth URL** — enter the token endpoint URL exposed by the MCP server, for example: `https://mcp.example.com/oauth2/token`.</li><li>**Client ID** — enter the application (client) ID issued by the MCP server.</li><li>**Client Secret** — enter the application secret issued by the MCP server. Keep it secure.</li><li>**Scope** — add a list of scopes that define the permissions required by your application, for example: `read write admin`. Separate multiple scopes with spaces.</li></ul> |
      </Accordion>

      <Accordion title="Debug Settings">
        | Parameter             | Type   | Description                                                                                                                                                                                                                                                               |
        | --------------------- | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
        | Debug Loaded Tools    | Toggle | Enable this parameter to display a debug message listing all tools loaded from the MCP server. The debug message also includes tools filtered out in the Advanced section. This parameter shows whether tools were loaded from the cache or directly from the MCP server. |
        | Debug with Parameters | Toggle | Enable this parameter to include the Tool Parameters in the debug message.                                                                                                                                                                                                |
        | Debug calling Tool    | Toggle | Enable the output of a debug message when the tool is called to provide detailed information about the tool call.                                                                                                                                                         |
      </Accordion>

      <Accordion title="Advanced">
        | Parameter      | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
        | -------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
        | Cache Tools    | Toggle        | Disables caching of loaded tools while developing. Ensure that caching is enabled in production for performance reasons. The caching time is 10 minutes. In debug mode, you can verify whether the cache was used. To do this, make sure **Debug Loaded Tools** is enabled. For more information, see the [Debug Settings](#debug-settings) section.                                                                                                                    |
        | Condition      | CognigyScript | Sets the condition under which the tool will be activated. If the condition is evaluated as false, the tool isn't part of the AI Agent's Tools during execution. For example, when using the `unlock_account` tool, you can specify a condition like `context.accountStatus === "locked"`. This checks the value in the context, and if it is missing or different, the tool will not be enabled.                                                                       |
        | Tool Filter    | Selector      | Controls if tools should be excluded from execution. You can select one of the following options:<ul><li>**None** — no tool filtering is applied, and all tools are available for execution. This option is selected by default.</li><li>**Whitelist** — only tools on the list are allowed for execution, while all other tools are excluded.</li><li>**Blacklist** — tools on the list are excluded from execution, while all other tools remain available.</li></ul> |
        | Blacklist      | CognigyScript | The parameter appears if you select **Blacklist** in **Tool Filter**. Specify the tools that should be blocked from execution. Specify only one tool per field.                                                                                                                                                                                                                                                                                                         |
        | Whitelist      | CognigyScript | This parameter appears if you select **Whitelist** in **Tool Filter**. Specify the tools you want to allow for execution. Specify only one tool per field.                                                                                                                                                                                                                                                                                                              |
        | Custom Headers | -             | Sets custom authentication headers to send with the request to the MCP server. Use the **Key** and **Value** fields to enter a header. The **Value** field supports CognigyScript. After entering the header key, new empty **Key** and **Value** fields are automatically added, in case you need to add more headers. Alternatively, you can click **Show JSON Editor** and enter the headers in the code field.                                                      |
      </Accordion>
    </AccordionGroup>

    You can add multiple tools to the AI Agent. The recommended maximum number of tools is 20. The following video shows an MCP Tool example:

    <iframe title="MCP Tool" />
  </Tab>

  <Tab title="Knowledge">
    A Knowledge tool is a tool that allows the AI Agent to directly access [Knowledge Stores](/ai/agents/develop/knowledge-ai/knowledge-store) to provide context-aware responses. Fill in the following fields:

    <AccordionGroup>
      <Accordion title="Tool">
        | Parameter       | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
        | --------------- | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
        | Knowledge Store | Selector      | Select a Knowledge Store.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
        | Tool ID         | CognigyScript | Enter a name as a tool ID. This ID can contain only letters, numbers, underscores (`_`), or dashes (`-`). If you use more than one Knowledge tool, the tool ID provides context to trigger the correct Knowledge tool. In this case, enter a clear tool ID for each Knowledge tool, for example, `search_appliances`.                                                                                                                                                                                                                                                                                                                       |
        | Description     | CognigyScript | Enter instructions to guide the AI Agent to call the Knowledge tool. The description field provides context to trigger the Knowledge tool. If you want to use more than one Knowledge tool, enter clear instructions for the cases when each Knowledge tool should be used. For example, `Find the answer to prompts or questions about appliances by searching the attached data sources. Use this tool when a customer asks about appliance items such as washing machines, dryers, and other household appliances. Focus exclusively on a knowledge search and does not execute tasks like small talk, calculations, or script running.` |
      </Accordion>

      <Accordion title="Debug Settings">
        | Parameter                 | Type   | Description                                                                                                       |
        | ------------------------- | ------ | ----------------------------------------------------------------------------------------------------------------- |
        | Debug Message when called | Toggle | Enable the output of a debug message when the tool is called to provide detailed information about the tool call. |
      </Accordion>

      <Accordion title="Advanced">
        | Parameter                   | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
        | --------------------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
        | Top K                       | Slider        | Set how many Knowledge Chunks to return. Providing more results gives the AI Agent additional context, but it also increases noise and token usage.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
        | Store Location              | Selector      | Select whether and where to store the knowledge search results. Select one of the following options:<ul><li>**Don't store** — the content isn't stored. This option is set by default.</li><li>**Input** — the content is stored in the Input object.</li><li>**Context** — the content is stored in the Context object.</li></ul>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
        | Input Key to store result   | CognigyScript | Appears when **Store Location** is set to **Input**. The property in the Input object where the result is stored. For example, `input.knowledgeSearch`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
        | Context Key to store result | CognigyScript | Appears when **Store Location** is set to **Context**. The property in the Context object where the result is stored. For example, `context.knowledgeSearch`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
        | Source Tags                 | CognigyScript | Enter [Knowledge Source Tags](/ai/agents/develop/knowledge-ai/knowledge-source/knowledge-source-tags) to refine the scope of your knowledge search, including only the most relevant Knowledge Chunks in the Knowledge Store. Before entering tags, ensure they are included in the Knowledge Sources. Add tags by entering each separately and pressing <kbd>Enter</kbd>. Max 5 tags. When multiple Source Tags are specified, the Search Extract Output Node defaults to the `AND` operator, meaning it only considers Sources that have all specified tags. To change this behavior, adjust the **Match Type for Source Tags** parameter.                                                                                                                                                                      |
        | Match type for Source Tags  | Selector      | The operator to filter Knowledge Sources by Source Tags. Select one of the following options:<ul><li>**AND** — the default value, requires all Source Tags to match across multiple Knowledge Sources. Consider the following example: there are Knowledge Sources with Source Tags `S-a`, `S-b`, and `S-c`. When you use the `AND` operator to filter by `S-a` and `S-b`, only Sources with both Tags `S-a` and `S-b` are included in the search results.</li><li>**OR** — requires at least one Source Tag to match across multiple Knowledge Sources. Consider the following example: there are Knowledge Sources with Source Tags `S-a`, `S-b`, and `S-c`. When you use the `OR` operator to filter by `S-a` or `S-b`, any Source with either Tag `S-a` or `S-b` is included in the search results.</li></ul> |
        | Condition                   | CognigyScript | The Knowledge tool is activated only if the condition is evaluated as true. If false, the tool isn't included in the current execution. For example, when using the Knowledge tool, you can enter a condition such as `context.productCategory === "appliances"`. This checks the value in the context, and if it is missing or different, the tool isn't activated.                                                                                                                                                                                                                                                                                                                                                                                                                                              |
      </Accordion>
    </AccordionGroup>
  </Tab>

  <Tab title="Send Email">
    The Send Email tool lets your AI Agent send emails directly to users. The Send Email tool has the capabilities and restrictions as the [Email Notification Node](/ai/agents/develop/node-reference/data/email-notification) but is more flexible. While the Email Notification Node sends emails at a fixed step in a Flow, the Send Email tool allows the AI Agent to send emails dynamically, based on user input, conversation context, or instructions. This tool makes automation more flexible and minimizes extra Flow steps.

    <AccordionGroup>
      <Accordion title="Tool">
        | Parameter                    | Type          | Description                                                                                                                                              |
        | ---------------------------- | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------- |
        | Tool ID                      | CognigyScript | Provide a meaningful name as a tool ID. This ID can contain only letters, numbers, underscores (`_`), or dashes (`-`). The default name is `send_email`. |
        | Description                  | CognigyScript | Provide a detailed description of what the tool should do. The default description is `Create and send a new email message.`                             |
        | Recipient TO Email Addresses | CognigyScript | A comma-separated list of email addresses to which the email will be sent.                                                                               |
      </Accordion>

      <Accordion title="Debug Settings">
        | Parameter                 | Type   | Description                                                                                                       |
        | ------------------------- | ------ | ----------------------------------------------------------------------------------------------------------------- |
        | Debug Message when called | Toggle | Enable the output of a debug message when the tool is called to provide detailed information about the tool call. |
      </Accordion>

      <Accordion title="Advanced">
        | Parameter | Type          | Description                                                                                                                                                                                                                                                                                                                                                                        |
        | --------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
        | Condition | CognigyScript | The tool will be enabled only if the condition is evaluated as true. If false, the tool isn't part of the AI Agent's Tools within this execution. For example, when using the `unlock_account` tool, you can specify a condition like `context.accountStatus === "locked"`. This checks the value in the context, and if it is missing or different, the tool will not be enabled. |
      </Accordion>
    </AccordionGroup>
  </Tab>

  <Tab title="Handover to AI Agent">
    The Handover to AI Agent tool lets you transfer a conversation to another AI Agent in the same or a different Flow. This approach ensures that different AI Agents share the same conversation context and can handle specific tasks, and that multi-step conversations run smoothly across Flows.

    <AccordionGroup>
      <Accordion title="Tool">
        | Parameter   | Type          | Description                                                                                                                                                                                                                                                                                       |
        | ----------- | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
        | Tool ID     | CognigyScript | Provide a meaningful name as a tool ID. This ID can contain only letters, numbers, underscores (`_`), or dashes (`-`). The default name is `handover_to_ai_agent`.                                                                                                                                |
        | Description | CognigyScript | Describe what the receiving AI Agent should do. For example: `Handle product recommendations`, `Perform technical troubleshooting`, or `Assist with billing questions`. Be specific so the handover rules clearly indicate when the conversation should be handed over to the receiving AI Agent. |
        | Select Flow | Selector      | Select the target Flow to hand over to. The list includes only Flows that contain AI Agent Nodes.                                                                                                                                                                                                 |
        | Select Node | Selector      | Select the AI Agent Node to hand over to.                                                                                                                                                                                                                                                         |
      </Accordion>

      <Accordion title="Debug Settings">
        | Parameter                 | Type   | Description                                                                                                       |
        | ------------------------- | ------ | ----------------------------------------------------------------------------------------------------------------- |
        | Debug Message when called | Toggle | Enable the output of a debug message when the tool is called to provide detailed information about the tool call. |
      </Accordion>

      <Accordion title="Advanced">
        | Parameter | Type          | Description                                                                                                                                                                                                                                                                                                                                                                        |
        | --------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
        | Condition | CognigyScript | The tool will be enabled only if the condition is evaluated as true. If false, the tool isn't part of the AI Agent's Tools within this execution. For example, when using the `unlock_account` tool, you can specify a condition like `context.accountStatus === "locked"`. This checks the value in the context, and if it is missing or different, the tool will not be enabled. |
      </Accordion>
    </AccordionGroup>
  </Tab>

  <Tab title="Handover to Human Agent">
    The Handover to Human Agent tool lets you transfer a conversation to a human agent through handover providers. This tool allows your AI Agents to refer complex tasks to a human agent if they can't address the user's request.

    <AccordionGroup>
      <Accordion title="Tool">
        | Parameter                        | Type          | Description                                                                                                                                                                                                                                                                                                                                            |
        | -------------------------------- | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
        | Tool ID                          | CognigyScript | Provide a meaningful name as a tool ID. This ID can contain only letters, numbers, underscores (`_`), or dashes (`-`). The default name is `handover_to_human_agent`.                                                                                                                                                                                  |
        | Description                      | CognigyScript | Describe the situation when the AI Agent should refer to a human agent. For example: `Use this tool if the user explicitly asks for a human.`, or `Use this tool if you fail to answer the user's question three times.`. Be specific so the handover rules clearly indicate when the conversation should be handed over to the receiving human agent. |
        | Handover Provider                | Selector      | Select the handover provider from the list and [configure the handover accordingly](/ai/escalate/handover-reference/overview). The list contains the handover providers created in **Deploy > Handover Providers**.                                                                                                                                    |
        | Handover Accepted Message        | CognigyScript | The message to be outputted when the handover is requested.                                                                                                                                                                                                                                                                                            |
        | Repeat Handover Accepted Message | Toggle        | Enable to repeatedly output the Handover Accepted Message if the user sends a message while waiting in the queue for an agent.                                                                                                                                                                                                                         |
      </Accordion>

      <Accordion title="Live Agent Settings">
        | Parameter                       | Description                                                                                                                                                                                                                                                                                                                                                                  |
        | ------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
        | Live Agent Inbox Id             | Enter the Inbox ID that you obtained in the [previous step](/ai/escalate/handover-reference/live-agent).                                                                                                                                                                                                                                                                     |
        | Skills                          | Enter a list of skills that should be used to filter the available agents. For example, `technical support, billing`.                                                                                                                                                                                                                                                        |
        | Languages                       | Enter a list of languages that should be used to filter the available agents. For example, `english, spanish`.                                                                                                                                                                                                                                                               |
        | Priority                        | Select a priority for the conversation from the list. For example, `Urgent`.                                                                                                                                                                                                                                                                                                 |
        | Get Queue Position              | Enable the system to retrieve the end user's current queue position. This parameter is inactive by default.                                                                                                                                                                                                                                                                  |
        | Update Interval                 | Specify the interval, in seconds, for how often to receive updates on the queue status. This parameter appears if **Get Queue Position** is selected.                                                                                                                                                                                                                        |
        | Get Estimated Wait Time         | Enable the retrieval of the estimated wait time for the end user. This parameter is inactive by default.                                                                                                                                                                                                                                                                     |
        | Update Interval                 | Set the interval, in seconds, for how often to receive updates on estimated wait time. This parameter appears if **Get Estimated Wait Time** is selected.                                                                                                                                                                                                                    |
        | Alternative Update              | The alternative update options when certain conditions are met. This parameter is inactive by default.                                                                                                                                                                                                                                                                       |
        | Maximum Queue Position          | Set the threshold for the maximum queue position. If this threshold is reached, the alternative text is sent. This parameter appears if **Alternative Update** is selected.                                                                                                                                                                                                  |
        | Maximum Estimated Wait Time     | Define the maximum estimated wait time in milliseconds. If this duration is exceeded, the alternative text is sent. This parameter appears if **Alternative Update** is selected.                                                                                                                                                                                            |
        | Alternative Text                | The message to be sent when either the maximum queue position or maximum wait time is reached. This parameter appears if **Alternative Update** is selected.                                                                                                                                                                                                                 |
        | Enable User Connects Message    | Notify human agents when an end user reconnects to the chat. The parameter is enabled by default. When the parameter is enabled, the message `User joined the conversation` appears in the chat as soon as the end user returns to the chat tab by clicking the **←** (back arrow) at the top bar in the browser, after having opened a new URL on the same tab as the chat. |
        | Enable User Disconnects Message | Notify human agents when an end user disconnects from the chat. The parameter is enabled by default. When the parameter is enabled, the message `User left the conversation` is sent as soon as the end user closes the tab with the chat or switches to a new URL address within the current tab.                                                                           |

        For more information, read the [Live Agent Handover Setup](/live-agent/getting-started/live-agent-setup/live-agent-setup-handover-flow) and [Real-Time Queue Notifications](/ai/escalate/handover-features/real-time-queue-notifications) documentation.
      </Accordion>

      <Accordion title="Chatwoot Settings">
        | Parameter         | Description                                                                                                                                                                                                                                                                                          |
        | ----------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
        | Chatwoot Inbox Id | Enter the Chatwoot Inbox ID. You can find the Inbox ID in your Chatwoot URL. For example, `https://app.chatwoot.com/app/account/6607/inbox/5446`, where the Inbox ID is `5446`. For more information, read the [Chatwoot handover provider](/ai/escalate/handover-reference/chatwoot) documentation. |
      </Accordion>

      <Accordion title="Salesforce Chat Settings">
        | Parameter                        | Description                                                                                                                                                                                                                                        |
        | -------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
        | Send Transcript as first message | Activate this parameter to allow Salesforce to receive the conversation transcript as the first message. This setting is turned off by default.                                                                                                    |
        | Button ID                        | Enter the ID of the Button you created in the Salesforce installation. This setting will override the default Button ID that you specified in the Endpoint configuration. You can use different Button IDs for individual Handover to Agent Nodes. |
        | Pre-chat Entities                | Enter the pre-chat entities. For more information, read the [Salesforce Handover](/ai/escalate/handover-reference/salesforce) documentation.                                                                                                       |
        | Pre-chat Details                 | Enter the pre-chat details. For more information, read the [Salesforce Handover](/ai/escalate/handover-reference/salesforce) documentation.                                                                                                        |
      </Accordion>

      <Accordion title="NiCE CXone Settings">
        | Parameter                        | Description                                                                                                                                                             |
        | -------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
        | Send Transcript as first message | Activate this parameter to allow NiCE CXone to receive the conversation transcript as the first message when the handover begins. This setting is turned on by default. |
      </Accordion>

      <Accordion title="Genesys Settings">
        | Parameter                       | Description                                                                                                                                                                                                                                                                                                                                                                  |
        | ------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
        | Language                        | Specify a language for the conversation. For example, `english`, `spanish`, `german`.                                                                                                                                                                                                                                                                                        |
        | Skills                          | Define skills for the conversation. For example, `escalation`.                                                                                                                                                                                                                                                                                                               |
        | Priority                        | Set the priority for the conversation. For example, `1`. <br /> If a priority is set, it triggers a flow in Genesys to prioritize or de-prioritize the conversation within the queue. <br /> Note that this functionality requires the appropriate flow to be set up in Genesys.                                                                                             |
        | Enable User Connects Message    | Notify human agents when an end user reconnects to the chat. The parameter is enabled by default. When the parameter is enabled, the message `User joined the conversation` appears in the chat as soon as the end user returns to the chat tab by clicking the **←** (back arrow) at the top bar in the browser, after having opened a new URL on the same tab as the chat. |
        | Enable User Disconnects Message | Notify human agents when an end user disconnects from the chat. The parameter is enabled by default. When the parameter is enabled, the message `User left the conversation` is sent as soon as the end user closes the tab with the chat or switches to a new URL address within the current tab.                                                                           |
        | Custom Attributes               | Add custom attributes in JSON format to include additional information the handover. For more information, refer to [Genesys Cloud Open Messaging](/ai/escalate/handover-reference/genesys-cloud-open-messaging).                                                                                                                                                            |
        | Get Estimated Wait Time         | Retrieve the estimated wait time for the end user. End users then receive wait time updates while they wait to be connected to a human agent. This parameter is inactive by default. It requires an active [Amazon EventBridge Source integration](/ai/escalate/handover-reference/genesys-cloud-open-messaging-with-amazon-eventbridge).                                    |
        | Update Interval                 | Specify the interval, in seconds, for how often to receive updates on the estimated wait time. This parameter appears if **Get Estimated Wait Time** is selected.                                                                                                                                                                                                            |
        | Alternative Update              | Send an alternative message to the end user when the maximum estimated wait time is exceeded. This parameter is inactive by default. This parameter appears if **Get Estimated Wait Time** is selected.                                                                                                                                                                      |
        | Maximum Estimated Wait Time     | Define the maximum estimated wait time in milliseconds. If this duration is exceeded, the alternative text is sent instead of the actual wait time. This parameter appears if **Alternative Update** is selected.                                                                                                                                                            |
        | Alternative Text                | Specify the message to be sent when the maximum estimated wait time is exceeded. This parameter appears if **Alternative Update** is selected.                                                                                                                                                                                                                               |

        Genesys Cloud Open Messaging doesn't provide a queue position. Estimated wait time is the only queue parameter available.

        Example for the **Custom Attributes** parameter:

        ```json theme={null}
        {
          "section": "section A",
          "region": "Europe"
        }
        ```

        For more information,
        read
        the [Genesys Cloud Guest Chat](/ai/escalate/handover-reference/genesys-cloud-guest-chat#prerequisites), [Genesys Cloud Open Messaging](/ai/escalate/handover-reference/genesys-cloud-open-messaging), and [Real-Time Queue Notifications](/ai/escalate/handover-features/real-time-queue-notifications#configure-settings-in-genesys-cloud-open-messaging) documentation.
      </Accordion>

      <Accordion title="8x8 Settings">
        | Parameter           | Description                                                                                                                                                                                                                                                                      |
        | ------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
        | 8x8 Channel Id      | Specify the channel ID that you created during the 8x8 configuration. <br /> You can find this ID in the 8x8 admin console. Navigate to **Channels > Chat API**. In the channel list, find the channel that you created. Copy the **ID** from the channel row.                   |
        | 8x8 Queue Id        | Specify the header value required for authenticating API calls to the 8x8 system. <br /> You can find this ID in the 8x8 admin console. Navigate to **Channels > Chat API**. In the channel list, find the channel that you created. Copy the **Queue ID** from the channel row. |
        | 8x8 JSON properties | Extend 8x8 properties with custom JSON.                                                                                                                                                                                                                                          |

        For more information, read the [8x8 Handover Setup](/ai/escalate/handover-reference/8x8) documentation.
      </Accordion>

      <Accordion title="RingCentral Engage Settings">
        | Parameter                       | Type          | Description                                                                                                                                                                                                                                                                                                                                                                  |
        | ------------------------------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
        | Additional Category Ids         | CognigyScript | IDs to be passed to the RingCentral Engage handover provider.                                                                                                                                                                                                                                                                                                                |
        | Enable User Connects Message    | Toggle        | Notify human agents when an end user reconnects to the chat. The parameter is enabled by default. When the parameter is enabled, the message `User joined the conversation` appears in the chat as soon as the end user returns to the chat tab by clicking the **←** (back arrow) at the top bar in the browser, after having opened a new URL on the same tab as the chat. |
        | Enable User Disconnects Message | Toggle        | Notify human agents when an end user disconnects from the chat. The parameter is enabled by default. When the parameter is enabled, the message `User left the conversation` is sent as soon as the end user closes the tab with the chat or switches to a new URL address within the current tab.                                                                           |

        For more information, read the [RingCentral Engage Integration](/ai/escalate/handover-reference/ring-central-engage) documentation.
      </Accordion>

      <Accordion title="Cancel Handover Options">
        | Parameter          | Type          | Description                                                                       |
        | ------------------ | ------------- | --------------------------------------------------------------------------------- |
        | Cancel Intent      | CognigyScript | The Intent the user has to trigger to cancel the Handover request.                |
        | Cancel Button Text | CognigyScript | The text for a quick reply button that cancels the Handover request when clicked. |
      </Accordion>

      <Accordion title="On Resolve Options">
        | Parameter                           | Type     | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
        | ----------------------------------- | -------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
        | Flow Continuation                   | Selector | Controls where the Flow continues after the human agent leaves the conversation.<br />When the user needs assistance completing the Flow with the AI Agent, the conversation is handed over to a human agent. The human agent can trigger the Flow by sending a `/bot` command as a reply in the conversation interface.<br />For example, the user enters their date of birth in an incorrect format, and the AI Agent can't process the response. The conversation is handed over to a human agent, who identifies the incorrect format and sends the corrected date using `/bot MM/DD/YYYY`. The AI Agent takes over and continues the Flow.<br />You can select one of the following options to determine where the Flow continues:<ul><li>**Below this Node** — the Flow continues from the first Node below the Human Agent Node, regardless of where the human agent stopped assisting the user. This option is selected by default.</li> <li>**At current AgentInject Entrypoint** — the Flow continues from the Node where the human agent stopped assisting the user.</li></ul> |
        | Send resolve event to Virtual Agent | Toggle   | Controls whether the Flow receives the `conversation_status_changed` event with the status set to `resolved`. This parameter is activated by default.<ul><li>When the parameter is activated, the event is sent when the human agent leaves the conversation. As a result, the Flow continues immediately, without waiting for the user to send another message. If the [Lookup Case](/ai/agents/develop/node-reference/logic/lookup) Node is configured with **Handover Status** set to **Completed**, the Node can detect the completed handover and continue the Flow.</li><li>When the parameter is deactivated, the event isn't sent when the human agent leaves the conversation. The Flow continues after the user sends their next message.</li></ul>                                                                                                                                                                                                                                                                                                                             |
      </Accordion>

      <Accordion title="Event Settings">
        | Parameter         | Type   | Description                                             |
        | ----------------- | ------ | ------------------------------------------------------- |
        | Send Queue Event  | Toggle | Sends an event to the Flow once the handover is queued. |
        | Send Active Event | Toggle | Sends an event to the Flow once the handover is active. |
      </Accordion>

      <Accordion title="Debug Settings">
        | Parameter                 | Type   | Description                                                                                                       |
        | ------------------------- | ------ | ----------------------------------------------------------------------------------------------------------------- |
        | Debug Message when called | Toggle | Enable the output of a debug message when the tool is called to provide detailed information about the tool call. |
      </Accordion>

      <Accordion title="Advanced">
        | Parameter | Type          | Description                                                                                                                                                                                                                                                                                                                                                                        |
        | --------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
        | Condition | CognigyScript | The tool will be enabled only if the condition is evaluated as true. If false, the tool isn't part of the AI Agent's Tools within this execution. For example, when using the `unlock_account` tool, you can specify a condition like `context.accountStatus === "locked"`. This checks the value in the context, and if it is missing or different, the tool will not be enabled. |
      </Accordion>
    </AccordionGroup>
  </Tab>

  <Tab title="Execute Workflow Tool">
    The Execute Workflow tool lets the AI Agent execute another Flow. After the target Flow executes, the conversation returns to the AI Agent.

    The target Flow inherits the [Context object](/ai/agents/develop/ai-agent-memory/context) from the AI Agent. If the Context object is changed during the target Flow execution, these changes are also available to the AI Agent.

    | Parameter   | Description                                                                                                                                                                                                                                                                                     |
    | ----------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Tool ID     | Provide a meaningful name as a tool ID. This ID can contain only letters, numbers, underscores (`_`), or dashes (`-`). The default name is `execute_workflow`.                                                                                                                                  |
    | Description | Describe when to trigger the AI Agent to execute the target Flow. For example, if you have a specific Flow for dealing with package returns, enter: `Use this tool if the user explicitly asks to return a package.`. Be specific so the AI Agent only executes the target Flow when necessary. |
    | Target Flow | Select the target Flow that the AI Agent executes when this tool is called. Click <img alt="target-flow" /> to open the target Flow.                                                                                                                                                            |
    | Select Node | Select the target Node in the target Flow that the AI Agent executes. You must set the Node you want to select as an Entrypoint with the [Set Entrypoint](/ai/agents/develop/nodes/overview#set-nodes-as-entrypoints) setting before you can select it.                                         |

    <AccordionGroup>
      <Accordion title="Debug Settings">
        | Parameter                 | Type   | Description                                                                                                         |
        | ------------------------- | ------ | ------------------------------------------------------------------------------------------------------------------- |
        | Debug Message when called | Toggle | Activate the output of a debug message when the tool is called to provide detailed information about the tool call. |
      </Accordion>

      <Accordion title="Advanced">
        | Parameter      | Type   | Description                                                                                                                                                                                                                                                                                                                                                                                                                              |
        | -------------- | ------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
        | Parse Intents  | Toggle | Sets whether Intents should be parsed in the target Flow. This parameter is activated by default.<br /><br />If this parameter is activated, Cognigy NLU reparses the Intents in the target Flow based on `text` in the [Input object](/ai/agents/develop/ai-agent-memory/input#accessing-input-properties). Then, Cognigy NLU generates Intent scores using the Intents from the target Flow instead of the ones from the current Flow. |
        | Parse Slots    | Toggle | Sets whether Slots should be parsed in the target Flow. This parameter is activated by default.<br /><br />If this parameter is activated, Cognigy NLU reparses the Slots in the target Flow based on `text` in the [Input object](/ai/agents/develop/ai-agent-memory/input#accessing-input-properties). Then, Cognigy NLU generates Slot scores using the Slots from the target Flow instead of the ones from the current Flow.         |
        | Absorb Context | Toggle | Sets whether the [default Context object](/ai/agents/develop/ai-agent-memory/context#default-context) of the target Flow is used in the context of the Flow the AI Agent is currently in.<br /><br />If the default Context of the target Flow has a value, it is copied into the Context object of the AI Agent as soon as the target Flow is executed.                                                                                 |
      </Accordion>
    </AccordionGroup>
  </Tab>
</Tabs>

### Configure AI Agent Settings for Each Job

Configure more settings for managing AI Agents:

<AccordionGroup>
  <Accordion title="Memory Handling">
    | Parameter                   | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
    | --------------------------- | ------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Long-Term Memory Injection  | Selector      | Allow the AI Agent to access [Contact Profile](/ai/agents/analyze/contact-profiles) information for the current user. Select one of the following options:<ul><li>**None** – no memory.</li><li>**Inherit from AI Agent** – use the settings specified in the [AI Agent creation settings](/ai/agents/develop/manage-ai-agents#create-ai-agents).</li><li>**Inject full Contact Profile** – use all information from the Contact Profile.</li><li>**Inject Contact Memories only** – use information only from the **Memories** field in the Contact Profile.</li><li>**Inject selected Profile fields** – use information from specific fields in the Contact Profile.</li></ul> |
    | Selected Profile Fields     | Text          | This parameter appears when the **Inject selected Profile fields** option is enabled. Enter specific fields from the Contact Profile for targeted data use. Specify the field using the [Profile keys](/ai/agents/analyze/contact-profiles#profile-schema) format and press <kbd>Enter</kbd> to apply it.                                                                                                                                                                                                                                                                                                                                                                         |
    | Short-Term Memory Injection | CognigyScript | Specify a static string or a dynamic value via CognigyScript to make available to the AI Agent in the current turn.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
  </Accordion>

  <Accordion title="Grounding Knowledge">
    | Parameter                  | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
    | -------------------------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Knowledge Injection        | Selector      | Use the Knowledge AI feature for the AI Agent. Select one of the following options:<ul><li>**Never** — do not use the Knowledge Stores.</li><li>**When Required** — let the AI Agent decide when querying the Knowledge Stores is required to help the user.</li><li>**Once for Each User Input** — query the Knowledge Store(s) after each user input. Note that executing a query on every user input can lead to increased costs and latency.</li></ul>                                                                                                                                                                                                                                                                                                                                                                                                                                     |
    | Use AI Agent Knowledge     | Toggle        | The parameter appears when you select either **When Required** or **Once for Each User Input**. Enable to use the Knowledge Store configured in the AI Agent. The Knowledge Store configured within the [AI Agent creation settings](#create-ai-agents) will be used.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
    | Use Job Knowledge          | Toggle        | The parameter appears when you select either **When Required** or **Once for Each User Input**. Enable this option to configure a specific Knowledge Store for this particular job, allowing the AI Agent to access job-specific data or resources.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | Job Knowledge Store        | Selector      | The parameter appears when you select either **When Required** or **Once for Each User Input**. The parameter appears when the **Use Job Knowledge** option is enabled. Select a specific Knowledge Store for this AI Agent's job.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | Top K                      | Slider        | The parameter appears when you select either **When Required** or **Once for Each User Input**. Specify how many Knowledge Chunks to return. Providing more results gives the AI Agent additional context, but it also increases noise and token usage.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
    | Source Tags                | CognigyScript | The parameter appears when you select either **When Required** or **Once for Each User Input**. The tags serve to refine the scope of your knowledge search, allowing you to include only the most pertinent sections of the knowledge base and, as a result, improve the accuracy of search outputs. Before specifying tags, ensure that they were provided during the creation of the Knowledge Sources. Add Tags by specifying each Tag separately and pressing ++enter++. The maximum number of tags is 5. When you specify multiple Source Tags, the Search Extract Output Node defaults to an `AND` operator, meaning it only considers Sources that have all the specified Tags. This approach ensures the search results are precise and highly relevant to the end user's query. To change this behavior, go to the **Match Types for Source Tags** parameter.                        |
    | Match type for Source Tags | Select        | The parameter appears when you select either **When Required** or **Once for Each User Input**. The operator to filter Knowledge Sources by Source Tags. Select one of the following options:<ul><li>**AND** — the default value, requires all tags to match across multiple Knowledge Sources. Consider the following example: there are Knowledge Sources with Tags `S-a`, `S-b`, and `S-c`. When you use the `AND` operator to filter by `S-a` and `S-b`, only Sources with both Tags `S-a` and `S-b` will be included in the search results.</li><li>**OR** — requires at least one tag to match across multiple Knowledge Sources. Consider the following example: there are Knowledge Sources with Tags `S-a`, `S-b`, and `S-c`. When you use the `OR` operator to filter by `S-a` or `S-b`, any Source with either Tag `S-a` or `S-b` will be included in the search results.</li></ul> |
    | Generate Search Prompt     | Toggle        | The parameter appears when you select **Once for Each User Input**. This parameter is enabled by default and allows you to generate a context-aware search prompt before executing the knowledge search. Note that enabling this parameter may lead to increased cost and latency.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
  </Accordion>

  <Accordion title="Storage and Streaming Options">
    | Parameter                   | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
    | --------------------------- | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | How to handle the result    | Select        | Determine how to handle the prompt result:<ul><li>**Store in Input** — stores the AI Agent result in the Input object. To print the prompt result, refer to the configured Context key in a Say Node or enable the **Output result immediately** option.</li><li>**Store in Context** — stores the result in the Context object. To print the prompt result, refer to the configured Context key in a Say Node or enable the **Output result immediately** option.</li><li>**Stream to Output** — streams the result directly into the output. This means that chunks coming from the prompt response will be output directly into the conversation chat as soon as a Stream Buffer Flush Token is matched, and you don't need to use the AI Agent Output Token and Say Node. By default, this result won't be stored in either the Input or the Context. You can change this behavior by activating the **Store Copy in Input** option.</li></ul> |
    | Input Key to store Result   | CognigyScript | The parameter appears when you select either **Store in Input** or **Stream to Output**. The result is stored in the `input.aiAgentOutput` object by default. You can specify another value, but the **AI Agent Output** Token will not work if the value is changed.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |
    | Context Key to store Result | CognigyScript | The parameter appears when **Store in Context** is selected. The result is stored in the `context.aiAgentOutput` object by default. You can specify another key.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
    | Stream Buffer Flush Tokens  | Text Array    | The parameter appears when **Stream to Output** is selected. It defines tokens that trigger the stream buffer to flush to the output. The tokens can be punctuation marks or symbols, such as `\n`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
    | Output result immediately   | Toggle        | The parameter appears when you select either **Store in Input** or **Store in Context**. This parameter allows you to output results immediately without using the Say Node and AI Agent Output token.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | Store Copy in Input         | Toggle        | The parameter appears when **Stream to Output** is selected. In addition to streaming the result to the output, store a copy in the Input object by specifying a value in the **Input Key to store Result** field.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
  </Accordion>

  <Accordion title="Voice">
    | Parameter                 | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | ------------------------- | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Voice Setting             | Select        | Configure the voice settings for the AI Agent Job. This parameter determines how the AI Agent selects the voice for text-to-speech (TTS) output. Select one of the following options:<ul><li>- **Inherit from AI Agent** — use the voice settings defined in the [AI Agent creation settings](#create-ai-agents).</li><li>- **Use Job Voice** – apply custom voice settings specific to this job, allowing the AI Agent to adapt to the particular role it performs.</li></ul> For example, if you create a marketing AI Agent, the voice can be more engaging, friendly, and persuasive. However, if the same AI Agent performs a different role, such as customer support, the voice might be more neutral, empathetic, and formal. |
    | TTS Vendor                | Dropdown      | Select a TTS vendor from the list or add a custom one. Note that the AI Agent Node doesn't support **TTS Labels** to distinguish configurations from the same TTS vendor. To use **TTS Labels**, add a [Set Session Config Node](/ai/agents/develop/node-reference/voice/voice-gateway/set-session-config) before the AI Agent Node in the Flow editor.                                                                                                                                                                                                                                                                                                                                                                               |
    | Custom (Vendor)           | CognigyScript | The **Custom** parameter appears when you select **Custom** from the **TTS Vendor** list. Specify the [custom TTS Vendor](/voice-gateway/webapp/speech-services#add-custom-speech-vendors). For preinstalled providers, use all lowercase letters, for example, `microsoft`, `google`, `aws`. For custom providers, use the name that you specified on the [Speech Service](/voice-gateway/webapp/speech-services) page in the Voice Gateway Self-Service Portal.                                                                                                                                                                                                                                                                     |
    | TTS Language              | Dropdown      | Define the language of the voice AI Agent output. Ensure this language aligns with the preferred language of the end user.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | Custom (Language)         | CognigyScript | The **Custom** parameter appears when you select **Custom** from the **TTS Language** list. Specify the language of the AI Agent output. The format depends on the option selected in the TTS vendor; check your TTS vendor documentation. The typical format is as follows: `de-DE`, `fr-FR`, `en-US`.                                                                                                                                                                                                                                                                                                                                                                                                                               |
    | TTS Voice                 | Dropdown      | Define the voice that should be used for the voice AI Agent output. This parameter allows you to customize the AI Agent's voice by defining its tone, gender, style, and regional specifics, making conversations more personalized and aligned with your brand and target audience.                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
    | Custom (Voice)            | CognigyScript | The **Custom** parameter appears when you select **Custom** from the **TTS Voice** list. Use this parameter to specify a custom voice, which is often required for region-specific voices. The format depends on the option selected in **TTS Vendor** and typically follows the pattern `language-region-VoiceName`. For example, `de-DE-ConradNeural` for German (Germany) male voice or `en-US-JennyNeural` for English (US) female voice.                                                                                                                                                                                                                                                                                         |
    | TTS Label                 | CognigyScript | The alternative name of the TTS vendor is the one you [specify in the Voice Gateway Self-Service Portal](/voice-gateway/webapp/applications#add-additional-tts-and-stt-vendor). If you have created multiple speech services from the same vendor, use the label to specify which service to use.                                                                                                                                                                                                                                                                                                                                                                                                                                     |
    | Disable TTS Audio Caching | Toggle        | Disables TTS audio caching. By default, the setting is deactivated. In this case, previously requested TTS audio results are stored in the AI Agent cache. When a new TTS request is made and the audio text has been previously requested, the AI Agent retrieves the cached result instead of sending another request to the TTS provider. When the setting is activated, the AI Agent caches TTS results but doesn't use them. In this case, each request is directly sent to your speech provider. Note that disabling caching can increase TTS costs. For detailed information, contact your speech provider.                                                                                                                    |
  </Accordion>

  <Accordion title="Tool Settings">
    | Parameter       | Type     | Description                                                                                                                                                                                                                                                                                                                         |
    | --------------- | -------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Tool Choice     | Selector | If supported by your LLM Model, this will determine how tools should be selected by the AI Agent:<ul><li>**Auto** — tools (or none) are automatically selected by the AI Agent when needed.</li><li>**Required** — your AI Agent will always use one of its Tools.</li><li>**None** — your AI Agent won't use a tool.</li></ul>     |
    | Use Strict mode | Toggle   | When the parameter is enabled, strict mode (if supported by the LLM provider) ensures that the arguments passed to a tool call precisely match the expected parameters. Enabling this feature can help prevent errors. However, it may cause a slight delay in the response, especially during the first call after making changes. |
  </Accordion>

  <Accordion title="Image Handling">
    | Parameter            | Type     | Description                                                                                                                                                                                                                                                                                                                                |
    | -------------------- | -------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Process Images       | Toggle   | Enable the AI Agent to read and understand images attachments. Make sure that your LLM provider supports image processing; refer to your provider's documentation. In addition, make sure that attachments are supported by and activated in your Endpoint, for example, Webchat.                                                          |
    | Images in Transcript | Selector | Configure how images older than the last turn are handled to reduce token usage: <ul><li>**Minify** — reduces the size of these images to 512x512px.</li><li>**Drop** — excludes the images.</li><li>**Keep** — sends the max size (this option consumes more tokens).</li></ul> Limitations and token consumption depend on the LLM used. |
  </Accordion>

  <Accordion title="Advanced">
    | Parameter                  | Type     | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
    | -------------------------- | -------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | LLM                        | Selector | Select a model that supports the [AI Agent Node feature](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature). The selected **Default** model is the model that you specified in **Settings > Generative AI Settings** of your Project. Select the model that you [added earlier](/ai/agents/overview#prerequisites) while configuring Agentic AI feature. This model will manage your AI Agent.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
    | AI Agent Base Version      | Selector | Select the base version of the AI Agent to use:<ul><li>**Fixed Version** — select a specific version, such as `1.0`, to ensure stability and avoid potential breaking changes. Use this version in production environments or for critical workflows. The version dropdown will be updated as future versions of the AI Agent Node are released.</li><li>**Latest** — use the most recent version of the AI Agent Node. While this version ensures access to the latest features, it may cause breaking changes that require manual updates.</li></ul>When upgrading to a fixed version or switching to the latest, always test your AI Agent carefully to ensure it works with the selected version.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
    | Timeout                    | Number   | Define the maximum number of milliseconds to wait for a response from the LLM provider.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
    | Maximum Completion Tokens  | Slider   | Define the maximum number of tokens that can be used during a process to manage costs. However, if the limit is set too low, the output may be incomplete, as the process could be cut off before it finishes. For example, if you set the maximum tokens to 100, the model will stop generating content once it reaches 100 tokens. This number would be roughly equal to 100 words, depending on the language and tokenization method.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
    | Temperature                | Slider   | Define the sampling temperature, which ranges between 0 and 1. Higher values, such as 0.8, make the output more random, while lower values, such as 0.2, make it more focused and deterministic.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | Include Rich Media Context | Toggle   | Controls whether the text from the [Textual Description](/ai/agents/develop/node-reference/basic/say#output-type) field in Say, Question, and Optional Questions Nodes is added to the AI Agent's prompt. This text provides AI Agents with additional information, improving their responses.<br /><br />If you have filled in a **Textual Description** field in other Nodes, this text is added to the prompt as context for rich media, such as Text with Quick Replies. If the **Textual Description** field is empty, the button titles and alt text from the rich media are added to the prompt. By default, the **Include Rich Media Context** parameter is active. When this parameter is inactive, nothing related to the rich media is added to the prompt.<br /><br />**Examples**:<ul><li>If **Textual Description** is filled in:<p>Textual Description: `Select your preferred delivery option: Standard Delivery or Express Delivery`.</p><p>Quick Replies buttons: `Standard Delivery`, `Express Delivery`.</p><p>Context added to the prompt: `Select your preferred delivery option: Standard Delivery or Express Delivery`.</p></li><li>If **Textual Description** is empty:<p>Textual Description: empty.</p><p>Quick Replies buttons: `Standard Delivery`, `Express Delivery`.</p><p>Context added to the prompt: `Standard Delivery`, `Express Delivery`.</p></li><li>If **Include Rich Media Context** is inactive:<p>No information related to the rich media is added to the prompt.</p></li></ul> |
  </Accordion>

  <Accordion title="Error Handling">
    | Parameter                      | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                              |
    | ------------------------------ | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Log to System Logs             | Toggle        | Log errors to the system logs. They can be viewed on the [Logs](/ai/agents/test/logs) page of your Project. The parameter is inactive by default.                                                                                                                                                                                                                                                                        |
    | Store in Input                 | Toggle        | Store errors in the Input object.                                                                                                                                                                                                                                                                                                                                                                                        |
    | Select Error Handling Approach | Select        | You can select one of the Error Handling options:<ul><li>**Stop Flow Execution** — terminate the current Flow execution.</li><li>**Continue Flow Execution** — allow the Flow to continue executing, bypassing the error and proceeding to the next steps.</li><li>**Go to Node** — redirect the workflow to a specific Node in the Flow, which can be useful for error recovery or customized error handling.</li></ul> |
    | Select Flow                    | Select        | The parameter appears when **Go to Node** is selected. Select a Flow from the available options.                                                                                                                                                                                                                                                                                                                         |
    | Select Node                    | Select        | The parameter appears when **Go to Node** is selected. Select a Node from the available options.                                                                                                                                                                                                                                                                                                                         |
    | Error Message (optional)       | CognigyScript | Add the optional message to the output if the AI Agent Node fails.                                                                                                                                                                                                                                                                                                                                                       |
  </Accordion>

  <Accordion title="Debug Settings">
    | Parameter             | Type   | Description                                                                                                                                                                                                                                                                                                                                           |
    | --------------------- | ------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Log Job Execution     | Toggle | Send a debug message with the current AI Agent Job configuration. The message appears in the Interaction Panel when [debug mode](/ai/agents/test/interaction-panel/chat#debug-mode) is enabled. The parameter is active by default.                                                                                                                   |
    | Log Knowledge Results | Toggle | Send a debug message containing the result from a knowledge search. The message appears in the Interaction Panel when [debug mode](/ai/agents/test/interaction-panel/chat#debug-mode) is enabled. The parameter is inactive by default.                                                                                                               |
    | Show Token Count      | Toggle | Send a debug message containing the input, output, and total token count. The message appears in the Interaction Panel when [debug mode](/ai/agents/test/interaction-panel/chat#debug-mode) is enabled. Cognigy uses the GPT-3 tokenizer algorithm, so actual token usage may vary depending on the model used. The parameter is inactive by default. |
    | Log System Prompt     | Toggle | Send a debug message containing the system prompt. The message appears in the Interaction Panel when [debug mode](/ai/agents/test/interaction-panel/chat#debug-mode) is enabled. The parameter is inactive by default.                                                                                                                                |
    | Log LLM Latency       | Toggle | Send a debug message containing key latency metrics for the request to the model, including the time taken for the first output and the total time to complete the request. The message appears in the Interaction Panel when [debug mode](/ai/agents/test/interaction-panel/chat#debug-mode) is enabled. The parameter is inactive by default.       |
    | Log Tool Definitions  | Toggle | Send a debug message containing information about the configured AI Agent tools. The message appears in the Interaction Panel when [debug mode](/ai/agents/test/interaction-panel/chat#debug-mode) is enabled. The parameter is inactive by default.                                                                                                  |

    To test and refine the job settings, use the Interaction Panel on the right side of the screen.

    To view the AI Agent's Flow in detail, click **Advanced Editor** in the upper-right corner.

    One AI Agent can have multiple jobs. To add another job, click **Add Job** in the bottom-left corner. Note that each job corresponds to one Flow.
  </Accordion>
</AccordionGroup>

## Other Operations

<Tabs>
  <Tab title="GUI">
    You can view and delete AI Agents, and copy their Reference ID in **Build > AI Agents**.
  </Tab>

  <Tab title="API">
    You can create, edit, and delete AI Agents using [Cognigy API](https://api-trial.cognigy.ai/openapi#get-/beta/aiagents).
  </Tab>

  <Tab title="CLI">
    You can clone and edit AI Agents using [Cognigy CLI](https://github.com/Cognigy/Cognigy-CLI).
  </Tab>
</Tabs>

## More Information

* [Overview](/ai/agents/overview)
* [Getting Started](/ai/agents/quickstart)


# Copilot: Identity Tile
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/agent-copilot/identity-tile



<a href="/release-notes/4.60"><Badge>Updated in 4.60</Badge></a>

<Frame>
  <img alt="Copilot Identity Tile Node configuration panel" />
</Frame>

## Description

This Node is designed to display and process customer-specific information directly within the widget of your Agent Copilot workspace, enhancing the context and usability of the identity-related data.

## Parameters

| Parameter     | Type          | Description                                                                                                                                 |
| ------------- | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------- |
| Tile ID       | CognigyScript | The ID that you created in the Agent Copilot configuration.                                                                                 |
| Widget Layout | Select        | The layout of how the widget should be placed in the Workspace (horizontal/vertical).                                                       |
| Image Shape   | Select        | The shape or format of the image (round or square).                                                                                         |
| Image URL     | URL           | The URL where the image is hosted or located.                                                                                               |
| Customer Data | CognigyScript | Additional customer-specific data for analysis in a key-value format. For example, the key is `Email` and the value is `{{profile.email}}`. |

## Use Case

Displaying the user's profile picture adds a human touch to the conversation, making it more engaging and personal.
One way to enhance the conversation further is by showing key customer data,
such as their name, account information, or preferences, directly within the chat conversation.
This allows human agents to have contextually relevant discussions without the need to search for customer data in multiple tools.

<Frame>
  <img alt="Identity Tile widget displaying customer profile data in Agent Copilot" />
</Frame>

## More Information

* [Copilot: Transcript Tile](/ai/agents/develop/node-reference/agent-copilot/transcript-tile)
* [Copilot: Sentiment Tile Node](/ai/agents/develop/node-reference/agent-copilot/sentiment-tile)
* [Copilot: Next Action Tile Node](/ai/agents/develop/node-reference/agent-copilot/next-action-tile)
* [Copilot Nodes](/ai/agents/develop/node-reference/agent-copilot/overview)
* [Copilot](/agent-copilot/overview)


# Copilot: Knowledge Tile
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/agent-copilot/knowledge-tile



<a href="/release-notes/4.72"><Badge>Updated in 4.72</Badge></a>

<Frame>
  <img alt="Copilot Knowledge Tile Node configuration panel" />
</Frame>

## Description

This Node provides potential answers to human agents regarding customer questions.
The Node creates an [Agent Copilot widget](/agent-copilot/configure/nodes) that consists of a search bar and a response form.
Additionally, it provides meta-source links,
information about the files from which knowledge was derived and hints for further questions.
The human agent uses the search bar to access a knowledge base and find information to respond to customer queries.
Responses are generated based on a Large Language Model (LLM) via Knowledge AI.

Before using the Copilot: Knowledge Tile Node, configure the LLM provider in the [Settings](/ai/agents/develop/gen-ai-and-llms/generative-ai) and select the appropriate model from the [supported model list](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature).
This Node supports models with the LLM Prompt Node and the Search Extract Output Node features.

To set up the Copilot: Knowledge Tile Node, you need two Nodes for the Flow configuration: one to initiate the knowledge search process and the second one to execute it.
Additionally, to prevent the content of other widgets from updating during the Node's operation,
move the Nodes under the Else child Node of the If Node.
For more information, refer to the [Configuration](#configuration) section.

## Parameters

<AccordionGroup>
  <Accordion title="General Settings">
    | Parameter       | Type          | Description                                                          |
    | --------------- | ------------- | -------------------------------------------------------------------- |
    | Tile ID         | CognigyScript | The ID that you created in the Agent Copilot configuration.          |
    | Knowledge Store | Select        | Select a store from the list to allow the Node to retrieve the data. |
  </Accordion>

  <Accordion title="Enable Copy to Clipboard Button">
    If you are using Live Agent as a handover provider, you can forward information from the widget to the **Reply to User** section by clicking <img alt="forward" /> and then sending it to the end user.

    For other providers, use the alternative option **Enable copy to clipboard button** to activate the Copy button on the tile.
    For Salesforce, to activate the Copy button, update your Salesforce integration using [this guide](https://github.com/Cognigy/salesforce-integrations?tab=readme-ov-file#update).
  </Accordion>

  <Accordion title="Search Settings">
    | Parameter                 | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                          |
    | ------------------------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Context-Aware Search      | Select        | Considering the transcript's context for the search. It allows an AI Agent to answer follow-up questions. You can select one of the following options:<ul><li>**On** — is the default option. When this parameter is enabled, it affects tokenization costs, which impact the amount of memory, computational power, and overall expenses needed for running an OpenAI or Azure OpenAI model.</li><li>**Off** — the setting is turned off.</li></ul> |
    | Transcript Steps          | Indicator     | Number of prior transcript steps to use in a context-aware search. This parameter is active when the Context-Aware Search setting is enabled.                                                                                                                                                                                                                                                                                                        |
    | Where to store the result | Select        | The storage for the search results. Select a storage:<ul><li>**default** — the content will be stored in the Input object (`input.knowledgeSearch`).</li><li>**Input** — the content will be stored in the Input object.</li><li>**Context** — the content will be stored in the Context.</li></ul>                                                                                                                                                  |
    | String to search for      | CognigyScript | The string to search for in the Knowledge Store. A default value is stored in `input.text`.                                                                                                                                                                                                                                                                                                                                                          |
    | Source Tags               | CognigyScript | The tags serve to refine the scope of your knowledge search, allowing you to include only the most pertinent sections of the knowledge base and, as a result, improve the accuracy of search outputs. <br /><br /> Before specifying tags, ensure that they were provided during the creation of the Knowledge Sources. Add tags by specifying each tag separately and pressing **enter**. The maximum number of tags is 5.                          |

    <Accordion title="Source Tags">
      * You can add CognigyScript within Source Tags.
      * Source Tags are always lower-case, only alpha-numerics (a-z, 0-9), underscores, and hyphens are allowed.
      * When using CognigyScript for Source Tags, it is important to lowercase them. For example, by using the following: `{{input.text.toLowerCase()}}`.
    </Accordion>
  </Accordion>

  <Accordion title="Error Handling">
    | Parameter                      | Type   | Description                                                                                                                                                                                                                                                                                                                                                                                                              |
    | ------------------------------ | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Select Error Handling Approach | Select | You can select one of the Error Handling options:<ul><li>**Stop Flow Execution** — terminate the current Flow execution.</li><li>**Continue Flow Execution** — allow the Flow to continue executing, bypassing the error and proceeding to the next steps.</li><li>**Go to Node** — redirect the workflow to a specific Node in the Flow, which can be useful for error recovery or customized error handling.</li></ul> |
    | Error Message (optional)       | Text   | Active when the **Continue Flow Execution** option is selected. Optional message to output if the Search or Extract service fails.                                                                                                                                                                                                                                                                                       |
    | Select Flow                    | Select | Active when the **Go to Node** option is selected. Select a Flow from the available options.                                                                                                                                                                                                                                                                                                                             |
    | Select Node                    | Select | Active when the **Go to Node** option is selected. Select a Node from the available options.                                                                                                                                                                                                                                                                                                                             |
  </Accordion>
</AccordionGroup>

## Configuration

<Accordion title="Configure the Copilot: Knowledge Tile widget">
  To use the Copilot: Knowledge AI widget, configure the following Flow:

  1. Add an **If** Node.
  2. Open the **If** Node editor.
  3. In the **Condition** field, specify the following condition: `input.data._cognigy._agentAssist.payload.tileId === "tile-id"`, where `tile-id` is a Knowledge tile ID that you created in the Agent Copilot configuration via API.
  4. In the **Knowledge Store** field, select the source that you want to use for searching knowledge.
  5. Below the **Then** child Node, place the **Copilot: Knowledge Tile** Node.
  6. Below the **Else** child Node, add the other Copilot tile Nodes. This step is necessary to ensure that the content of these Nodes does not get updated during the operation of the **Copilot: Knowledge Tile** Node.
  7. Below the **Copilot: Knowledge Tile** Node and other tile Nodes, place the **Once** child Node.
  8. Below the **On First Time** child Node, add a duplicate of the **Copilot: Knowledge Tile** Node.

  <Check>
    If the Flow is not working or issues arise with the Knowledge Source, try creating a new Agent and running the Flow in it.
  </Check>

  <Frame>
    <img alt="Knowledge Tile Flow configuration with If, Once, and Knowledge Tile Nodes" />
  </Frame>

  Test your Flow by using a [Demo Webchat](/webchat/demo) in the Webchat Endpoint. You can use Live Agent as a handover provider or any other available provider.

  <Frame>
    <img alt="Knowledge Tile widget showing search results in Agent Copilot" />
  </Frame>
</Accordion>

## More Information

* [Copilot: Transcript Tile](/ai/agents/develop/node-reference/agent-copilot/transcript-tile)
* [Copilot: Sentiment Tile Node](/ai/agents/develop/node-reference/agent-copilot/sentiment-tile)
* [Copilot: Identity Tile](/ai/agents/develop/node-reference/agent-copilot/identity-tile)
* [Agent Copilot Nodes](/ai/agents/develop/node-reference/agent-copilot/overview)
* [Agent Copilot](/agent-copilot/overview)


# Copilot: Next Action Tile
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/agent-copilot/next-action-tile



<a href="/release-notes/4.72"><Badge>Updated in 4.72</Badge></a>

<Frame>
  <img alt="Copilot Next Action Tile Node configuration panel" />
</Frame>

## Description

This Node is designed to determine the appropriate next step or action in a conversation or process based on the user's input or the current context.
This Node helps guide a conversation flow and ensures that the system responds effectively to user requests or queries.

## Parameters

<AccordionGroup>
  <Accordion title="General Settings">
    | Parameter       | Type          | Description                                                                                                                                                                                                                  |
    | --------------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Tile ID         | CognigyScript | The ID created in the Agent Copilot configuration.                                                                                                                                                                           |
    | Type            | Select        | Choose from the following types: <ul><li>**Text** — the text you want to display in the next action widget.</li><li>**Knowledge Assist** — the text from the Knowledge Store to display in the next action widget.</li></ul> |
    | Text            | CognigyScript | Enter the text you want to display in the next action widget.                                                                                                                                                                |
    | Knowledge Store | Select        | Choose a store from the list to allow the Node to retrieve the data.                                                                                                                                                         |
  </Accordion>

  <Accordion title="Enable Copy to Clipboard Button">
    If you are using Live Agent as a handover provider, you can forward information from the widget to the **Reply to User** section by clicking <img alt="forward" /> and then sending it to the end user.

    For other providers, use the alternative option **Enable copy to clipboard button** to activate the Copy button on the tile.
    For Salesforce, to activate the Copy button, update your Salesforce integration using [this guide](https://github.com/Cognigy/salesforce-integrations?tab=readme-ov-file#update).
  </Accordion>

  <Accordion title="Search Settings">
    | Parameter                 | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                          |
    | ------------------------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Context-Aware Search      | Select        | Considering the transcript's context for the search. It allows an AI Agent to answer follow-up questions. You can select one of the following options:<ul><li>**On** — is the default option. When this parameter is enabled, it affects tokenization costs, which impact the amount of memory, computational power, and overall expenses needed for running an OpenAI or Azure OpenAI model.</li><li>**Off** — the setting is turned off.</li></ul> |
    | Transcript Steps          | Indicator     | Number of prior transcript steps to use in a context-aware search. This parameter is active when the Context-Aware Search setting is enabled.                                                                                                                                                                                                                                                                                                        |
    | Where to store the result | Select        | The storage for the search results. Select a storage:<ul><li>**default** — the content will be stored in the Input object (`input.knowledgeSearch`).</li><li>**Input** — the content will be stored in the Input object.</li><li>**Context** — the content will be stored in the Context.</li></ul>                                                                                                                                                  |
    | String to search for      | CognigyScript | The string to search for in the Knowledge Store. A default value is stored in `input.text`.                                                                                                                                                                                                                                                                                                                                                          |
    | Source Tags               | CognigyScript | The tags serve to refine the scope of your knowledge search, allowing you to include only the most pertinent sections of the knowledge base and, as a result, improve the accuracy of search outputs. <br /><br /> Before specifying tags, ensure that they were provided during the creation of the Knowledge Sources. Add tags by specifying each tag separately and pressing **enter**. The maximum number of tags is 5.                          |

    <Accordion title="Source Tags">
      * You can add CognigyScript within Source Tags.
      * Source Tags are always lower-case, only alpha-numerics (a-z, 0-9), underscores, and hyphens are allowed.
      * When using CognigyScript for Source Tags, it is important to lowercase them. For example, by using the following: `{{input.text.toLowerCase()}}`.
    </Accordion>
  </Accordion>

  <Accordion title="Error Handling">
    | Parameter                      | Type   | Description                                                                                                                                                                                                                                                                                                                                                                                                              |
    | ------------------------------ | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Select Error Handling Approach | Select | You can select one of the Error Handling options:<ul><li>**Stop Flow Execution** — terminate the current Flow execution.</li><li>**Continue Flow Execution** — allow the Flow to continue executing, bypassing the error and proceeding to the next steps.</li><li>**Go to Node** — redirect the workflow to a specific Node in the Flow, which can be useful for error recovery or customized error handling.</li></ul> |
    | Error Message (optional)       | Text   | Active when the **Continue Flow Execution** option is selected. Optional message to output if the Search or Extract service fails.                                                                                                                                                                                                                                                                                       |
    | Select Flow                    | Select | Active when the **Go to Node** option is selected. Select a Flow from the available options.                                                                                                                                                                                                                                                                                                                             |
    | Select Node                    | Select | Active when the **Go to Node** option is selected. Select a Node from the available options.                                                                                                                                                                                                                                                                                                                             |
  </Accordion>
</AccordionGroup>

## Use Case

This Node helps the human agent by providing guidance and predefined instructions based on the user's question.
For instance, if a user asks about the latest release version, the AI Agent can give them both the answer and links to important updates in that release.

<Frame>
  <img alt="Next Action Tile widget with guidance and links in Agent Copilot" />
</Frame>

## More Information

* [Copilot: Transcript Tile](/ai/agents/develop/node-reference/agent-copilot/transcript-tile)
* [Copilot: Sentiment Tile Node](/ai/agents/develop/node-reference/agent-copilot/sentiment-tile)
* [Copilot: Identity Tile](/ai/agents/develop/node-reference/agent-copilot/identity-tile)
* [Agent Copilot Nodes](/ai/agents/develop/node-reference/agent-copilot/overview)
* [Agent Copilot](/agent-copilot/overview)


# Agent Copilot Nodes
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/agent-copilot/overview



<a href="/release-notes/4.66"><Badge>Updated in 4.66</Badge></a>

The Agent Copilot Nodes allow users to customize the Agent Copilot workspace by adding HTML, Adaptive Cards, and IFrame content.

<CardGroup>
  <Card title="Copilot: Adaptive Card Tile" href="/ai/agents/develop/node-reference/agent-copilot/set-adaptive-card-tile">
    This Node allows you to create and display adaptive cards in your Agent Copilot workspace.
  </Card>

  <Card title="Copilot: HTML Tile" href="/ai/agents/develop/node-reference/agent-copilot/set-html-tile">
    This Node allows you to display customized HTML content in the Agent Copilot workspace.
  </Card>

  <Card title="Copilot: IFrame Tile" href="/ai/agents/develop/node-reference/agent-copilot/set-iframe-tile">
    This Node allows you to embed external websites directly into the Agent Copilot workspace.
  </Card>

  <Card title="Copilot: Identity Tile" href="/ai/agents/develop/node-reference/agent-copilot/identity-tile">
    This Node displays and processes customer-specific information directly within the widget of your Agent Copilot workspace.
  </Card>

  <Card title="Copilot: Knowledge Tile" href="/ai/agents/develop/node-reference/agent-copilot/knowledge-tile">
    This Node provides potential answers to human agents regarding customer questions.
  </Card>

  <Card title="Copilot: Next Action Tile" href="/ai/agents/develop/node-reference/agent-copilot/next-action-tile">
    This Node determines the appropriate next step or action in a conversation or process based on the user's input or the current context.
  </Card>

  <Card title="Copilot: Send Data" href="/ai/agents/develop/node-reference/agent-copilot/send-data">
    This Node enables real-time updates of dynamic content to Agent Copilot widgets without reloading the entire widget.
  </Card>

  <Card title="Copilot: Sentiment Tile" href="/ai/agents/develop/node-reference/agent-copilot/sentiment-tile">
    This Node performs sentiment analysis on the most recent input and provides insights into the emotional tone of the user.
  </Card>

  <Card title="Copilot: Set Grid" href="/ai/agents/develop/node-reference/agent-copilot/set-grid">
    This Node updates the Agent Copilot workspace grid via an [API POST request](https://api-trial.cognigy.ai/openapi#post-/v2.0/agentassistconfigs) and provides a preview of the grid.
  </Card>

  <Card title="Copilot: Transcript Tile" href="/ai/agents/develop/node-reference/agent-copilot/transcript-tile">
    This Node captures the most recent customer response and displays it on a designated widget in the Agent Copilot workspace.
  </Card>
</CardGroup>

## More Information

* [Agent Copilot](/agent-copilot/overview)
* [Getting started with Agent Copilot](/agent-copilot/getting-started/chat)
* [Agent Copilot configuration](/agent-copilot/configure/overview)
* [Agent Copilot embedding](/agent-copilot/deploy/embedding)
* [Agent Copilot for Voice](/agent-copilot/getting-started/voice)
* [Agent Copilot for Chat](/agent-copilot/getting-started/chat)


# Copilot: Send Data
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/agent-copilot/send-data



<a href="/release-notes/4.90"><Badge>Updated in 4.90</Badge></a>

<Frame>
  <img alt="Copilot Send Data Node configuration panel" />
</Frame>

## Description

This Node updates the content within Agent Copilot widgets (iframe or HTML) without reloading the entire widget, making it easier to modify and refresh dynamic content in real time.

## Parameters

| Parameter | Type          | Description                                                                                                                                                                                                                         |
| --------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Tile ID   | CognigyScript | The ID of the [Copilot: IFrame Tile](/ai/agents/develop/node-reference/agent-copilot/set-iframe-tile) or [Copilot: HTML Tile](/ai/agents/develop/node-reference/agent-copilot/set-html-tile) Node you want to update with new data. |
| JSON Data | JSON          | The data to send to the [Copilot: IFrame Tile](/ai/agents/develop/node-reference/agent-copilot/set-iframe-tile) or [Copilot: HTML Tile](/ai/agents/develop/node-reference/agent-copilot/set-html-tile) Node as a postMessage event. |

## Example

This example shows how to dynamically update metadata in the Copilot: HTML Tile Node using the Copilot: Send Data and a processing script.

<Accordion title="Update metadata in the Copilot: HTML Tile Node using the Copilot: Send Data">
  **Copilot: HTML Tile Node**

  1. Enter the following script in the **HTML Content** field:

     ```html theme={null}
     <!DOCTYPE html>
     <html>
       <head></head>
       <body>
         <!-- Script to listen for messages sent to this HTML Tile -->
         <script>
           // Listen for 'message' events (data sent to this HTML tile)
           window.addEventListener("message", function (event) {
             // Log the content of the received message to the console
             console.log("Content of message: " + JSON.stringify(event.data));

             // If data is received, display it inside the metadataDisplay div
             if (event.data) {
                 document.getElementById("metadataDisplay").innerText = "Metadata: " + JSON.stringify(event.data);
             }
           });
         </script>

         <!-- Header for the HTML Tile -->
         <h1>HTML Tile</h1>

         <!-- Placeholder for dynamic content -->
         <div id="test">
             Here will be your content
         </div>

         <!-- Div to display received metadata -->
         <div id="metadataDisplay">Metadata: None</div>
       </body>
     </html>
     ```

     The JavaScript in the `<script>` tag listens for messages sent via the `postMessage` API. Once a message is received, the content of the message (stored in `event.data`) is logged to the browser's console. If the message contains data, the content of the `metadataDisplay` div updates to show the received data.

  2. Define the current version of the Copilot: HTML Tile Node content on the **JSON** field:

     ```json theme={null}
     {
       "version": 1
     }
     ```

  **Copilot: Send Data Node**

  Configure the Copilot: Send Data Node:

  1. Below the Copilot: HTML Tile Node, add a Copilot: Send Data Node.
  2. Open the Node editor and enter the Tile ID of the Copilot: HTML Tile Node in the **Tile ID** field.
  3. In the **JSON** field, enter the following JSON:

     ```json theme={null}
     {
       "version": 2,  // Version of the Copilot: HTML Tile Node
       "additionalInfo": "Updated metadata"  // The updated metadata that will be sent to the Copilot: HTML Tile Node
     }
     ```

  The Copilot: Send Data Node sends updated metadata to the Copilot: HTML Tile Node when triggered. The listener script inside the Copilot: HTML Tile Node dynamically displays this metadata on the widget.
</Accordion>

## More Information

* [Copilot: HTML Tile](/ai/agents/develop/node-reference/agent-copilot/set-html-tile)
* [Copilot: IFrame Tile](/ai/agents/develop/node-reference/agent-copilot/set-iframe-tile)
* [Agent Copilot Nodes](/ai/agents/develop/node-reference/agent-copilot/overview)
* [Agent Copilot](/agent-copilot/overview)


# Copilot: Sentiment Tile
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/agent-copilot/sentiment-tile



<a href="/release-notes/4.60"><Badge>Updated in 4.60</Badge></a>

<Frame>
  <img alt="Copilot Sentiment Tile Node configuration panel" />
</Frame>

## Description

This Node performs sentiment analysis on the most recent input and provides insights into the emotional tone (positive, negative, or neutral) of the last user inputs. This real-time analysis offers immediate feedback to human agents regarding the emotional tone of the customer within the Agent Copilot workspace. For instance, if the sentiment becomes negative during a conversation, the system can promptly alert the human agent to take appropriate action.

## Prerequisites

* Add a [compatible LLM](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) and select the LLM in the [Generative AI settings](/ai/agents/develop/gen-ai-and-llms/generative-ai).

## Parameters

| Parameter                | Type          | Description                                                                                                                                    |
| ------------------------ | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------- |
| Tile ID                  | CognigyScript | The ID that you created in the Agent Copilot configuration.                                                                                    |
| Sentiment Analysis Input | Select        | The number of user inputs that should be analyzed by a model. You can choose whether you want to analyze the last input or the last 10 inputs. |

## Use Case

The Copilot: Sentiment Tile Node allows you to analyze a larger part of the conversation, extending beyond the user's latest message. This Node helps provide a more objective conversation assessment and empowers the management of the customer's overall impression.

<Frame>
  <img alt="Sentiment Tile widget showing emotional tone analysis in Agent Copilot" />
</Frame>

## More Information

* [Copilot: Transcript Tile](/ai/agents/develop/node-reference/agent-copilot/transcript-tile)
* [Copilot: Next Action Tile](/ai/agents/develop/node-reference/agent-copilot/next-action-tile)
* [Copilot: Identity Tile](/ai/agents/develop/node-reference/agent-copilot/identity-tile)
* [Agent Copilot Nodes](/ai/agents/develop/node-reference/agent-copilot/overview)
* [Agent Copilot](/agent-copilot/overview)


# Copilot: Adaptive Card Tile
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/agent-copilot/set-adaptive-card-tile



<a href="/release-notes/4.79"><Badge>Updated in 4.79</Badge></a>

<Frame>
  <img alt="Copilot Adaptive Card Tile Node configuration panel" />
</Frame>

## Description

This Node allows you to create and display adaptive cards in your Agent Copilot workspace. Adaptive cards can display rich content and interactive elements, such as images, text, buttons, and forms.

The [Adaptive Card Designer](https://adaptivecards.io/designer) can help you to define an Adaptive Card. Copy the JSON card payload and paste it into the form. Alternatively, you can use Generative AI to create a new adaptive card or improve an existing one.

## Parameters

| Parameter            | Type          | Description                                                                                        |
| -------------------- | ------------- | -------------------------------------------------------------------------------------------------- |
| Tile ID              | CognigyScript | The ID that you created in the Agent Copilot configuration.                                        |
| Generate Node Output | Text          | A prompt or instruction for generating the adaptive card. To apply the prompt, click **Generate**. |
| Adaptive Card JSON   | JSON          | The JSON payload to render inside the tile. Supported schema version: `1.5` or earlier.            |

<Accordion title="Create an Adaptive Card with Generative AI">
  Before creating an Adaptive Card with Generative AI, ensure that you are connected to one of the [LLM Providers](/ai/agents/develop/gen-ai-and-llms/generative-ai#prerequisites).

  To use this feature, follow these steps:

  1. In the **Generate Node Output** section, instruct the Generative AI model how to improve the current adaptive card. For example, `Create a form with a customer name field and a date input field`.
  2. Click **Generate**. The adaptive card will be generated.
  3. Iteratively improve the resulting adaptive card by giving further instructions in the **Generate Node Output** section. For example, `Add a flight number field`.
  4. Click **Generate**. The existing adaptive card will be updated.

  To navigate between your inputs, use <img alt="previous input" /> and <img alt="next input" />.

  <Frame>
    <img alt="Adaptive Card Tile editor with Generative AI prompt" />
  </Frame>
</Accordion>

## More Information

* [Copilot: HTML Tile](/ai/agents/develop/node-reference/agent-copilot/set-html-tile)
* [Copilot: IFrame tile](/ai/agents/develop/node-reference/agent-copilot/set-iframe-tile)
* [Copilot: Set Grid](/ai/agents/develop/node-reference/agent-copilot/set-grid)
* [Agent Copilot Nodes](/ai/agents/develop/node-reference/agent-copilot/overview)
* [Agent Copilot](/agent-copilot/overview)


# Copilot: Set Grid
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/agent-copilot/set-grid



<a href="/release-notes/4.69"><Badge>Updated in 4.69</Badge></a>

<Frame>
  <img alt="Copilot Set Grid Node configuration panel" />
</Frame>

## Description

The Copilot: Set Grid Node updates the existing Agent Copilot workspace grid in the Agent Copilot Configuration, created via an [API POST request](https://api-trial.cognigy.ai/openapi#post-/v2.0/agentassistconfigs). Additionally, this Node can be used instead of the Agent Copilot Configuration to preview, edit, and modify the grid.

To edit the grid configuration, open the Node Editor and modify the default Agent Copilot Grid Configuration in JSON, then check the preview below to confirm the configuration is correct.

Multiple Nodes can be utilized within a single Agent Copilot Flow, enabling you to modify the grid according to your specific requirements.

## Parameters

| Parameter                        | Type          | Description                                                                                                                                                                                                                                                                                                                              |
| -------------------------------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Agent Copilot Grid Configuration | JSON          | Updates the Agent Copilot workspace grid configuration. By default, there is a grid example in the code editor. You can view a grid with Tile IDs in the Node editor.                                                                                                                                                                    |
| Background Selector              | Select        | Customizes the background color of the Agent Copilot workspace. You can select one of the following options:<ul><li>**Default** — the default white color.</li><li>**Solid Background Color** — specify the color in RGB or Hex format.</li><li>**Custom Image** — specify an image URL.</li></ul>                                       |
| Color (RGB/Hex)                  | CognigyScript | This field appears if the **Solid Background Color** option is selected.<br /><br />Enter the color in one of the following formats:<ul><li>**RGB**. For example, green can be represented as `rgb(34, 85, 85)`.</li><li>**Hex**. For example, red can be represented as `#FF3333`.</li></ul>The color will be rendered under the field. |
| Image URL                        | CognigyScript | This field appears if the **Custom Image** option is selected.<br /><br />Provide the image URL for your custom background. The image will be rendered under the field.                                                                                                                                                                  |

## More Information

* [Set HTML file](/ai/agents/develop/node-reference/agent-copilot/set-html-tile)
* [Set Adaptive card tile](/ai/agents/develop/node-reference/agent-copilot/set-adaptive-card-tile)
* [Set IFrame tile](/ai/agents/develop/node-reference/agent-copilot/set-iframe-tile)
* [Agent Copilot Nodes](/ai/agents/develop/node-reference/agent-copilot/overview)
* [Agent Copilot](/agent-copilot/overview)


# Copilot: HTML Tile
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/agent-copilot/set-html-tile



<a href="/release-notes/4.90"><Badge>Updated in 4.90</Badge></a>

<Frame>
  <img alt="Copilot HTML Tile Node configuration panel" />
</Frame>

## Description

This Node displays HTML content in the Agent Copilot workspace. You can design custom layouts to present information to a human agent. HTML content may include text, images, videos, links, and scripts. The Set HTML Tile Node is useful for displaying complex information.

Technically, the HTML content will be injected into an [iFrame](https://developer.mozilla.org/en-US/docs/Web/HTML/Element/iframe) on the workspace side, so any general restrictions on iFrames also apply to this case.

## Parameters

| Parameter    | Type          | Description                                                 |
| ------------ | ------------- | ----------------------------------------------------------- |
| Tile ID      | CognigyScript | The ID that you created in the Agent Copilot configuration. |
| HTML Content | HTML          | The HTML content to render inside the Widget.               |
| JSON Data    | JSON          | The data to send to the iframe as a postMessage event.      |

## Using React or Angular

In the Copilot: HTML Tile Node, you can enhance your HTML content by utilizing the [React](https://react.dev/) or [Angular](https://angular.dev/) libraries.
These libraries allow you to seamlessly display dynamic widgets within your workspace.
To use these libraries in the Copilot: HTML Tile Node, you need to include them in your HTML code.

<Accordion title="Example">
  ```html theme={null}
  <!DOCTYPE html>
  <html lang="en">
    <head>
      <meta charset="UTF-8" />
      <meta name="viewport" content="width=device-width, initial-scale=1.0"/>
      <title>Minimal React Widget</title>
      <!-- External React libraries -->
      <script src="https://unpkg.com/react@18/umd/react.production.min.js"></script>
      <script src="https://unpkg.com/react-dom@18/umd/react-dom.production.min.js"></script>
      <script src="https://unpkg.com/@babel/standalone/babel.min.js"></script>
    </head>
    <body>
      <div id="root"></div>

      <script type="text/babel">
        const { useState } = React;

        // this event listener processes incoming data updates
        window.addEventListener("message", function (event) {
          console.log("Content of message: " + event.data);
        });

        function ReactWidget() {
          const [name, setName] = useState('');
          const [message, setMessage] = useState('');

          const handleData = () => {
            if (name) {
              setMessage('Your name is ' + name);
            } else {
              setMessage('Please enter your name.');
            }
          };

          return (
            <div>
              <input
                type="text"
                value={name}
                onChange={(e) => setName(e.target.value)}
                placeholder="Name"
              />
              <button onClick={handleData}>Submit</button>
              <p>{message}</p>
            </div>
          );
        }

        ReactDOM.render(<ReactWidget />, document.getElementById('root'));
      </script>
    </body>
  </html>
  ```
</Accordion>

## Receiving JSON Data from the Flow

When you use a Copilot: HTML Tile, JSON data is passed into the HTML code using the [postMessage](https://developer.mozilla.org/en-US/docs/Web/API/Window/postMessage) method.
To receive and process the passed JSON in your widget, add an event listener into the Copilot: HTML Tile Node and handle the data accordingly.

```js theme={null}
window.addEventListener("message", function (event) {
  console.log("Content of message: " + event.data);
});
```

## Sending JSON Data back to the Flow

You can send data back to the Flow by using the Postback feature.

<Accordion title="Postback">
  The Postback feature allows you
  to send data either back to the Flow or chat, or directly to the widget interface.
  This feature can be useful for updating the state of your widget based on human agent interactions.
  These interactions can include actions such as submitting forms,
  updating display elements, or notifying the system of changes made by the human agent.

  The Postback feature can be configured in the following ways:

  * [Using SDK.postback](#using-sdkpostback)
  * [Using the Copilot: Send Data](#using-the-copilot-send-data)

  ### Using SDK.postback

  Follow these steps to implement postback for your widget:

  1. Add the `SDK.postback` function to the HTML page. Make sure this code is included after the SDK script is loaded:

  <Tabs>
    <Tab title="Send a String">
      ```html theme={null}
      <button type="button" onclick="SDK.postback('value');">Yes</button>
      ```
    </Tab>

    <Tab title="Send an Object">
      ```js theme={null}
      <button type="button" onclick="SDK.postback({ "key": "value" });">Yes</button>
      ```
    </Tab>
  </Tabs>

  2. Click the button in the widget in the Agent Copilot workspace. Triggering this action will send an injected message formatted as follows:
     ```json theme={null}
     {
     "data": {
         "_cognigy": {
         "_agentAssist": {
             "type": "submit",
             "payload": "value"
         }
         }
     }
     }
     ```

  To allow human agents to view this object, add `{{ input.data._cognigy._agentAssist }}` to the **Text** field in the Say Node of your Agent Copilot Flow.
  This Node should be placed below the Node with the `SDK.postback` function.
  The output will be available in the chat message and visible only to human agents.

  #### Example

  In this example, the `SDK.postback` function sends the value of `name` back to the user.
  If `name` is falsy, meaning the user did not enter a name, the `else` block is executed, and the function calls `setMessage('Please enter your name.');`.
  This message will appear in the widget, prompting the user to provide their name and reminding them of the required input.

  ```html theme={null}
  <!DOCTYPE html>
  <html lang="en">
  <head>
  <meta charset="UTF-8" />
  <meta name="viewport" content="width=device-width, initial-scale=1.0"/>
  <title>Minimal React Widget</title>
  <!-- External React libraries -->
  <script src="https://unpkg.com/react@18/umd/react.production.min.js"></script>
  <script src="https://unpkg.com/react-dom@18/umd/react-dom.production.min.js"></script>
  <script src="https://unpkg.com/@babel/standalone/babel.min.js"></script>
  </head>
  <body>
  <div id="root"></div>
  <script type="text/babel">
          const { useState } = React;

          // This event listener processes incoming data updates
          window.addEventListener("message", function (event) {
              console.log("Content of message: " + event.data);
          });

          function ReactWidget() {
              const [name, setName] = useState('');
              const [message, setMessage] = useState('');

              const handleData = () => {
                  if (name) {
                      setMessage('Your name is ' + name);
                      SDK.postback(name);
                  } else {
                      setMessage('Please enter your name.');
                  }
              };

              return (
                  <div>
                      <input
                          type="text"
                          value={name}
                          onChange={(e) => setName(e.target.value)}
                          placeholder="Name"
                      />
                      <button onClick={handleData}>Submit</button>
                      <p>{message}</p>
                  </div>
              );
          }

          ReactDOM.render(<ReactWidget />, document.getElementById('root'));
      </script>
  <button type="button" onclick="SDK.postback('yes');">yes</button>
  </body>
  </html>
  ```

  ### Using the Copilot: Send Data

  In the Agent Copilot Flow, below the Copilot: HTML Tile Node, add a [Copilot: Send Data](/ai/agents/develop/node-reference/agent-copilot/send-data).
  In the **JSON** field of the Copilot: Send Data Node,
  specify the parameters you want to pass as metadata to the Copilot: HTML Tile Node to update the content.
</Accordion>

## More Information

* [Copilot: Adaptive Card Tile](/ai/agents/develop/node-reference/agent-copilot/set-adaptive-card-tile)
* [Copilot: IFrame Tile](/ai/agents/develop/node-reference/agent-copilot/set-iframe-tile)
* [Copilot: Send Data](/ai/agents/develop/node-reference/agent-copilot/send-data)
* [Copilot: Set Grid](/ai/agents/develop/node-reference/agent-copilot/set-grid)
* [Agent Copilot Nodes](/ai/agents/develop/node-reference/agent-copilot/overview)
* [Agent Copilot](/agent-copilot/overview)


# Copilot: IFrame Tile
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/agent-copilot/set-iframe-tile



<a href="/release-notes/4.90"><Badge>Updated in 4.90</Badge></a>

<Frame>
  <img alt="Copilot IFrame Tile Node configuration panel" />
</Frame>

## Description

This Node embeds external websites directly into the Agent Copilot workspace. To do that, specify a URL for the external content you want to display. You can present web pages, forms, or applications within the Agent Copilot workspace without redirecting human agents to an external site.

## Parameters

| Parameter  | Type          | Description                                                 |
| ---------- | ------------- | ----------------------------------------------------------- |
| Tile ID    | CognigyScript | The ID that you created in the Agent Copilot configuration. |
| IFrame URL | URL           | The URL to render inside the Widget.                        |
| JSON Data  | JSON          | The data to send to the iframe as a postMessage event.      |

## Passing JSON Data into the IFrame

When you use a Copilot: IFrame Tile Node, JSON data is passed into the iframe code using the [postMessage](https://developer.mozilla.org/en-US/docs/Web/API/Window/postMessage) method.

To receive and process the passed JSON in your widget, add an event listener inside the iframe content:

```js theme={null}
window.addEventListener("message", function (event) {
  console.log("Content of message:", event.data);
  // Handle the incoming JSON data here
});
```

<Accordion title="Example: Dynamic IFrame Content">
  ```html theme={null}
  <!DOCTYPE html>
  <html>
    <head>
      <title>IFrame Example</title>
    </head>
    <body>
      <script>
        let counter = 0;

        // Update the body content every second
        setInterval(function() {
          counter++;
          document.body.innerHTML = '<h1>iframe Tile ' + counter + '</h1>';
        }, 1000);

        // Listen for messages from the parent window
        window.addEventListener('message', function(event) {
          console.log('Content of message from parent:', event.data);
          // Process the received data here
        });
      </script>
      <h1>iframe Tile</h1>
    </body>
  </html>
  ```
</Accordion>

At this stage, the iframe can get data from the Agent Copilot workspace. Next, you can set up the iframe to send data back to the Flow.

## Send Data from the IFrame Back to the Flow

User actions inside the iframe, such as form submissions or button clicks, can trigger logic in the Agent Copilot Flow.

To send data from an iframe back to the Agent Copilot Flow, use the browser messaging API method `window.parent.postMessage()`. The Agent Copilot workspace receives this message and forwards it to the Flow using `SDK.postback`.

<Note>
  Only the workspace can call `SDK.postback`. The iframe itself can't call it directly because it runs in an isolated context.
</Note>

<Accordion title="Send Data back to the Flow">
  <Steps>
    <Step title="IFrame sends data">
      The iframe sends a JSON object as data to the Agent Copilot workspace via `window.parent.postMessage()`.

      ```html theme={null}
      <button type="button" onclick="window.parent.postMessage({ name: 'John' }, '*')">Submit</button>
      ```
    </Step>

    <Step title="Workspace receives the message">
      The Agent Copilot workspace receives messages via an HTML tile controlled by the [Copilot: HTML Tile Node](/ai/agents/develop/node-reference/agent-copilot/set-html-tile), which listens for messages from the iframe.
      The HTML Tile doesn't need to be visible, but it must exist in the workspace.

      In this example:

      * `window.parent.addEventListener` ensures the message is caught from the parent workspace.
      * Only data with the `name` property is sent to the Flow.
      * `SDK.postback` automatically wraps the data into the `_agentAssist` payload for the Flow.

      ```html theme={null}
      <script>
      window.parent.addEventListener("message", function (event) {
        console.log("Content of parent message:", event.data);
        if (event.data.name) {
          SDK.postback(event.data.name);
        }
      });
      </script>
      ```
    </Step>

    <Step title="HTML Tile forwards data to the Flow">
      The HTML Tile forwards the received data to the Agent Copilot Flow using `SDK.postback`.
    </Step>

    <Step title="Flow receives the message">
      The Flow receives the message payload in this JSON structure:

      ```json theme={null}
      {
        "data": {
          "_cognigy": {
            "_agentAssist": {
              "type": "submit",
              "payload": { "name": "John" }
            }
          }
        }
      }
      ```
    </Step>

    <Step title="Accessing data in the Flow">
      You can access the data received from the iframe in the Flow via the Input object.

      In the Agent Copilot Flow, add a Say Node and reference the payload using the following syntax:

      ```txt theme={null}
      {{ input.data._cognigy._agentAssist.payload.name }}
      ```

      This output is visible only to human agents in the Copilot workspace.
    </Step>
  </Steps>

  The sequence diagram illustrates how the data is sent from the iframe to the Flow via the workspace:

  ```mermaid theme={null}
  sequenceDiagram
      participant IFrame
      participant Parent as Copilot Workspace
      participant Flow as Agent Copilot Flow

      IFrame->>Parent: window.parent.postMessage({ name: 'John' })
      Parent->>Parent: HTML Tile receives message
      Parent->>Flow: SDK.postback(event.data.name)
      Flow-->>Flow: Process data
      Flow-->>Flow: Access data via input object
  ```
</Accordion>

<Accordion title="Example: Name Submission from an IFrame">
  ```html theme={null}
  <input id="name" type="text" placeholder="Enter your name" />
  <p id="message"></p>
  <button type="button" onclick="
      const name = document.getElementById('name').value;
      if (name) {
        window.parent.postMessage({ name }, '*');
      } else {
        document.getElementById('message').innerText = 'Please enter your name.';
      }
    ">
    Submit
  </button>
  ```
</Accordion>

## More Information

* [Copilot: HTML Tile](/ai/agents/develop/node-reference/agent-copilot/set-html-tile)
* [Copilot: Adaptive Card Tile](/ai/agents/develop/node-reference/agent-copilot/set-adaptive-card-tile)
* [Copilot: Send Data](/ai/agents/develop/node-reference/agent-copilot/send-data)
* [Copilot: Set Grid](/ai/agents/develop/node-reference/agent-copilot/set-grid)
* [Copilot Nodes](/ai/agents/develop/node-reference/agent-copilot/overview)
* [Copilot](/agent-copilot/overview)


# Copilot: Transcript Tile
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/agent-copilot/transcript-tile



<a href="/release-notes/4.60"><Badge>Updated in 4.60</Badge></a>

<Frame>
  <img alt="Copilot Transcript Tile Node configuration panel" />
</Frame>

## Description

This Node captures the most recent customer response and displays it on a designated Widget within the Agent Copilot workspace.

Additionally, the Copilot: Transcript Tile offers the option to enable Sentiment Analysis,
which allows for assessing the user's speech tone.
This sentiment analysis can provide insights into whether the user's response is positive, negative, or neutral.
Before using this option,
set the LLM provider in the [Settings](/ai/agents/develop/gen-ai-and-llms/generative-ai)
and select the appropriate model in the [supported model list](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature).

## Parameters

| Parameter        | Type          | Description                                                                                                                                                                                                                |
| ---------------- | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Tile ID          | CognigyScript | The ID that you created in the Agent Copilot configuration.                                                                                                                                                                |
| Enable sentiment | Toggle        | This option allows you to decide if sentiment analysis should be applied to the customer's voice input. Enabling this option means that Cognigy assesses the emotional tone (positive, negative, or neutral) of the input. |

## Use Case

This Node is well-suited for scenarios involving voice calls,
where a human agent needs to listen to the user's speech and make quick decisions.
In such cases, the Copilot: Transcript Tile plays a crucial role in recording and presenting the user's spoken responses.
This approach allows human agents to have immediate access to this information,
thereby facilitating more effective and efficient communication and decision-making during voice call interactions.

<Frame>
  <img alt="Transcript Tile widget displaying voice call transcript in Agent Copilot" />
</Frame>

## More Information

* [Copilot: Next Action Tile](/ai/agents/develop/node-reference/agent-copilot/next-action-tile)
* [Copilot: Sentiment Tile](/ai/agents/develop/node-reference/agent-copilot/sentiment-tile)
* [Copilot: Identity Tile](/ai/agents/develop/node-reference/agent-copilot/identity-tile)
* [Agent Copilot Nodes](/ai/agents/develop/node-reference/agent-copilot/overview)
* [Agent Copilot](/agent-copilot/overview)


# Add Lexicon Keyphrase
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/ai/add-lexicon-keyphrase



<Frame>
  <img alt="Add Lexicon Keyphrase Node configuration panel" />
</Frame>

## Description

The Add Lexicon Keyphrase Node adds new Keyphrases to [Lexicons](/ai/platform-features/nlu/slots/user-defined/lexicon) from the Flow using [CognigyScript](/ai/platform-features/cognigyscript).

## Parameters

| Parameter   | Type                         | Description                                   |
| ----------- | ---------------------------- | --------------------------------------------- |
| LexiconName | CognigyScript                | The name of the Lexicon you want to update    |
| Keyphrase   | CognigyScript                | The Keyphrase you want to add to the Lexicon  |
| Tags        | String, one tag per line     | Tags you want to attach to the Keyphrase      |
| Synonyms    | String, one synonym per line | Synonyms you want to attach to the Keyphrase. |


# AI Agent
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/ai/ai-agent



<a href="/release-notes/2026.18"><Badge>Updated in 2026.18</Badge></a>

<Frame>
  <img alt="AI Agent Node" />
</Frame>

## Description

<Warning>
  As of Cognigy 2026.18.0, the **Maximum Completion Tokens** and **Temperature** parameters are deprecated. The removal is planned for April 2027. For models that don't support these parameters, activate the **Ignore deprecated parameters** toggle to exclude them from the request to the LLM provider. Otherwise, the request to the LLM provider might fail. If you plan to use these parameters, migrate them to the [Custom Options section](#custom-options) and activate the **Ignore deprecated parameters** toggle. Otherwise, Custom Options might not override them for every provider and model combination.

  For example, OpenAI models such as `gpt-5`, `gpt-5.1`, and `gpt-5.2` don't support these parameters. Similar models from other LLM providers might also not support these parameters. Check your LLM provider's documentation for model-specific support.
</Warning>

The AI Agent Node assigns a job to an AI Agent, provides instructions and tools for that job, and access to the knowledge the AI Agent can use when holding a conversation with a user.

To configure this Node, follow these steps:

1. [Define an AI Agent job](#parent-node).
2. [Define the tool actions to perform this task](#child-nodes).

## Limitations

* By default, the number of transcript turns is set to 50. In the AI Agent Node, this number is fixed and can't be adjusted, unlike in the [LLM Prompt](/ai/agents/develop/node-reference/service/llm-prompt#advanced) and [Get Transcript](https://docs.cognigy.com/ai/agents/develop/node-reference/service/get-transcript#parameters) Nodes, where customization is possible.

- The **Description** and **Enum** fields in the **Parameters** section of an [AI Agent tool](#parameters) don't support CognigyScript.

## Parameters

### Parent Node

This configuration assigns a job to an AI Agent, defines its role and responsibilities, and provides additional instructions or context to guide its actions.

<AccordionGroup>
  <Accordion title="AI Agent">
    | Parameter                | Type          | Description                                                                                                                                                                                                                                                                                                 |
    | ------------------------ | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | AI Agent                 | Selector      | Select the AI Agent.                                                                                                                                                                                                                                                                                        |
    | Job Name                 | CognigyScript | Specify the name of the job. For example, `Customer Support Specialist`.                                                                                                                                                                                                                                    |
    | Job Description          | CognigyScript | Provide a description of the job responsibilities to guide the AI Agent's interactions. For example, `Assist customers with product issues, escalate complex cases, and provide guidance on best practices`.                                                                                                |
    | Instructions and Context | Toggle        | Add specific instructions or context as a system message to help the AI Agent better fulfill the job requirements. For example, `Stay professional and friendly; focus on problem-solving and clarity`. These instructions are considered in addition to those specified in the AI Agent creation settings. |
  </Accordion>

  <Accordion title="Memory Handling">
    | Parameter                   | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
    | --------------------------- | ------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Long-Term Memory Injection  | Selector      | Allow the AI Agent to access [Contact Profile](/ai/agents/analyze/contact-profiles) information for the current user. Select one of the following options:<ul><li>**None** – no memory.</li><li>**Inherit from AI Agent** – use the settings specified in the [AI Agent creation settings](/ai/agents/develop/manage-ai-agents#create-ai-agents).</li><li>**Inject full Contact Profile** – use all information from the Contact Profile.</li><li>**Inject Contact Memories only** – use information only from the **Memories** field in the Contact Profile.</li><li>**Inject selected Profile fields** – use information from specific fields in the Contact Profile.</li></ul> |
    | Selected Profile Fields     | Text          | This parameter appears when the **Inject selected Profile fields** option is enabled. Enter specific fields from the Contact Profile for targeted data use. Specify the field using the [Profile keys](/ai/agents/analyze/contact-profiles#profile-schema) format and press <kbd>Enter</kbd> to apply it.                                                                                                                                                                                                                                                                                                                                                                         |
    | Short-Term Memory Injection | CognigyScript | Specify a static string or a dynamic value via CognigyScript to make available to the AI Agent in the current turn.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
  </Accordion>

  <Accordion title="Grounding Knowledge">
    | Parameter                  | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | -------------------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Knowledge Injection        | Selector      | Use the Knowledge AI feature for the AI Agent. Select one of the following options:<ul><li>**Never** — do not use the Knowledge Stores.</li><li>**When Required** — let the AI Agent decide when querying the Knowledge Stores is required to help the user.</li><li>**Once for Each User Input** — query the Knowledge Stores after each user input. Note that executing a query on every user input can lead to increased costs and latency.</li></ul>                                                                                                                                                                                              |
    | Use AI Agent Knowledge     | Toggle        | Appears when you select **When Required** or **Once for Each User Input**. Enable to use the Knowledge Store configured in the AI Agent. The Knowledge Store configured within the [AI Agent creation settings](/ai/agents/develop/manage-ai-agents#create-ai-agents) will be used.                                                                                                                                                                                                                                                                                                                                                                   |
    | Use Job Knowledge          | Toggle        | Appears when you select **When Required** or **Once for Each User Input**. Enable this option to configure a specific Knowledge Store for this particular job, allowing the AI Agent to access job-specific data or resources.                                                                                                                                                                                                                                                                                                                                                                                                                        |
    | Job Knowledge Store        | Selector      | Appears when you select **When Required** or **Once for Each User Input** and **Use Job Knowledge** is enabled. Select a specific Knowledge Store for this AI Agent's job.                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | Top K                      | Slider        | Appears when you select **When Required** or **Once for Each User Input**. Specify how many knowledge chunks to return. Providing more results gives the AI Agent additional context but may increase noise and token usage.                                                                                                                                                                                                                                                                                                                                                                                                                          |
    | Source Tags                | CognigyScript | Appears when you select **When Required** or **Once for Each User Input**. Tags refine the scope of your knowledge search, including only the most relevant sections of the knowledge base to improve accuracy. Before specifying tags, ensure they were provided during the creation of the Knowledge Sources. Add tags by entering each separately and pressing <kbd>Enter</kbd>. Max 5 tags. When multiple Source Tags are specified, the Search Extract Output Node defaults to the `AND` operator, meaning it only considers Sources that have all specified tags. To change this behavior, adjust the **Match Type for Source Tags** parameter. |
    | Match Type for Source Tags | Select        | Appears when you select **When Required** or **Once for Each User Input**. Defines the operator for filtering Knowledge Sources by tags:<ul><li>**AND** — default; requires all tags to match. Example: filtering by `S-a` and `S-b` only includes Sources with both tags.</li><li>**OR** — requires at least one tag to match. Example: filtering by `S-a` or `S-b` includes Sources with either tag.</li></ul>                                                                                                                                                                                                                                      |
    | Generate Search Prompt     | Toggle        | Appears when you select **Once for Each User Input**. Enabled by default. Generates a context-aware search prompt before executing the knowledge search. May increase cost and latency. By default, the prompt includes the last 7 conversation turns. This value can't be changed.                                                                                                                                                                                                                                                                                                                                                                   |
  </Accordion>

  <Accordion title="Storage and Streaming Options">
    | Parameter                   | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
    | --------------------------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | How to handle the result    | Select        | Determine how to handle the prompt result:<ul><li>**Store in Input** — stores the AI Agent result in the Input object. To print the prompt result, refer to the configured Context key in a Say Node or enable the **Output result immediately** option.</li><li>**Store in Context** — stores the result in the Context object. To print the prompt result, refer to the configured Context key in a Say Node or enable the **Output result immediately** option.</li><li>**Stream to Output** — streams the result directly into the output. Chunks from the prompt response will be output into the conversation chat as soon as a Stream Buffer Flush Token is matched. You don't need to use the AI Agent Output token and a Say Node. By default, the result is not stored in the Input or Context. To store it, enable **Store Copy in Input**.</li></ul> |
    | Input Key to store Result   | CognigyScript | Appears when **Store in Input** or **Stream to Output** is selected. The result is stored in `input.aiAgentOutput` by default. You can specify another value, but the **AI Agent Output** token will not work if the value is changed.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | Context Key to store Result | CognigyScript | Appears when **Store in Context** is selected. The result is stored in `context.aiAgentOutput` by default. You can specify another key.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
    | Stream Buffer Flush Tokens  | Text Array    | Appears when **Stream to Output** is selected. Defines tokens that trigger the stream buffer to flush to the output. Tokens can be punctuation marks or symbols, such as `\n`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
    | Output result immediately   | Toggle        | Appears when **Store in Input** or **Store in Context** is selected. Allows immediate output of results without using the Say Node and AI Agent Output token.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
    | Store Copy in Input         | Toggle        | Appears when **Stream to Output** is selected. In addition to streaming the result, stores a copy in the Input object by specifying a value in the **Input Key to store Result** field.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
  </Accordion>

  <Accordion title="Voice">
    | Parameter                 | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
    | ------------------------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Voice Setting             | Select        | Configure the voice settings for the AI Agent Job. This parameter determines how the AI Agent selects the voice for text-to-speech (TTS) output. Select one of the following options:<ul><li>**Inherit from AI Agent** — use the voice settings defined in the [AI Agent creation settings](/ai/agents/develop/manage-ai-agents#create-ai-agents).</li><li>**Use Job Voice** — apply custom voice settings specific to this job, allowing the AI Agent to adapt to the particular role it performs. For example, for a marketing AI Agent, the voice can be engaging, friendly, and persuasive. For customer support, it might be neutral, empathetic, and formal.</li></ul> |
    | TTS Vendor                | Dropdown      | Select a TTS vendor from the list or add a custom one. Note: The AI Agent Node doesn't support **TTS Labels** to distinguish configurations from the same vendor. To use **TTS Labels**, add a [Set Session Config Node](/ai/agents/develop/node-reference/voice/voice-gateway/set-session-config) before the AI Agent Node in the Flow editor.                                                                                                                                                                                                                                                                                                                              |
    | Custom (Vendor)           | CognigyScript | Appears when **Custom** is selected in **TTS Vendor**. Specify the [custom TTS Vendor](/voice-gateway/webapp/speech-services#add-custom-speech-vendors). For preinstalled providers, use lowercase, for example, `microsoft`, `google`, `aws`. For custom providers, use the name defined on the [Speech Service](/voice-gateway/webapp/speech-services) page in the Voice Gateway Self-Service Portal.                                                                                                                                                                                                                                                                      |
    | TTS Language              | Dropdown      | Define the language of the AI Agent output. Ensure it aligns with the preferred language of the end user.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
    | Custom (Language)         | CognigyScript | Appears when **Custom** is selected in **TTS Language**. Specify the output language. Format depends on the TTS vendor; check vendor documentation. Typical format: `de-DE`, `fr-FR`, `en-US`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
    | TTS Voice                 | Dropdown      | Define the voice for AI Agent output. Customize tone, gender, style, and regional specifics to align conversations with your brand and audience.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | Custom (Voice)            | CognigyScript | Appears when **Custom** is selected in **TTS Voice**. Specify a custom voice, often required for region-specific voices. Format depends on **TTS Vendor** and typically follows `language-region-VoiceName`, for example, `de-DE-ConradNeural`, `en-US-JennyNeural`.                                                                                                                                                                                                                                                                                                                                                                                                         |
    | TTS Label                 | CognigyScript | Alternative name for the TTS vendor, as specified in the [Voice Gateway Self-Service Portal](/voice-gateway/webapp/applications#add-additional-tts-and-stt-vendor). Use this when multiple speech services from the same vendor exist.                                                                                                                                                                                                                                                                                                                                                                                                                                       |
    | Disable TTS Audio Caching | Toggle        | Disables TTS audio caching. By default, the setting is deactivated. In this case, previously requested TTS audio results are stored in the AI Agent cache. When a new TTS request is made and the audio text has been previously requested, the AI Agent retrieves the cached result instead of sending another request to the TTS provider.<br />When the setting is activated, the AI Agent caches TTS results but doesn't use them. In this case, each request is directly sent to your speech provider.<br />Note that disabling caching can increase TTS costs. For detailed information, contact your speech provider.                                                 |
  </Accordion>

  <Accordion title="Tool Settings">
    | Parameter       | Type     | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
    | --------------- | -------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Tool Choice     | Selector | Determines how the AI Agent selects tools:<ul><li>**Auto** — the model decides whether to call a tool or return a plain-text response. Use this option for general-purpose AI Agents. For example, a support AI Agent that answers FAQs directly but calls a Knowledge tool when the question needs it.</li><li>**Required** — the model must call one of the provided tools on every turn and won't return a plain-text response. Use this option when the Flow must always pass through a structured tool call. For example, when the tool's output must be retrieved and included in the response before anything is returned to the user. Selecting this option can cause an infinite tool-calling loop. See [Avoiding infinite loops when Tool Choice is Required](#avoiding-infinite-loops-when-tool-choice-is-required) for mitigation options.</li><li>**None** — the model ignores all configured tools and returns only a plain-text response. Use this option to temporarily disable tool use without removing the tool configuration, for example, during testing or debugging.</li></ul> <br /> The options in **Tool Choice** map to the following LLM function-calling modes: <ul><li>**Auto** — `mode: AUTO`</li><li>**Required** — `mode: ANY`</li><li>**None** — `mode: NONE`</li></ul>Availability of these modes depends on your LLM provider. Refer to your provider's documentation to confirm support. |
    | Use Strict Mode | Toggle   | When enabled, strict mode (if supported by the LLM provider) ensures that arguments passed to a tool call precisely match the expected parameters. This helps prevent errors but may slightly delay responses, especially during the first call after making changes.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         |
  </Accordion>

  <Accordion title="Image Handling">
    | Parameter            | Type     | Description                                                                                                                                                                                                                                                                                                                         |
    | -------------------- | -------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Process Images       | Toggle   | Enables the AI Agent to read and understand image attachments. Ensure that your LLM provider supports image processing (refer to your provider's documentation). Also verify that attachments are supported and activated in your Endpoint, such as Webchat.                                                                        |
    | Images in Transcript | Selector | Configures how images older than the last turn are handled to reduce token usage:<ul><li>**Minify** — reduces the size of these images to 512×512 px.</li><li>**Drop** — excludes the images.</li><li>**Keep** — sends the maximum size (consumes more tokens).</li></ul> Limitations and token consumption depend on the LLM used. |
  </Accordion>

  <Accordion title="Advanced">
    | Parameter                    | Type     | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
    | ---------------------------- | -------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Ignore deprecated parameters | Toggle   | Excludes the deprecated **Maximum Completion Tokens** and **Temperature** parameters from requests to the LLM provider. Activate this toggle if your model doesn't support these parameters. If you activate this toggle for a model that supports them, the model uses default values.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
    | LLM                          | Selector | Select a model that supports the [AI Agent Node feature](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature). The selected **Default** model is the one specified in **Settings > Generative AI Settings** of your Project. Choose the model you [added earlier](/ai/agents/overview#prerequisites) while configuring the Agentic AI feature. This model will manage your AI Agent.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
    | AI Agent Base Version        | Selector | Select the base version of the AI Agent to use:<ul><li>**Fixed Version** — choose a specific version, for example, `1.0`, to ensure stability and avoid potential breaking changes. Use this version in production environments or critical workflows. The version dropdown will be updated as future AI Agent Node versions are released.</li><li>**Latest** — use the most recent version of the AI Agent Node. This gives access to the latest features but may require manual updates if breaking changes occur.</li></ul>When upgrading to a fixed version or switching to the latest, always test your AI Agent carefully to ensure compatibility.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
    | Timeout                      | Number   | Define the maximum number of milliseconds to wait for a response from the LLM provider.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
    | Maximum Completion Tokens    | Slider   | Set the maximum number of tokens that can be used during a process to manage costs. If the limit is too low, the output may be incomplete. For example, setting 100 tokens roughly corresponds to 100 words, depending on language and tokenization.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         |
    | Temperature                  | Slider   | Define the sampling temperature, ranging from 0 to 1. Higher values, for example, 0.8, make output more random; lower values, for example, 0.2, make it more focused and deterministic.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
    | Include Rich Media Context   | Toggle   | Controls whether the text from the [Textual Description](/ai/agents/develop/node-reference/basic/say#output-type) field in Say, Question, and Optional Questions Nodes is added to the AI Agent's prompt. This text provides AI Agents with additional information, improving their responses.<br /><br />If you have filled in a **Textual Description** field in other Nodes, this text is added to the prompt as context for rich media, such as Text with Quick Replies. If the **Textual Description** field is empty, the button titles and alt text from the rich media are added to the prompt. By default, the **Include Rich Media Context** parameter is active. When this parameter is inactive, nothing related to the rich media is added to the prompt.<br /><br />**Examples**:<ul><li>If **Textual Description** is filled in:<p>Textual Description: `Select your preferred delivery option: Standard Delivery or Express Delivery`.</p><p>Quick Replies buttons: `Standard Delivery`, `Express Delivery`.</p><p>Context added to the prompt: `Select your preferred delivery option: Standard Delivery or Express Delivery`.</p></li><li>If **Textual Description** is empty:<p>Textual Description: empty.</p><p>Quick Replies buttons: `Standard Delivery`, `Express Delivery`.</p><p>Context added to the prompt: `Standard Delivery`, `Express Delivery`.</p></li><li>If **Include Rich Media Context** is inactive:<p>No information related to the rich media is added to the prompt.</p></li></ul> |
  </Accordion>

  <Accordion title="Error Handling">
    | Parameter                      | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                     |
    | ------------------------------ | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Log to System Logs             | Toggle        | Log errors to the system logs. They can be viewed on the [Logs](/ai/agents/test/logs) page of your Project. This parameter is inactive by default.                                                                                                                                                                                                                                              |
    | Store in Input                 | Toggle        | Store errors in the Input object.                                                                                                                                                                                                                                                                                                                                                               |
    | Select Error Handling Approach | Select        | Choose one of the Error Handling options:<ul><li>**Stop Flow Execution** — terminate the current Flow execution.</li><li>**Continue Flow Execution** — allow the Flow to continue executing, bypassing the error and proceeding to the next steps.</li><li>**Go to Node** — redirect the workflow to a specific Node in the Flow, useful for error recovery or custom error handling.</li></ul> |
    | Select Flow                    | Select        | Appears when **Go to Node** is selected. Choose a Flow from the available options.                                                                                                                                                                                                                                                                                                              |
    | Select Node                    | Select        | Appears when **Go to Node** is selected. Choose a Node from the available options.                                                                                                                                                                                                                                                                                                              |
    | Error Message (optional)       | CognigyScript | Add an optional message to the output if the AI Agent Node fails.                                                                                                                                                                                                                                                                                                                               |
  </Accordion>

  <Accordion title="Custom Options">
    These options let you use parameters that aren't included in the AI Agent Node or overwrite set configurations.

    | Parameter              | Type | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        | Example                                                                      |
    | ---------------------- | ---- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------- |
    | Custom Options         | JSON | Set custom parameters to the LLM configuration. You can set individual parameters as well as entire functions. These parameters customize the behavior of the model, such as adjusting `temperature`, `top_k`, or `presence_penalty`. This field overrides the parameters set in the AI Agent Node. To view the full list of available parameters for your model, refer to the LLM provider's API documentation, for example, [OpenAI](https://platform.openai.com/docs/api-reference/chat/create) or [Azure OpenAI](https://learn.microsoft.com/en-us/azure/ai-services/openai/). | `{ "model": "gpt-4o", "temperature": 0.7, "max_tokens": 4000 }`              |
    | Custom Request Options | JSON | Set custom parameters to customize the request to the LLM provider. These parameters override the configurations set in this Node. For example, setting a `{ "stream": true }` parameter overrides the **Output result immediately** option. For more information, refer to the LLM provider's API documentation.                                                                                                                                                                                                                                                                  | `{ "stream": true }`, `{ "headers": { "Authorization": "Bearer <token>" } }` |

    **Forcing Model Versions**

    You can force the AI Agent Node to use a specific model version by including it in the Custom Options.
    This means that the AI Agent Node will use the specified version of the language model instead of the default or any other available versions. This allows for more control over the behavior of the AI Agent Node, ensuring it uses the desired model version for generating prompts or responses.

    You can use models from any LLM provider supported by Cognigy, including those not yet directly integrated.
    However, you can only replace a model with another from the same provider.

    Let's consider an example with Anthropic:
    You can force the AI Agent Node to use the model version `claude-opus-4-1-20250805`, despite the LLM resource defaulting to the `claude-opus-4-5-20251101` model:

    1. [Create an Anthropic LLM resource](/ai/agents/develop/gen-ai-and-llms/providers/anthropic) for Claude, for example, `claude-opus-4-5-20251101`.
    2. Create a Flow and add an AI Agent Node to it.
    3. In the AI Agent Node, select the model `claude-opus-4-5-20251101` from the **Large Language Model** list.
    4. In the **Custom Model Options** field, add `{ "model": "claude-opus-4-1-20250805" }` to force the use of the `claude-opus-4-1-20250805` model.
    5. Click **Save Node**.

    The AI Agent Node now utilizes the `claude-opus-4-1-20250805` model.

    Below, you'll find documentation for supported models:

    * [Aleph Alpha Models](https://docs.aleph-alpha.com/)
    * [Anthropic Models](https://docs.anthropic.com/claude/docs/models-overview)
    * [Azure OpenAI](https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/models)
    * [OpenAI Models](https://platform.openai.com/docs/models)
  </Accordion>

  <Accordion title="Debug Settings">
    | Parameter                     | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                             |
    | ----------------------------- | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Log Job Execution             | Toggle        | Send a debug message with the current AI Agent Job configuration. The message appears in the Interaction Panel when debug mode is enabled. The parameter is active by default.                                                                                                                                                                                                                          |
    | Log Knowledge Results         | Toggle        | Send a debug message containing the result from a knowledge search. The message appears in the Interaction Panel when debug mode is enabled. The parameter is inactive by default.                                                                                                                                                                                                                      |
    | Show Token Count              | Toggle        | Send a debug message containing the input, output, and total token count. The message appears in the Interaction Panel when debug mode is enabled. Cognigy uses the GPT-3 tokenizer algorithm, so actual token usage may vary depending on the model used. The parameter is inactive by default.                                                                                                        |
    | Log System Prompt             | Toggle        | Send a debug message containing the system prompt. The message appears in the Interaction Panel when debug mode is enabled. The parameter is inactive by default.                                                                                                                                                                                                                                       |
    | Log Tool Definitions          | Toggle        | Send a debug message containing information about the configured AI Agent tools. The message appears in the Interaction Panel when debug mode is enabled. The parameter is inactive by default.                                                                                                                                                                                                         |
    | Log LLM Latency               | Toggle        | Send a debug message containing key latency metrics for the request to the model, including the time taken for the first output and the total time to complete the request. The message appears in the Interaction Panel when [debug mode](/ai/agents/test/interaction-panel/chat#debug-mode) is enabled. The parameter is inactive by default.                                                         |
    | Send request logs to Webhook  | Toggle        | Send the request sent to the LLM provider and the subsequent completion to a webhook service, including metadata, the request body, and custom logging data. With this parameter, you can use a webhook service to view detailed logs of the request to the LLM. The parameter is inactive by default.                                                                                                  |
    | Webhook URL                   | CognigyScript | Enter the URL of the webhook service to send the request logs to.                                                                                                                                                                                                                                                                                                                                       |
    | Custom Logging Data           | CognigyScript | Enter custom data to send with the request to the webhook service.                                                                                                                                                                                                                                                                                                                                      |
    | Condition for Webhook Logging | CognigyScript | Enter the condition for the webhook logging.                                                                                                                                                                                                                                                                                                                                                            |
    | Webhook Headers               | Input fields  | Enter the headers to send with the request to the webhook service. Use the **Key** and **Value** fields to enter a header. The **Value** field supports CognigyScript. After entering the header key, new empty **Key** and **Value** fields are automatically added, in case you need to add more headers. Alternatively, you can click **Show JSON Editor** and add input examples in the code field. |
  </Accordion>
</AccordionGroup>

### Child Nodes

#### Tool

Tools are child Nodes of AI Agent Nodes. They define the actions the AI Agent can take. If an AI Agent wants to execute the tool, the branch below the child Node is executed.

At the end of a tool branch, it is advisable to use a [Resolve Tool Action](/ai/agents/develop/node-reference/ai/resolve-tool-action) Node to return to the AI Agent.

Clicking the Tool Node lets you define a tool, set its parameters, and allows debugging by enabling detailed messages about the tool's execution.

<AccordionGroup>
  <Accordion title="Tool">
    | Parameter   | Type          | Description                                                                                                                                          |
    | ----------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Tool ID     | CognigyScript | Provide a meaningful name as a Tool ID. This ID can contain only letters, numbers, underscores (`_`), or dashes (`-`). For example, `update_user-1`. |
    | Description | CognigyScript | Provide a detailed description of what the tool does, when it should be used, and its parameters.                                                    |
  </Accordion>

  <Accordion title="Parameters">
    Configure the parameters that the AI Agent will collect before the tool is called. You can switch between the Graphical and JSON editors. When editing the JSON, follow the [JSON Schema specification](https://json-schema.org).

    | Parameter       | Type     | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         |
    | --------------- | -------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Use Parameters  | Toggle   | Activate this toggle to add parameters in addition to the tool name and description. The AI Agent will collect all data it needs and call a Tool with these parameters filled as arguments. These values can be accessed directly in the `input.aiAgent.toolArgs` object.                                                                                                                                                                                                                                                                                                                                                                           |
    | Name            | Text     | Specify the name of the parameter. The name should be clear and concise, and describe the purpose of the parameter.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
    | Type            | Selector | Select a type of the parameter:<ul><li>**String** — a sequence of characters. For example, `"hello"`, `"123"`.</li><li>**Number** — a numerical value, which can be either an integer (for example,`5`) or a floating point number (for example, `3.14`).</li><li>**Boolean** — a logical value representing `true` or `false`.</li><li>**Array** — a collection of elements, which can contain multiple values of any type. For example, `["apple", "banana", "cherry"]`.</li><li>**Object** — a collection of key-value pairs, where each key is a string and the value can be of any type. For example, `{"name": "John", "age": 30}`.</li></ul> |
    | Description     | Text     | Explain what the parameter means by providing a brief description of the parameter's usage.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         |
    | Enum (optional) | Enum     | Define a set of values that the parameter can accept. The enum restricts the input to one of the specified values, ensuring only valid options are chosen. The enum is only available for string-type parameters in the Graphical editor. For other types, use the JSON editor. May not be supported by all LLM providers.                                                                                                                                                                                                                                                                                                                          |
    | Add Parameter   | Button   | Add a new parameter.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
  </Accordion>

  <Accordion title="Debug Settings">
    | Parameter                 | Type   | Description                                                                                                       |
    | ------------------------- | ------ | ----------------------------------------------------------------------------------------------------------------- |
    | Debug Message when called | Toggle | Enable the output of a debug message when the tool is called to provide detailed information about the tool call. |
  </Accordion>

  <Accordion title="Advanced">
    | Parameter | Type          | Description                                                                                                                                                                                                                                                                                                                                                                        |
    | --------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Condition | CognigyScript | The tool will be enabled only if the condition is evaluated as true. If false, the tool isn't part of the AI Agent's Tools within this execution. For example, when using the `unlock_account` tool, you can specify a condition like `context.accountStatus === "locked"`. This checks the value in the context, and if it is missing or different, the tool will not be enabled. |
  </Accordion>
</AccordionGroup>

#### MCP Tool

MCP Tool Nodes are child Nodes of AI Agent Nodes. The MCP Tool Nodes connect to a remote [MCP server](https://modelcontextprotocol.io/introduction) to load tools that the AI Agent can execute. If an AI Agent wants to execute one of the loaded tools, the branch below the MCP Tool Node is triggered.

Clicking the MCP Tool Node lets you define the connection, filter loaded tools, and allows debugging by enabling detailed messages about the tool's execution.

<AccordionGroup>
  <Accordion title="MCP Tool">
    | Parameter          | Type          | Description                                                                                                                                                                            |
    | ------------------ | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Name               | CognigyScript | Provide a name for the MCP connection. This name helps you identify the source of the loaded tool.                                                                                     |
    | MCP Server SSE URL | CognigyScript | Provide the URL to an SSE (Server-Sent Events) endpoint from a remote [MCP server](https://modelcontextprotocol.io/introduction). Ensure that you connect only to trusted MCP servers. |
    | Timeout            | Slider        | Set the timeout time for the MCP connection in seconds.                                                                                                                                |
  </Accordion>

  <Accordion title="Authentication">
    | Parameter           | Type     | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
    | ------------------- | -------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Authentication Type | Selector | The authentication type to use for the MCP server connection. Select one of the following options: <ul><li>**None** — no authentication. This option is used by default.</li><li>**OAuth2** — authenticate using OAuth 2.0, which provides improved security and control compared to API keys. OAuth 2.0 supports scoped permissions, expiring tokens, and reduced exposure through short-lived tokens.</li></ul> If you want to use a different authentication method, such as an API key, you can add it as a custom header in **Custom Headers** of the **Advanced** section.                                                                                                                                                                                                                                                             |
    | OAuth2 Parameters   | Selector | Opens the fields required to configure OAuth 2.0 authentication. These fields allow the AI Agent to securely request and use access tokens when communicating with the MCP server. Configure the following fields: <ul><li>**Connection Name** — enter a descriptive name for the connection, for example, `MCP Tool OAuth`.</li><li>**OAuth URL** — enter the token endpoint URL exposed by the MCP server, for example: `https://mcp.example.com/oauth2/token`.</li><li>**Client ID** — enter the application (client) ID issued by the MCP server.</li><li>**Client Secret** — enter the application secret issued by the MCP server. Keep it secure.</li><li>**Scope** — add a list of scopes that define the permissions required by your application, for example: `read write admin`. Separate multiple scopes with spaces.</li></ul> |
  </Accordion>

  <Accordion title="Debug Settings">
    | Parameter             | Type   | Description                                                                                                                                                                                                                                                               |
    | --------------------- | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Debug Loaded Tools    | Toggle | Enable this parameter to display a debug message listing all tools loaded from the MCP server. The debug message also includes tools filtered out in the Advanced section. This parameter shows whether tools were loaded from the cache or directly from the MCP server. |
    | Debug with Parameters | Toggle | Enable this parameter to include the Tool Parameters in the debug message.                                                                                                                                                                                                |
    | Debug calling Tool    | Toggle | Enable the output of a debug message when the tool is called to provide detailed information about the tool call.                                                                                                                                                         |
  </Accordion>

  <Accordion title="Advanced">
    | Parameter      | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | -------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Cache Tools    | Toggle        | Disables caching of loaded tools while developing. Ensure that caching is enabled in production for performance reasons. The caching time is 10 minutes. In debug mode, you can verify whether the cache was used. To do this, make sure **Debug Loaded Tools** is enabled. For more information, see the [Debug Settings](#debug-settings) section.                                                                                                                    |
    | Condition      | CognigyScript | Sets the condition under which the tool will be activated. If the condition is evaluated as false, the tool isn't part of the AI Agent's Tools during execution. For example, when using the `unlock_account` tool, you can specify a condition like `context.accountStatus === "locked"`. This checks the value in the context, and if it is missing or different, the tool will not be enabled.                                                                       |
    | Tool Filter    | Select        | Controls if tools should be excluded from execution. You can select one of the following options:<ul><li>**None** — no tool filtering is applied, and all tools are available for execution. This option is selected by default.</li><li>**Whitelist** — only tools on the list are allowed for execution, while all other tools are excluded.</li><li>**Blacklist** — tools on the list are excluded from execution, while all other tools remain available.</li></ul> |
    | Blacklist      | CognigyScript | This parameter appears if you select **Blacklist** in **Tool Filter**. Specify the tools that should be blocked from execution. Specify only one tool per field.                                                                                                                                                                                                                                                                                                        |
    | Whitelist      | CognigyScript | This parameter appears if you select **Whitelist** in **Tool Filter**. Specify the tools you want to allow for execution. Specify only one tool per field.                                                                                                                                                                                                                                                                                                              |
    | Custom Headers | -             | Sets custom authentication headers to send with the request to the MCP server. Use the **Key** and **Value** fields to enter a header. The **Value** field supports CognigyScript. After entering the header key, new empty **Key** and **Value** fields are automatically added, in case you need to add more headers. Alternatively, you can click **Show JSON Editor** and enter the headers in the code field.                                                      |
  </Accordion>
</AccordionGroup>

#### Call MCP Tool

In the Flow editor, when you add an MCP Tool Node, a Call MCP Tool Node is automatically created below it. These two Nodes work together to define and execute the chosen tool. The Call MCP Tool Node sets the actual execution point of the chosen tool. This way, you can verify or modify the tool call arguments in the `input.aiAgent.toolArgs` object, or add a Say Node before the tool call. When the Call MCP Tool Node is executed, the tool call is sent to the remote MCP server, where the Tool is executed remotely with any arguments set by the AI Agent.

To return the tool result to the AI Agent, the **Resolve Immediately** setting can be enabled to send the full result returned from the remote MCP server to the AI Agent.

As an alternative, use a [Resolve Tool Action](/ai/agents/develop/node-reference/ai/resolve-tool-action) Node to return a specific result to the AI Agent.

<AccordionGroup>
  <Accordion title="Call MCP Tool">
    | Parameter           | Type   | Description                                                                                           |
    | ------------------- | ------ | ----------------------------------------------------------------------------------------------------- |
    | Resolve Immediately | Toggle | Enable this parameter to immediately resolve the tool action with the full result as the tool answer. |
  </Accordion>

  <Accordion title="Storage Options">
    | Parameter                   | Type          | Description                                                                                                                                                                                                                               |
    | --------------------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | How to handle the result    | Select        | Determine how to handle the MCP tool call result:<ul><li>**Store in Input** — stores the result in the Input object.</li><li>**Store in Context** — stores the result in the Context object.</li></ul>                                    |
    | Input Key to store Result   | CognigyScript | The parameter appears when **Store in Input** is selected. The result is stored in the `input.aiAgent.toolResult` object by default. You can specify another value, but the **MCP Tool Result** Token won't work if the value is changed. |
    | Context Key to store Result | CognigyScript | The parameter appears when **Store in Context** is selected. The result is stored in the `context.aiAgent.toolResult` object by default.                                                                                                  |
  </Accordion>

  <Accordion title="Debug Settings">
    | Parameter         | Type   | Description                                                                             |
    | ----------------- | ------ | --------------------------------------------------------------------------------------- |
    | Debug Tool Result | Toggle | Enable the output of a debug message with the tool call result after a successful call. |
  </Accordion>
</AccordionGroup>

#### Knowledge Tool

The Knowledge Tool Node is a child Node of the AI Agent Node. It allows the AI Agent to directly access [Knowledge Stores](/ai/agents/develop/knowledge-ai/knowledge-store) to provide context-aware responses.

In the Knowledge Tool Node, you can select the [Knowledge Store](/ai/agents/develop/knowledge-ai/knowledge-store) to search, [Source Tags](/ai/agents/develop/knowledge-ai/knowledge-source/knowledge-source-tags) to refine the search, and allows for detailed debug messages about the tool's execution.

<AccordionGroup>
  <Accordion title="Knowledge Tool">
    | Parameter       | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
    | --------------- | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Knowledge Store | Selector      | Select a Knowledge Store.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
    | Tool ID         | CognigyScript | Enter a name as a tool ID. This ID can contain only letters, numbers, underscores (`_`), or dashes (`-`). If you use more than one Knowledge tool, the tool ID provides context to trigger the correct Knowledge tool. In this case, enter a clear tool ID for each Knowledge tool, for example, `search_appliances`                                                                                                                                                                                                                                                                                                                        |
    | Description     | CognigyScript | Enter instructions to guide the AI Agent to call the Knowledge tool. The description field provides context to trigger the Knowledge tool. If you want to use more than one Knowledge tool, enter clear instructions for the cases when each Knowledge tool should be used. For example, `Find the answer to prompts or questions about appliances by searching the attached data sources. Use this tool when a customer asks about appliance items such as washing machines, dryers, and other household appliances. Focus exclusively on a knowledge search and does not execute tasks like small talk, calculations, or script running.` |
  </Accordion>

  <Accordion title="Debug Settings">
    | Parameter                 | Type   | Description                                                                                                       |
    | ------------------------- | ------ | ----------------------------------------------------------------------------------------------------------------- |
    | Debug Message when called | Toggle | Enable the output of a debug message when the tool is called to provide detailed information about the tool call. |
  </Accordion>

  <Accordion title="Advanced">
    | Parameter                   | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
    | --------------------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Top K                       | Slider        | Set how many Knowledge Chunks to return. Providing more results gives the AI Agent additional context, but it also increases noise and token usage.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
    | Store Location              | Selector      | Select whether and where to store the knowledge search results. Select one of the following options:<ul><li>**Don't store** — the content isn't stored. This option is set by default.</li><li>**Input** — the content is stored in the Input object.</li><li>**Context** — the content is stored in the Context object.</li></ul>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
    | Input Key to store result   | CognigyScript | Appears when **Store Location** is set to **Input**. The property in the Input object where the result is stored. For example, `input.knowledgeSearch`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | Context Key to store result | CognigyScript | Appears when **Store Location** is set to **Context**. The property in the Context object where the result is stored. For example, `context.knowledgeSearch`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
    | Source Tags                 | CognigyScript | Enter [Knowledge Source Tags](/ai/agents/develop/knowledge-ai/knowledge-source/knowledge-source-tags) to refine the scope of your knowledge search, including only the most relevant Knowledge Chunks in the Knowledge Store. Before entering tags, ensure they are included in the Knowledge Sources. Add tags by entering each separately and pressing <kbd>Enter</kbd>. Max 5 tags. When multiple Source Tags are specified, the Search Extract Output Node defaults to the `AND` operator, meaning it only considers Sources that have all specified tags. To change this behavior, adjust the **Match Type for Source Tags** parameter.                                                                                                                                                                      |
    | Match type for Source Tags  | Selector      | The operator to filter Knowledge Sources by Source Tags. Select one of the following options:<ul><li>**AND** — the default value, requires all Source Tags to match across multiple Knowledge Sources. Consider the following example: there are Knowledge Sources with Source Tags `S-a`, `S-b`, and `S-c`. When you use the `AND` operator to filter by `S-a` and `S-b`, only Sources with both Tags `S-a` and `S-b` are included in the search results.</li><li>**OR** — requires at least one Source Tag to match across multiple Knowledge Sources. Consider the following example: there are Knowledge Sources with Source Tags `S-a`, `S-b`, and `S-c`. When you use the `OR` operator to filter by `S-a` or `S-b`, any Source with either Tag `S-a` or `S-b` is included in the search results.</li></ul> |
    | Condition                   | CognigyScript | The Knowledge tool is activated only if the condition is evaluated as true. If false, the tool isn't included in the current execution. For example, when using the Knowledge tool, you can enter a condition such as `context.productCategory === "appliances"`. This checks the value in the context, and if it is missing or different, the tool isn't activated.                                                                                                                                                                                                                                                                                                                                                                                                                                              |
  </Accordion>
</AccordionGroup>

#### Send Email Tool

The Send Email tool lets your AI Agent send emails directly to users. The Send Email tool uses the same configuration and restrictions as the [Email Notification Node](/ai/agents/develop/node-reference/data/email-notification) but adds more flexibility. While the Email Notification Node sends emails at a fixed step in a Flow, the Send Email tool allows the AI Agent to send emails dynamically, based on user input, conversation context, or instructions. This tool makes automation more flexible and minimizes extra Flow steps.

<AccordionGroup>
  <Accordion title="Tool">
    | Parameter                    | Type          | Description                                                                                                                                              |
    | ---------------------------- | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Tool ID                      | CognigyScript | Provide a meaningful name as a tool ID. This ID can contain only letters, numbers, underscores (`_`), or dashes (`-`). The default name is `send_email`. |
    | Description                  | CognigyScript | Provide a detailed description of what the tool should do. The default description is `Create and send a new email message.`                             |
    | Recipient TO Email Addresses | CognigyScript | A comma-separated list of email addresses to which the email will be sent.                                                                               |
  </Accordion>

  <Accordion title="Debug Settings">
    | Parameter                 | Type   | Description                                                                                                       |
    | ------------------------- | ------ | ----------------------------------------------------------------------------------------------------------------- |
    | Debug Message when called | Toggle | Enable the output of a debug message when the tool is called to provide detailed information about the tool call. |
  </Accordion>

  <Accordion title="Advanced">
    | Parameter | Type          | Description                                                                                                                                                                                                                                                                                                                                                                        |
    | --------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Condition | CognigyScript | The tool will be enabled only if the condition is evaluated as true. If false, the tool isn't part of the AI Agent's Tools within this execution. For example, when using the `unlock_account` tool, you can specify a condition like `context.accountStatus === "locked"`. This checks the value in the context, and if it is missing or different, the tool will not be enabled. |
  </Accordion>
</AccordionGroup>

#### Handover to AI Agent Tool

The Handover to AI Agent tool lets you transfer a conversation to another AI Agent in the same or a different Flow. This approach ensures the conversation keeps its context, different AI Agents can handle specific tasks, and multi-step conversations run smoothly across Flows.

<AccordionGroup>
  <Accordion title="Tool">
    | Parameter   | Type          | Description                                                                                                                                                                                                                                                     |
    | ----------- | ------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Tool ID     | CognigyScript | Provide a meaningful name as a tool ID. This ID can contain only letters, numbers, underscores (`_`), or dashes (`-`). The default name is `handover_to_ai_agent`.                                                                                              |
    | Description | CognigyScript | Describe what the receiving AI Agent should do. For example: `Handle product recommendations`, `Perform technical troubleshooting`, or `Assist with billing questions`. Be specific so the handover rules clearly indicate when this AI Agent should take over. |
    | Select Flow | Selector      | Select the target Flow to hand over to. By default, all Flows in the Project are displayed in the list. The selected Flow must contain an AI Agent Node.                                                                                                        |
    | Select Node | Selector      | Select the AI Agent Node to hand over to.                                                                                                                                                                                                                       |
  </Accordion>

  <Accordion title="Debug Settings">
    | Parameter                 | Type   | Description                                                                                                       |
    | ------------------------- | ------ | ----------------------------------------------------------------------------------------------------------------- |
    | Debug Message when called | Toggle | Enable the output of a debug message when the tool is called to provide detailed information about the tool call. |
  </Accordion>

  <Accordion title="Advanced">
    | Parameter | Type          | Description                                                                                                                                                                                                                                                                                                                                                                        |
    | --------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Condition | CognigyScript | The tool will be enabled only if the condition is evaluated as true. If false, the tool isn't part of the AI Agent's Tools within this execution. For example, when using the `unlock_account` tool, you can specify a condition like `context.accountStatus === "locked"`. This checks the value in the context, and if it is missing or different, the tool will not be enabled. |
  </Accordion>
</AccordionGroup>

#### Handover to Human Agent Tool

The Handover to Human Agent tool lets you transfer a conversation to a human agent through handover providers. This tool allows your AI Agents to refer complex tasks to a human agent if they can't address the user's request.

<AccordionGroup>
  <Accordion title="Tool">
    | Parameter                        | Type          | Description                                                                                                                                                                                                                                                                                                                                            |
    | -------------------------------- | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Tool ID                          | CognigyScript | Provide a meaningful name as a tool ID. This ID can contain only letters, numbers, underscores (`_`), or dashes (`-`). The default name is `handover_to_human_agent`.                                                                                                                                                                                  |
    | Description                      | CognigyScript | Describe the situation when the AI Agent should refer to a human agent. For example: `Use this tool if the user explicitly asks for a human.`, or `Use this tool if you fail to answer the user's question three times.`. Be specific so the handover rules clearly indicate when the conversation should be handed over to the receiving human agent. |
    | Handover Provider                | Selector      | Select the handover provider from the list. The list contains the handover providers created in **Deploy > Handover Providers**.                                                                                                                                                                                                                       |
    | Handover Accepted Message        | CognigyScript | The message to be outputted when the handover is requested.                                                                                                                                                                                                                                                                                            |
    | Repeat Handover Accepted Message | Toggle        | Enable to repeatedly output the Handover Accepted Message if the user sends a message while waiting in the queue for an agent.                                                                                                                                                                                                                         |
  </Accordion>

  <Accordion title="Live Agent Settings">
    | Parameter                       | Description                                                                                                                                                                                                                                                                                                                                                                  |
    | ------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Live Agent Inbox Id             | Enter the Inbox ID that you obtained in the [previous step](/ai/escalate/handover-reference/live-agent).                                                                                                                                                                                                                                                                     |
    | Skills                          | Enter a list of skills that should be used to filter the available agents. For example, `technical support, billing`.                                                                                                                                                                                                                                                        |
    | Languages                       | Enter a list of languages that should be used to filter the available agents. For example, `english, spanish`.                                                                                                                                                                                                                                                               |
    | Priority                        | Select a priority for the conversation from the list. For example, `Urgent`.                                                                                                                                                                                                                                                                                                 |
    | Get Queue Position              | Enable the system to retrieve the end user's current queue position. This parameter is inactive by default.                                                                                                                                                                                                                                                                  |
    | Update Interval                 | Specify the interval, in seconds, for how often to receive updates on the queue status. This parameter appears if **Get Queue Position** is selected.                                                                                                                                                                                                                        |
    | Get Estimated Wait Time         | Enable the retrieval of the estimated wait time for the end user. This parameter is inactive by default.                                                                                                                                                                                                                                                                     |
    | Update Interval                 | Set the interval, in seconds, for how often to receive updates on estimated wait time. This parameter appears if **Get Estimated Wait Time** is selected.                                                                                                                                                                                                                    |
    | Alternative Update              | The alternative update options when certain conditions are met. This parameter is inactive by default.                                                                                                                                                                                                                                                                       |
    | Maximum Queue Position          | Set the threshold for the maximum queue position. If this threshold is reached, the alternative text is sent. This parameter appears if **Alternative Update** is selected.                                                                                                                                                                                                  |
    | Maximum Estimated Wait Time     | Define the maximum estimated wait time in milliseconds. If this duration is exceeded, the alternative text is sent. This parameter appears if **Alternative Update** is selected.                                                                                                                                                                                            |
    | Alternative Text                | The message to be sent when either the maximum queue position or maximum wait time is reached. This parameter appears if **Alternative Update** is selected.                                                                                                                                                                                                                 |
    | Enable User Connects Message    | Notify human agents when an end user reconnects to the chat. The parameter is enabled by default. When the parameter is enabled, the message `User joined the conversation` appears in the chat as soon as the end user returns to the chat tab by clicking the **←** (back arrow) at the top bar in the browser, after having opened a new URL on the same tab as the chat. |
    | Enable User Disconnects Message | Notify human agents when an end user disconnects from the chat. The parameter is enabled by default. When the parameter is enabled, the message `User left the conversation` is sent as soon as the end user closes the tab with the chat or switches to a new URL address within the current tab.                                                                           |

    For more information, read the [Live Agent Handover Setup](/live-agent/getting-started/live-agent-setup/live-agent-setup-handover-flow) and [Real-Time Queue Notifications](/ai/escalate/handover-features/real-time-queue-notifications) documentation.
  </Accordion>

  <Accordion title="Chatwoot Settings">
    | Parameter         | Description                                                                                                                                                                                                                                                                                          |
    | ----------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Chatwoot Inbox Id | Enter the Chatwoot Inbox ID. You can find the Inbox ID in your Chatwoot URL. For example, `https://app.chatwoot.com/app/account/6607/inbox/5446`, where the Inbox ID is `5446`. For more information, read the [Chatwoot handover provider](/ai/escalate/handover-reference/chatwoot) documentation. |
  </Accordion>

  <Accordion title="Salesforce Chat Settings">
    | Parameter                        | Description                                                                                                                                                                                                                                        |
    | -------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Send Transcript as first message | Activate this parameter to allow Salesforce to receive the conversation transcript as the first message. This setting is turned off by default.                                                                                                    |
    | Button ID                        | Enter the ID of the Button you created in the Salesforce installation. This setting will override the default Button ID that you specified in the Endpoint configuration. You can use different Button IDs for individual Handover to Agent Nodes. |
    | Pre-chat Entities                | Enter the pre-chat entities. For more information, read the [Salesforce Handover](/ai/escalate/handover-reference/salesforce) documentation.                                                                                                       |
    | Pre-chat Details                 | Enter the pre-chat details. For more information, read the [Salesforce Handover](/ai/escalate/handover-reference/salesforce) documentation.                                                                                                        |
  </Accordion>

  <Accordion title="NiCE CXone Settings">
    | Parameter                        | Description                                                                                                                                                                            |
    | -------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Send Transcript as first message | Activate this parameter to allow NiCE CXone to receive the conversation transcript as the first message when the handover begins. This setting is turned on by default.                |
    | Display Agent Details            | Display the human agent's name and avatar from NiCE CXone in the chat for the end user. This parameter is disabled by default.                                                         |
    | Custom Attributes                | Add custom attributes in JSON format to include additional information the handover. For more information, refer to [NiCE CXone](/ai/escalate/handover-reference/nice-cxone/overview). |

    Example for the **Custom Attributes** parameter:

    ```json theme={null}
    {
      "section": "section A",
      "region": "Europe"
    }
    ```
  </Accordion>

  <Accordion title="Genesys Settings">
    | Parameter                       | Description                                                                                                                                                                                                                                                                                                                                                                  |
    | ------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Language                        | Specify a language for the conversation. For example, `english`, `spanish`, `german`.                                                                                                                                                                                                                                                                                        |
    | Skills                          | Define skills for the conversation. For example, `escalation`.                                                                                                                                                                                                                                                                                                               |
    | Priority                        | Set the priority for the conversation. For example, `1`. <br /> If a priority is set, it triggers a flow in Genesys to prioritize or de-prioritize the conversation within the queue. <br /> Note that this functionality requires the appropriate flow to be set up in Genesys.                                                                                             |
    | Enable User Connects Message    | Notify human agents when an end user reconnects to the chat. The parameter is enabled by default. When the parameter is enabled, the message `User joined the conversation` appears in the chat as soon as the end user returns to the chat tab by clicking the **←** (back arrow) at the top bar in the browser, after having opened a new URL on the same tab as the chat. |
    | Enable User Disconnects Message | Notify human agents when an end user disconnects from the chat. The parameter is enabled by default. When the parameter is enabled, the message `User left the conversation` is sent as soon as the end user closes the tab with the chat or switches to a new URL address within the current tab.                                                                           |
    | Custom Attributes               | Add custom attributes in the `Key: Value` format, which allows you to include additional information.                                                                                                                                                                                                                                                                        |
    | Get Estimated Wait Time         | Retrieve the estimated wait time for the end user. End users then receive wait time updates while they wait to be connected to a human agent. This parameter is inactive by default. It requires an active [Amazon EventBridge Source integration](/ai/escalate/handover-reference/genesys-cloud-open-messaging-with-amazon-eventbridge).                                    |
    | Update Interval                 | Specify the interval, in seconds, for how often to receive updates on the estimated wait time. This parameter appears if **Get Estimated Wait Time** is selected.                                                                                                                                                                                                            |
    | Alternative Update              | Send an alternative message to the end user when the maximum estimated wait time is exceeded. This parameter is inactive by default. This parameter appears if **Get Estimated Wait Time** is selected.                                                                                                                                                                      |
    | Maximum Estimated Wait Time     | Define the maximum estimated wait time in milliseconds. If this duration is exceeded, the alternative text is sent instead of the actual wait time. This parameter appears if **Alternative Update** is selected.                                                                                                                                                            |
    | Alternative Text                | Specify the message to be sent when the maximum estimated wait time is exceeded. This parameter appears if **Alternative Update** is selected.                                                                                                                                                                                                                               |

    Genesys Cloud Open Messaging doesn't provide a queue position. Estimated wait time is the only queue parameter available.

    Example for the **Custom Attributes** parameter:

    ```json theme={null}
    {
      "section": "section A",
      "region": "Europe"
    }
    ```

    For more information,
    read
    the [Genesys Cloud Guest Chat](/ai/escalate/handover-reference/genesys-cloud-guest-chat#prerequisites), [Genesys Cloud Open Messaging](/ai/escalate/handover-reference/genesys-cloud-open-messaging), and [Real-Time Queue Notifications](/ai/escalate/handover-features/real-time-queue-notifications#configure-settings-in-genesys-cloud-open-messaging) documentation.
  </Accordion>

  <Accordion title="8x8 Settings">
    | Parameter           | Description                                                                                                                                                                                                                                                                      |
    | ------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | 8x8 Channel Id      | Specify the channel ID that you created during the 8x8 configuration. <br /> You can find this ID in the 8x8 admin console. Navigate to **Channels > Chat API**. In the channel list, find the channel that you created. Copy the **ID** from the channel row.                   |
    | 8x8 Queue Id        | Specify the header value required for authenticating API calls to the 8x8 system. <br /> You can find this ID in the 8x8 admin console. Navigate to **Channels > Chat API**. In the channel list, find the channel that you created. Copy the **Queue ID** from the channel row. |
    | 8x8 JSON properties | Extend 8x8 properties with custom JSON.                                                                                                                                                                                                                                          |

    For more information, read the [8x8 Handover Setup](/ai/escalate/handover-reference/8x8) documentation.
  </Accordion>

  <Accordion title="RingCentral Engage Settings">
    | Parameter                       | Type          | Description                                                                                                                                                                                                                                                                                                                                                                  |
    | ------------------------------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Additional Category Ids         | CognigyScript | IDs to be passed to the RingCentral Engage handover provider.                                                                                                                                                                                                                                                                                                                |
    | Enable User Connects Message    | Toggle        | Notify human agents when an end user reconnects to the chat. The parameter is enabled by default. When the parameter is enabled, the message `User joined the conversation` appears in the chat as soon as the end user returns to the chat tab by clicking the **←** (back arrow) at the top bar in the browser, after having opened a new URL on the same tab as the chat. |
    | Enable User Disconnects Message | Toggle        | Notify human agents when an end user disconnects from the chat. The parameter is enabled by default. When the parameter is enabled, the message `User left the conversation` is sent as soon as the end user closes the tab with the chat or switches to a new URL address within the current tab.                                                                           |

    For more information, read the [RingCentral Engage Integration](/ai/escalate/handover-reference/ring-central-engage) documentation.
  </Accordion>

  <Accordion title="Cancel Handover Options">
    | Parameter          | Type          | Description                                                                                          |
    | ------------------ | ------------- | ---------------------------------------------------------------------------------------------------- |
    | Cancel Intent      | CognigyScript | The Intent the user has to trigger to cancel the Handover request.                                   |
    | Cancel Button Text | CognigyScript | The text for a quick reply button that cancels the handover request when the user clicks the button. |
  </Accordion>

  <Accordion title="On Resolve Options">
    | Parameter                           | Type     | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | ----------------------------------- | -------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Flow Continuation                   | Selector | Controls where the Flow continues after the human agent leaves the conversation.<br />When the user needs assistance completing the Flow with the AI Agent, the conversation is handed over to a human agent. The human agent can trigger the Flow by sending a `/bot` command as a reply in the conversation interface.<br />For example, the user enters their date of birth in an incorrect format, and the AI Agent can't process the response. The conversation is handed over to a human agent, who identifies the incorrect format and sends the corrected date using `/bot MM/DD/YYYY`. The AI Agent takes over and continues the Flow.<br /><ul><li>**Below this Node** — the Flow continues from the first Node below the Human Agent Node, regardless of where the human agent stopped assisting the user. This option is selected by default.</li> <li>**At current AgentInject Entrypoint** — the Flow continues from the Node where the human agent stopped assisting the user.</li></ul> |
    | Send resolve event to Virtual Agent | Toggle   | Controls whether the Flow receives the `conversation_status_changed` event with the status set to `resolved`. This parameter is activated by default.<ul><li>When the parameter is activated, the event is sent when the human agent leaves the conversation. As a result, the Flow continues immediately, without waiting for the user to send another message. If the [Lookup Case](/ai/agents/develop/node-reference/logic/lookup) Node is configured with **Handover Status** set to **Completed**, the Node can detect the completed handover and continue the Flow.</li><li>When the parameter is deactivated, the event isn't sent when the human agent leaves the conversation. The Flow continues after the user sends their next message.</li></ul>                                                                                                                                                                                                                                           |
  </Accordion>

  <Accordion title="Event Settings">
    | Parameter         | Type   | Description                                             |
    | ----------------- | ------ | ------------------------------------------------------- |
    | Send Queue Event  | Toggle | Sends an event to the Flow once the handover is queued. |
    | Send Active Event | Toggle | Sends an event to the Flow once the handover is active. |
  </Accordion>

  <Accordion title="Debug Settings">
    | Parameter                 | Type   | Description                                                                                                       |
    | ------------------------- | ------ | ----------------------------------------------------------------------------------------------------------------- |
    | Debug Message when called | Toggle | Enable the output of a debug message when the tool is called to provide detailed information about the tool call. |
  </Accordion>

  <Accordion title="Advanced">
    | Parameter | Type          | Description                                                                                                                                                                                                                                                                                                                                                                        |
    | --------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Condition | CognigyScript | The tool will be enabled only if the condition is evaluated as true. If false, the tool isn't part of the AI Agent's Tools within this execution. For example, when using the `unlock_account` tool, you can specify a condition like `context.accountStatus === "locked"`. This checks the value in the context, and if it is missing or different, the tool will not be enabled. |
  </Accordion>
</AccordionGroup>

#### Execute Workflow Tool

The Execute Workflow tool lets the AI Agent execute another Flow. After the target Flow executes, the conversation returns to the AI Agent.

The target Flow inherits the [Context object](/ai/agents/develop/ai-agent-memory/context) from the AI Agent. If the Context object is changed during the target Flow execution, these changes are also available to the AI Agent.

| Parameter   | Description                                                                                                                                                                                                                                                                                     |
| ----------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Tool ID     | Enter a meaningful name as a tool ID. This ID can contain only letters, numbers, underscores (`_`), or dashes (`-`). The default name is `execute_workflow`.                                                                                                                                    |
| Description | Describe when to trigger the AI Agent to execute the target Flow. For example, if you have a specific Flow for dealing with package returns, enter: `Use this tool if the user explicitly asks to return a package.`. Be specific so the AI Agent only executes the target Flow when necessary. |
| Target Flow | Select the target Flow that the AI Agent executes when this tool is called. Click <img alt="open target flow" /> to open the target Flow.                                                                                                                                                       |
| Select Node | Select the target Node that the AI Agent executes when the target Flow is called. You must set the Node you want to select as an Entrypoint with the [Set Entrypoint](/ai/agents/develop/nodes/overview#set-nodes-as-entrypoints) setting before you can select it.                             |

<AccordionGroup>
  <Accordion title="Debug Settings">
    | Parameter                 | Type   | Description                                                                                                         |
    | ------------------------- | ------ | ------------------------------------------------------------------------------------------------------------------- |
    | Debug Message when called | Toggle | Activate the output of a debug message when the tool is called to provide detailed information about the tool call. |
  </Accordion>

  <Accordion title="Advanced">
    | Parameter      | Type   | Description                                                                                                                                                                                                                                                                                                                                                                                                                              |
    | -------------- | ------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Parse Intents  | Toggle | Sets whether Intents should be parsed in the target Flow. This parameter is activated by default.<br /><br />If this parameter is enabled, Cognigy NLU reparses the Intents in the target Flow based on `text` in the [Input object](/ai/agents/develop/ai-agent-memory/input#accessing-input-properties) and generates Intent scores using the Intents in the target Flow instead of the ones in the Flow the AI Agent is currently in. |
    | Parse Slots    | Toggle | Sets whether Slots should be parsed in the target Flow. This parameter is activated by default.<br /><br />If this parameter is enabled, Cognigy NLU reparses the Slots in the target Flow based on `text` in the [Input object](/ai/agents/develop/ai-agent-memory/input#accessing-input-properties) and generates Slot scores using the Slots in the target Flow instead of the ones in the Flow the AI Agent is currently in.         |
    | Absorb Context | Toggle | Sets whether the [default Context object](/ai/agents/develop/ai-agent-memory/context#default-context) of the target Flow is used in the context of the Flow the AI Agent is currently in.<br /><br />If the default Context of the target Flow has a value, it is copied into the Context object of the AI Agent as soon as the target Flow is executed.                                                                                 |
  </Accordion>
</AccordionGroup>

## Examples

### Tool

In this example, the `unlock_account` tool unlocks a user account by providing the email and specifying the reason for the unlocking.

Parameter configuration in JSON:

```json theme={null}
{
  "type": "object",
  "properties": {
    "email": {
      "type": "string",
      "description": "User's login email for their account."
    }
  },
  "required": ["email"],
  "additionalProperties": false
}
```

where:

* `type` — the type for a tool parameter schema, which must always be `object`.
* `properties` — defines the parameters for the tool configuration:
  * `email` — a required tool parameter for unlocking the account.
    * `type` — defines the data type for the tool parameter.
    * `description` — a brief explanation of what the property represents.
* `required` — lists `email` as a required parameter, ensuring that this value is always provided when the tool is called.
* `additionalProperties` — ensures that the input contains only the `email` tool parameter, and no others are allowed.

### MCP Tool and Call MCP Tool

#### Use Zapier's Remote MCP server

You can create a custom MCP server with personalized tools by using one of the [provided SDKs](https://github.com/modelcontextprotocol). For a quicker setup, you can use a third-party provider. For example, [Zapier](https://zapier.com/) allows you to [configure your MCP server](https://actions.zapier.com/settings/mcp/), which can be connected to multiple application APIs.

To use Zapier as a remote MCP server, follow these steps:

1. Log in to your Zapier account, go to the [MCP settings](https://actions.zapier.com/settings/mcp/) page, and configure your MCP server.
2. Copy the SSE URL and paste it into the **MCP Server SSE URL** field of your MCP Tool Node.
3. In the Zapier MCP settings, create an action to connect to various APIs. For example, you can create a Zapier action to automatically generate a Google Doc.

Once the setup is complete, the configured MCP Actions will be loaded when the AI Agent is executed.

You will see the following debug message in the Interaction Panel, indicating the result of the tool call after a successful execution:

```txt theme={null}
AI Agent: MCP Tool
Fetched tools from MCP Tool "zapier"

- google_docs_create_document_from_: Create a new document from text. Also supports limited HTML.
  Parameters:
  - title (string): Document Name
  - file (string): Document Content
```

### Knowledge Tool

In this example, you have two Knowledge tools to search two different Knowledge Stores. Each Knowledge Store refers to a different product category, in this example, appliances and furniture. To guide the AI Agent to use the correct Knowledge tool, enter a clear tool ID and description:

* Knowledge Tool 1:
  * Tool ID: `search_appliances`
  * Description: `Find the answer to prompts or questions about appliances by searching the attached data sources. Use this tool when a customer asks about appliance items such as washing machines, dryers, and other household appliances. Focus exclusively on a knowledge search and does not execute tasks like small talk, calculations, or script running.`
* Knowledge Tool 2:
  * Tool ID: `search_furniture`
  * Description: `Find the answer to prompts or questions about furniture by searching the attached data sources. Use this tool when a customer asks about furniture items such as sofas, tables, and other items. Focus exclusively on a knowledge search and does not execute tasks like small talk, calculations, or script running.`

### Avoiding Infinite Loops when Tool Choice is Required

When **Tool Choice** is set to **Required**, the model must call a tool on every turn. After a tool branch executes and a [Resolve Tool Action](/ai/agents/develop/node-reference/ai/resolve-tool-action) Node returns the result to the AI Agent Node, the model still can't respond with plain text. Instead, it must call another tool. If every tool branch loops back through a **Resolve Tool Action** Node, the AI Agent keeps calling tools without ever producing a response. This loop continues until the Flow times out and can consume a large number of tokens.

<Warning>
  Avoid setting **Tool Choice** to **Required** unless your use case requires the model to always call a tool. If you do use this setting, ensure that at least one tool branch exits the loop without passing through a Resolve Tool Action Node.
</Warning>

To prevent an infinite loop, use one of the following approaches:

<AccordionGroup>
  <Accordion title="Option 1: Change Tool Choice to Auto">
    Set **Tool Choice** to **Auto** unless your use case strictly requires the model to always use a tool. With **Auto**, the model calls a tool when relevant and returns a plain-text response when the tool isn't needed, allowing the Flow to exit normally.
  </Accordion>

  <Accordion title="Option 2: Add an Exit Tool">
    Keep **Tool Choice** set to **Required** and add a dedicated tool that the model can call to end the loop. This tool branch must not include a Resolve Tool Action Node, so execution doesn't return to the AI Agent Node. Instead, the branch must end with a Say Node or another output Node that produces the final response.

    Example exit tool configuration:

    | Field       | Value                                                                                                    |
    | ----------- | -------------------------------------------------------------------------------------------------------- |
    | Tool ID     | `finish_response`                                                                                        |
    | Description | `Call this tool when you have gathered all the information needed and are ready to respond to the user.` |

    The resulting Flow structure looks like this:

    ```text theme={null}
    AI Agent Node (Tool Choice: Required)
    ├── search_knowledge → Resolve Tool Action → [returns to AI Agent Node]
    └── finish_response  → Say Node            → [end of Flow, no Resolve Tool Action]
    ```

    The model calls `search_knowledge` one or more times to gather information, then calls `finish_response` to exit the loop and deliver the response.
  </Accordion>
</AccordionGroup>

## Get AI Agent Jobs and Tools via API

You can retrieve all job configurations and associated tools for a specific AI Agent via the Cognigy API [`GET /v2.0/aiagents/{aiAgentId}/jobs`](https://api-trial.cognigy.ai/openapi#get-/v2.0/aiagents/-aiAgentId-/jobs) request. The response includes each job's configuration details and a list of available tools.

## More Information

* [AI Agent Handover](/ai/agents/develop/node-reference/ai/ai-agent-handover)
* [Resolve Tool Action](/ai/agents/develop/node-reference/ai/resolve-tool-action)
* [Add Memories](/ai/agents/develop/node-reference/analytics/add-memory)
* [Agentic AI](/ai/agents/overview)


# AI Agent Handover
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/ai/ai-agent-handover



<Frame>
  <img alt="AI Agent Handover Node configuration panel" />
</Frame>

## Description

The AI Agent Handover Node transfers the conversation to another AI Agent Node or Flow and executes it.

## Parameters

| Parameter   | Type     | Description                                                                                                                                                                                                                       |
| ----------- | -------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Select Flow | Selector | Select the target Flow to hand over to.                                                                                                                                                                                           |
| Select Node | Selector | Select the Node to hand over to. You must set the Node as an Entrypoint before you can select it. To set the Node as an Entrypoint, use the [Set Entrypoint](/ai/agents/develop/nodes/overview#set-nodes-as-entrypoints) setting. |

### Advanced

| Parameter     | Type   | Description                                                                                       |
| ------------- | ------ | ------------------------------------------------------------------------------------------------- |
| Parse Intents | Toggle | Enable the system to check for Intents in the correct child Flow, rather than in the parent Flow. |
| Parse Slots   | Toggle | Enable the system to check for Slots in the correct child Flow, rather than in the parent Flow.   |

## More Information

* [AI Agent](/ai/agents/develop/node-reference/ai/ai-agent)
* [Resolve Tool Action](/ai/agents/develop/node-reference/ai/resolve-tool-action)
* [Agentic AI](/ai/agents/overview)


# Clean Text
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/ai/clean-text



<a href="/release-notes/4.51"><Badge>Updated in 4.51</Badge></a>

<Frame>
  <img alt="Clean Text Node configuration panel" />
</Frame>

## Description

The Clean Text Node enables a user to clean a provided text using the Text Cleaner class, which exposes various functions for cleaning text.

The result is either written to the Input or Context objects.

See the [Text Cleaner](/ai/platform-features/nlu/text-cleaner) page for more information on the individual functions.


# Copy Slots to Context
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/ai/copy-slots-to-context



<Frame>
  <img alt="Copy Slots to Context Node configuration panel" />
</Frame>

## Description

The Copy Slots to Context Node stores a found Slot in a specific place in the Context.

## Parameters

| Parameter | Type          | Description                                                                                                                                                                                                                                                                                                        |
| --------- | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| Key       | CognigyScript | The place in the Context where the found Slot should be stored.                                                                                                                                                                                                                                                    |
| Tag       | CognigyScript | The tag of the Slot that should be stored in the Context if found (e.g., a food).                                                                                                                                                                                                                                  |
| Mode      | Select        | <ul><li>`simple` – only the last found Slot is stored.</li><li>`array` – all found Slots with the correct tag are stored as an array.</li></ul>                                                                                                                                                                    |
| UseNeg    | Toggle        | <ul><li>`true`– only positive Slots are added to the Context. This means that in the sentence `I don't like pizza`, `pizza` is not stored in the Context because it is negated.</li><li>`false` – negative Slots are also added to the context. In the sentence `I don't like pizza`, `pizza` is stored.</li></ul> |

## Example

You create a Copy Slots To Context Node and set:

* `key` to `foods`.
* `tag` to `food`.
* `mode` to `array`.
* `UseNeg` to `false`.

You attach a Lexicon with the Keyphrases `salad` and `pizza` with the `food` tag. If a user says `I want to order a pizza and a salad`, then the Keyphrases `pizza` and `salad` are stored in an array in the Context under the key `foods`.


# Disable Slot Fillers
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/ai/disable-slot-fillers



<Frame>
  <img alt="Disable Slot Fillers Node configuration panel" />
</Frame>

## Description

The Disable Slot Fillers Node disables any active Slot Fillers. For more information on Slot Fillers, refer to [Natural Language Understanding (NLU)](/ai/platform-features/nlu/overview).


# Enable Slot Fillers
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/ai/enable-slot-fillers



<Frame>
  <img alt="Enable Slot Fillers Node configuration panel" />
</Frame>

## Description

The Enable Slot Fillers Node enables a particular Slot Filler. Use the dropdown menu to select the Slot Filler that you want to enable. For more information on Slot Fillers, refer to the [Natural Language Understanding (NLU)](/ai/platform-features/nlu/overview) page.


# Execute Cognigy NLU
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/ai/execute-cognigy-nlu



<Frame>
  <img alt="Execute Cognigy NLU Node configuration panel" />
</Frame>

## Description

The Execute Cognigy NLU Node executes the Cognigy NLU pipeline and updates the [Input](/ai/agents/develop/ai-agent-memory/input) or [Context](/ai/agents/develop/ai-agent-memory/context) object. For example, you can use a Code Node to modify a user's input and subsequently execute the NLU pipeline with the modified user input to obtain fresh NLU results.

The Execute Cognigy NLU Node doesn't trigger standard responses in Intents nor execute the Flow based on a triggered Intent.

<AccordionGroup>
  <Accordion title="Settings">
    | Parameter               | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | ----------------------- | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Text                    | CognigyScript | Use `{{ input.text }}` or the Answer Token.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | Data                    | JSON          | Data to send to the NLU pipeline. Can be used to modify `input.data`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
    | Mode                    | Select        | Three modes are available to control how the output of the NLU pipeline is stored. The available selections are:<ul><li>**Overwrite** — overwrites the Input object with the result of the NLU pipeline.</li><li>**Input** — stores the result of the NLU pipeline in the Input object under the key specified in the **Input Key to write to** field. This option is useful if you are running the NLU pipeline a second time with different text than the user input.</li><li>**Context** — stores the result of the NLU pipeline in the Context object under the key specified in the **Context Key to write to** field.</li></ul> |
    | Context Key to write to | CognigyScript | This field appears if the **Write to Context** option is selected.<br /><br /> Store the result of the NLU Pipeline in the Context object.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | Input Key to write to   | CognigyScript | This field appears if the **Write to Input** option is selected.<br /><br /> Store the result of the NLU Pipeline in the Input object.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
  </Accordion>

  <Accordion title="Advanced">
    | Parameter                    | Type   | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
    | ---------------------------- | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Parse Intents                | Toggle | Determine whether Cognigy should parse Intents in the child Flow. The parameter is enabled by default.<br /><br />If the parameter is enabled, the NLU pipeline reparses the Intents in the target Flow, using either [input.text](/ai/agents/develop/ai-agent-memory/input#accessing-input-properties) or the [Optionally Injected Text](/ai/agents/develop/node-reference/logic/go-to).<br /><br />The NLU pipeline parses the input into the child Flow, and the NLU pipeline then scores this text using the available Intents in the child Flow, not based on Intents in the parent Flow.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |
    | Parse System Slots           | Toggle | Determine whether Cognigy should parse System Slots in the child Flow. This parameter is enabled by default.<br /><br />If the parameter is enabled, the NLU pipeline reparses the Slots in the target Flow, using either [input.text](/ai/agents/develop/ai-agent-memory/input#accessing-input-properties) or the [Optionally Injected Text](/ai/agents/develop/node-reference/logic/go-to).<br /><br />The NLU pipeline parses the input into the child Flow, and the NLU scores this text using the available Slots in the child Flow, not based on System Slots in the parent Flow.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
    | Parse Slots                  | Toggle | Determine whether Cognigy should parse Slots in the child Flow. This parameter is enabled by default.<br /><br />If the parameter is enabled, the NLU pipeline reparses the Slots in the target Flow, using either [input.text](/ai/agents/develop/ai-agent-memory/input#accessing-input-properties) or the [Optionally Injected Text](/ai/agents/develop/node-reference/logic/go-to).<br /><br />The NLU pipeline parses the input into the child Flow, and the NLU scores this text using the available Slots in the child Flow, not based on Slots in the parent Flow.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
    | Find Type                    | Toggle | Identify the type of user input. This parameter is enabled by default.<br /><br />If the user's answer contains words associated with a particular type of question, the system identifies the type of user input accordingly. The following types can be recognized:<ul><li>**whQuestion** — questions about specific details that include question words. For example, questions with `who`, `whom`, `when`, `where`, `why`, `whose`, or `what`.</li><li>**ynQuestion** — questions that are answered with yes or no. These questions contain auxiliary verbs. For example, questions with `am`, `are`, `is`, `was`, `were`, `do`, or `does`.</li><li>**howQuestion** — questions that start with the word `how`.</li><li>**Greeting** — common expressions used to greet someone. For example, `hi`, `hello`, `greetings`, `hey`, or `howdy`.</li><li>**BGreeting** — common expressions used to say farewell or goodbye. For example, `bye`, `goodbye`, or `see you`.</li><li>**nAnswer** — negative responses or negations, including `no`, `nope`, `na`, `nah`, `not`, or `false`.</li><li>**pAnswer** — positive responses or confirmations, such as `yes` or `super`.</li><li>**Statement** — any other user responses. If the system cannot identify the type of user input or the **Find Type** is disabled, it classifies it as a statement.</li></ul><br />Note that certain user responses may be ambiguous and could potentially be misinterpreted by the agent. For example:<br /><ul><li>The user input `Why not` is interpreted as a **whQuestion**.</li><li>The user input `My name only, yes` is interpreted as a **Statement**.</li><li>The user input `Could you help me?` is interpreted as a **Statement**</li></ul> |
    | Process Intent Default Reply | Toggle | By default, this parameter is disabled, so the default reply isn't processed. If the option is enabled, the default reply is handled in the same manner as when triggered by means other than the Execute Cognigy NLU Node.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
  </Accordion>
</AccordionGroup>

## Alternatives

When you need to process user input or other text in the middle of a Flow to retrieve an NLU result, and you want to restart the Flow from the Start Node, use the [Think Node](/ai/agents/develop/node-reference/logic/think).


# Fuzzy Search
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/ai/fuzzy-search



<Frame>
  <img alt="Fuzzy Search Node configuration panel" />
</Frame>

## Description

The Fuzzy Search Node searches through a list of strings of source data using a search pattern during the Flow execution. This Node returns the best possible matches based on a set of parameters.

The Node assigns each match a score between 0 and 1, where 1 indicates the best match and 0 the worst based on the search pattern.

The matches can be stored in either the [Context](/ai/agents/develop/ai-agent-memory/context) or [Input](/ai/agents/develop/ai-agent-memory/input) object.

## Parameters

| Parameter      | Type   | Description                                                                                                                                                                                                                              |
| -------------- | ------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Search Pattern | String | The string to search for in the source data. The Fuzzy Search Node compares this pattern against each element in the array and returns the best matches. Example: `Aple`, `Jon`.                                                         |
| Source Data    | JSON   | A CognigyScript reference to the array that should be searched. The reference can point to the Context, Input, or Profile object. The type must be set to `array`. Example: `{ "$cs": { "script": "context.names", "type": "array" } }`. |

<AccordionGroup>
  <Accordion title="Basic Options">
    | Parameter          | Type    | Description                                                                                                                                                  |
    | ------------------ | ------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Is Case Sensitive  | Boolean | Specify whether comparisons should be case-sensitive.                                                                                                        |
    | Include Score      | Boolean | Specify whether the score should be included in the result set.                                                                                              |
    | Include Matches    | Boolean | Specify whether the matches should be included in the result set. When active, each record in the result set includes the indices of the matched characters. |
    | Minimum Characters | Number  | Only the matches whose length exceeds this value is returned.                                                                                                |
    | Should Sort        | Boolean | Specify whether to sort the result list by score.                                                                                                            |
  </Accordion>

  <Accordion title="Fuzzy Matching Options">
    | Parameter        | Type    | Description                                                                                                                                                        |
    | ---------------- | ------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Find All Matches | Boolean | When activated, the matching function continues to the end of a search pattern even if a perfect match has already been located in the string.                     |
    | Location         | Number  | Determine approximately where in the text the pattern is expected to be found.                                                                                     |
    | Threshold        | Number  | At what point does the match algorithm give up. A threshold of 1 requires a perfect match (of both letters and location), a threshold of 0.0 would match anything. |
    | Distance         | Number  | Determine how close the match must be to the fuzzy location (specified by Location).                                                                               |
    | Ignore Location  | Boolean | When activated, the search ignores Location and Distance, so it won't matter where in the string the pattern appears.                                              |
  </Accordion>

  <Accordion title="Storage Options">
    | Parameter                 | Type   | Description                                                          |
    | ------------------------- | ------ | -------------------------------------------------------------------- |
    | Where to store the result | List   | Specify whether the result is stored in the Input or Context object. |
    | Input Key to store Result | String | The key to store the result in.                                      |
  </Accordion>
</AccordionGroup>

## Example

In this example, the Fuzzy Search Node searches for the pattern `Jon` in an array of names stored in the Context object at `context.names`. The Node is configured to store the results in the Input object at `input.search`.
The Node returns each name in the array along with a score indicating how closely it matches the search pattern.

<Accordion title="View Example">
  1. Add data to the Context object.
     ```json theme={null}
     {
       "names": ["John", "Jonathan", "Johnny", "Alice", "Bob"]
     }
     ```
  2. Specify the search pattern in the Fuzzy Search Node.
     ```txt theme={null}
     Jon
     ```
  3. Specify the path for the source data in the Fuzzy Search Node.
     ```json theme={null}
     {
       "$cs":{
         "script":"context.names",

         "type":"array"
       }
     }
     ```
  4. Check the search results in the Input object.
     ```json theme={null}
     {
       "search": [
         {
           "item": "Jonathan",
           "refIndex": 1,
           "score": 0.999
         },
         {
           "item": "John",
           "refIndex": 0,
           "score": 0.6666666666666667
         },
         {
           "item": "Johnny",
           "refIndex": 2,
           "score": 0.6666666666666667
         },
         {
           "item": "Bob",
           "refIndex": 4,
           "score": 0.33333333333333337
         }
       ]
     }
     ```
</Accordion>

## Troubleshooting

<Accordion title="Error: this.docs.forEach is not a function">
  This error occurs when the source data provided to the Fuzzy Search Node isn't an array. The Node expects an array of strings to iterate over and calculate similarity scores.
</Accordion>


# Load AI Agent
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/ai/load-ai-agent



<Frame>
  <img alt="Load AI Agent Node configuration panel" />
</Frame>

## Description

The Load AI Agent Node lets you load the AI Agent's configuration and persona data into the Input or Context object.
The primary use case is to enhance personalization in prompt generation with the [LLM Prompt Node](/ai/agents/develop/node-reference/service/llm-prompt).
To do so, place the Load AI Agent Node above the LLM Prompt Node in the Flow editor. You can also place this Node above any other Nodes to access AI Agent information.

## Parameters

| Parameter                | Type          | Description                                                                                                                                                                                     |
| ------------------------ | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| AI Agent                 | Selector      | Select the AI Agent.                                                                                                                                                                            |
| How to handle the result | Selector      | Determine how to handle the prompt result:<ul><li>**Store in Input** — stores the result in the Input object.</li><li>**Store in Context** — stores the result in the Context object.</li></ul> |
| Input Key                | CognigyScript | The parameter appears when **Store in Input** is selected. The result is stored in the `aiAgentConfiguration` Input object by default. You can specify another key.                             |
| Context Key              | CognigyScript | The parameter appears when **Store in Context** is selected. The result is stored in the `aiAgentConfiguration` key in the Context object by default. You can specify another key.              |


# Match Pattern
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/ai/match-pattern



<Frame>
  <img alt="Match Pattern Node configuration panel" />
</Frame>

## Description

The Match Pattern Node enables a user to find patterns in text and expose them in the [Input object](/ai/agents/develop/ai-agent-memory/input).

You can define patterns to search for in the input text or in the alternate text that was provided. If a pattern is found, a `matchedPatterns` group is created in the Input object.

## Patterns

Patterns can be any text and can contain references to Slots by using the `@` symbol.

Example: `@color @product`

In `I need a green shirt`, two Slots are found and assigned to `color = green` and `product = shirt`.

### Tags

You can tag Slots in a pattern to make it easier to identify them later.

Example: `from @city>origin to @city>destination`.

In `I want to go from Düsseldorf to Tokyo`, the `city` Slot is found in a matched pattern group with origin and destination tags.

## Parameters

| Parameter               | Type        | Description                                                                   |
| ----------------------- | ----------- | ----------------------------------------------------------------------------- |
| Patterns                | TextArray   | The patterns to be applied to the text.                                       |
| Pattern Group Name      | CognigyText | The name of the pattern matching group.                                       |
| Create New Slots        | Boolean     | Indicate whether new Slots are created from tags or not.                      |
| Tag Existing Slots      | Boolean     | Indicate whether existing Slots are tagged or not.                            |
| Detailed Compound Slots | Boolean     | Indicate whether detailed results for the matched patterns are stored or not. |
| Alternate Input         | CognigyText | The input text to use instead of the current input text.                      |

## Example

Text: `I want to fly from bErlin to NEW York with 3 or 4 people on October 1st 2021 if its 30 degrees and I am 22 years old 100%`

Pattern: `from @city>origin to @city>destination with @NUMBER>ppl1 or @NUMBER>ppl2 people @DATE>date if its @TEMPERATURE>celsius degrees and I am @AGE>userage years old @PERCENTAGE>perc`

Result:

```JSON theme={null}
"matchedPatterns": {
        "group": [
            {
                "matchedPhrase": "from bErlin to NEW York with 3 or 4 people on October 1st 2021 if its 30 degrees and I am 22 years old 100%",
                "origin": "Berlin",
                "destination": "New York",
                "ppl1": 3,
                "ppl2": 4,
                "date": {
                    "day": 1,
                    "month": 10,
                    "year": 2021,
                    "hour": 12,
                    "minute": 0,
                    "second": 0,
                    "millisecond": 0,
                    "weekday": 1,
                    "dayOfWeek": "Monday",
                    "ISODate": "2020-06-15T12:00:00"
                },
                "celsius": 30,
                "userage": 22,
                "perc": 100
            }
        ]
    }
```


# AI
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/ai/overview



AI Nodes extend the toolset provided by the Basic Nodes to a more in-depth level, enabling interactions with system resources.

[Watch this Episode](https://support.cognigy.com/hc/en-us/articles/360019857220-Cognigy-Sessions-Cognigy-NLU) of Cognigy Sessions for a technical deep dive.

<CardGroup>
  <Card title="Add Lexicon Keyphrase" href="/ai/agents/develop/node-reference/ai/add-lexicon-keyphrase">
    The Add Lexicon Keyphrase Node dynamically creates a [Lexicons](/ai/platform-features/nlu/slots/user-defined/lexicon) Keyphrase entry, either based on hardcoded or dynamic data.
  </Card>

  <Card title="AI Agent" href="/ai/agents/develop/node-reference/ai/ai-agent">
    The AI Agent Node lets you assign a job to an AI Agent, provide them instructions, tools, and access to knowledge.
  </Card>

  <Card title="AI Agent Handover" href="/ai/agents/develop/node-reference/ai/ai-agent-handover">
    The AI Agent Handover Node transfers the conversation to another AI Agent Node or Flow.
  </Card>

  <Card title="Clean Text" href="/ai/agents/develop/node-reference/ai/clean-text">
    The Clean Text Node enables a user to clean a provided text using the Text Cleaner class.
  </Card>

  <Card title="Copy Slots to Context" href="/ai/agents/develop/node-reference/ai/copy-slots-to-context">
    The Copy Slots to Context Node copies Slots that were detected and written to the [Input](/ai/agents/develop/ai-agent-memory/input) or [Context](/ai/agents/develop/ai-agent-memory/context) object.
  </Card>

  <Card title="Disable Slot Fillers" href="/ai/agents/develop/node-reference/ai/disable-slot-fillers">
    The Disable Slot Fillers Node disables any active [Slot Fillers](/ai/platform-features/nlu/overview).
  </Card>

  <Card title="Enable Slot Fillers" href="/ai/agents/develop/node-reference/ai/enable-slot-fillers">
    The Enable Slot Fillers Node enables a particular [Slot Filler](/ai/platform-features/nlu/overview).
  </Card>

  <Card title="Execute Cognigy NLU" href="/ai/agents/develop/node-reference/ai/execute-cognigy-nlu">
    The Execute Cognigy NLU Node explicitly executes [Cognigy NLU](/ai/platform-features/nlu/overview).
  </Card>

  <Card title="Fuzzy Search" href="/ai/agents/develop/node-reference/ai/fuzzy-search">
    The Fuzzy Search Node enables a Flow to search fuzzy string patterns in list of source data and returns the best possible matches.
  </Card>

  <Card title="Load AI Agent" href="/ai/agents/develop/node-reference/ai/load-ai-agent">
    The Load AI Agent Node lets you load the AI Agent's configuration and persona data into the Input or Context object.
  </Card>

  <Card title="Match Pattern" href="/ai/agents/develop/node-reference/ai/match-pattern">
    The Match Pattern Node enables a user to find patterns in text and expose them in the [Input Object](/ai/agents/develop/ai-agent-memory/input).
  </Card>

  <Card title="Regex Slot Filler" href="/ai/agents/develop/node-reference/ai/regex-slot-filler">
    The Regex Slot Filler Node creates a Slot based on a regex pattern. When the pattern is detected, a Slot is written to the [Input](/ai/agents/develop/ai-agent-memory/input) object.
  </Card>

  <Card title="Resolve Tool Action" href="/ai/agents/develop/node-reference/ai/resolve-tool-action">
    The Resolve Tool Action Node sends an answer as the response to a tool call to the AI Agent after completing a [tool action](/ai/agents/develop/node-reference/ai/ai-agent#tool).
  </Card>
</CardGroup>


# Regex Slot Filler
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/ai/regex-slot-filler



<Frame>
  <img alt="Regex Slot Filler Node configuration panel" />
</Frame>

## Description

Use this Node to fill Slots using regular expressions. To learn more about Slots and Tags, see [Lexicons & Slots](/ai/platform-features/nlu/slots/user-defined/lexicon).

## Parameters

| Parameter | Type                                                | Description                                                                                                                                                                                                                                                                                                  |
| --------- | --------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| Regex     | Regular expression                                  | The regular expression applied to the user input. The regular expression must start and end with /, for example, `/^1\d{7}$/`. Compared to [regular expressions](/ai/platform-features/nlu/slots/user-defined/regex) in **Slot Fillers** on the **NLU** tab, the Regex Slot Filler Node also supports flags. |
| Flags     | Regular expression flag characters in single string | Regular expression flags you want to set. When the `g` flag is used, for example, `/^1\d{7}$/g`, and the input contains multiple matches, all matched values are stored as an array. The array is stored under the `slots` key in the Input object.                                                          |
| Tag       | CognigyScript                                       | The Tag or Slot you want to fill.                                                                                                                                                                                                                                                                            |


# Resolve Tool Action
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/ai/resolve-tool-action



<a href="/release-notes/2026.4"><Badge>Updated in 2026.4</Badge></a>

<Frame>
  <img alt="Resolve Tool Action Node configuration panel" />
</Frame>

## Description

The Resolve Tool Action Node sends an answer as the response to a tool call to the AI Agent after completing a [tool](/ai/agents/develop/node-reference/ai/ai-agent).

When the AI Agent calls a tool to perform an action, the tool is executed. Once the action is completed, the Resolve Tool Action Node sends a response back to the AI Agent, indicating the result of that action. This approach helps the AI Agent process and make decisions based on the results of the tools' actions.

### Restrictions

* When using an [Execute Flow Node](/ai/agents/develop/node-reference/logic/execute-flow) with a Resolve Tool Action Node, consider the following: The Resolve Tool Action Node returns the interaction to the tool's AI Agent or LLM Prompt Node. That Node doesn't preserve the execution order determined by the Execute Flow Node. If the Resolve Tool Action Node is executed in a Flow after an Execute Flow Node, use a [Go To Node](/ai/agents/develop/node-reference/logic/go-to) after the Default child Node of that AI Agent or LLM Prompt Node to switch back to the original Flow.

## Parameters

| Parameter   | Type          | Description                                                                                                                                    |
| ----------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------- |
| Answer Text | CognigyScript | Specify the text that is sent to the AI Agent in response to the tool call to inform about the success, failure, or other details of the tool. |

### Debug Settings

| Parameter     | Type   | Description                                                                     |
| ------------- | ------ | ------------------------------------------------------------------------------- |
| Debug Message | Toggle | Enable sending a debug message that displays the tool name and the answer text. |

### Advanced

| Parameter     | Type   | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
| ------------- | ------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Maximum Loops | Slider | The maximum number of times an AI Agent can loop back to continue processing after completing a tool action. The parameter allows the AI Agent to perform multi-step reasoning or use multiple tools in sequence while preventing infinite loops that could waste resources or slow down responses. <br /> Each Resolve Tool Action Node tracks its own loop count. When the maximum is reached, the AI Agent stops and displays an `Infinite Loop Detected` error. The default value is `4`, and the maximum value is `50`. Adjust the value based on the use case complexity:<ul><li>**1–4** – for tasks where the AI Agent only needs to call one tool or complete a single step. The default is usually sufficient.</li><li>**5–15** – for tasks that require the AI Agent to call multiple tools in sequence or perform several reasoning steps to reach a result.</li><li>**15–50** – for tasks where the AI Agent must plan and execute long sequences of actions, make decisions, or handle highly complex reasoning. Higher loop numbers let the AI Agent do more work but make it slower and use more resources, so only raise it when necessary.</li></ul> |

## More Information

* [AI Agent](/ai/agents/develop/node-reference/ai/ai-agent)
* [AI Agent Handover](/ai/agents/develop/node-reference/ai/ai-agent-handover)
* [Agentic AI](/ai/agents/overview)


# Activate Profile
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/analytics/activate-profile



<Frame>
  <img alt="Activate Profile Node configuration panel" />
</Frame>

## Description

This Node is used to reactivate a [specific Contact Profile](/ai/agents/analyze/contact-profiles#working-with-contact-profiles) that has been deactivated, in case the user disabled data collection by mistake or wants to allow data collection again.

## Use Case

Consider placing this Node below an [If](/ai/agents/develop/node-reference/logic/if) Node to trigger it when the contact allows collecting their data.

## More Information

* [Update Profile](/ai/agents/develop/node-reference/analytics/update-profile)
* [Deactivate Profile](/ai/agents/develop/node-reference/analytics/deactivate-profile)
* [Delete Profile](/ai/agents/develop/node-reference/analytics/delete-profile)
* [Merge Profile](/ai/agents/develop/node-reference/analytics/merge-profile)


# Add Memory
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/analytics/add-memory



<Frame>
  <img alt="Add Memory Node configuration panel" />
</Frame>

## Description

<Note>
  This feature is in beta. We encourage you to try it out and provide us with feedback.
</Note>

The Add Memory Node stores memories within the [Contact Profile](/ai/agents/analyze/contact-profiles).
A memory entry can be any text, such as user input and other relevant details.

Recorded memories are also accessible within the [Profile](/ai/agents/develop/ai-agent-memory/profile) object in the Interaction Panel.
The array of memories consists of multiple memory objects.
Each memory object contains:

* `id` — a unique identifier for the memory.
* `timestamp` — the exact time the memory was created, recorded in ISO 8601 format.
* `text` — the content of the memory.

Each time the Add Memory Node is triggered, it adds a new memory object to the array, updating the AI Agent's long-term memory.

By referencing stored memories, the AI Agent can generate responses based on past interactions, offering a more personalized and context-aware experience for the end user.

## Parameters

| Parameter  | Type          | Description                                                                    |
| ---------- | ------------- | ------------------------------------------------------------------------------ |
| New Memory | CognigyScript | The text that will be stored in the **Memories** field of the Contact Profile. |

## More Information

* [Agent AI](/ai/agents/develop/node-reference/ai/ai-agent)
* [Agentic AI](/ai/agents/overview)
* [Contact Profile](/ai/agents/analyze/contact-profiles)


# Blind Mode
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/analytics/blind-mode



<Frame>
  <img alt="Blind Mode Node configuration panel" />
</Frame>

## Description

<Warning>
  As of Cognigy 2026.18.0, the Blind Mode Node is deprecated. The removal is planned for March 2027. Use the [Data Redaction](/ai/platform-features/data-redaction) feature instead.
</Warning>

Cognigy logs conversations to operate, maintain, and improve the service, but user inputs can sometimes contain sensitive information, such as names, addresses, phone numbers, or other private details. To protect this information, you can mask or exclude it from logs or analytics by using a Blind Mode Node.
The Node lets you control when sensitive data is masked or excluded, ensuring it isn't exposed.

## How to Use the Blind Mode Node

In the Flow editor, add two Blind Mode Nodes:

1. Add a Blind Mode Node before the Node where the sensitive input is expected. Rename the Blind Mode Node to `Blind Mode On`.
2. Add a Blind Mode Node after the Node with the sensitive input. Rename the Blind Mode Node to `Blind Mode Off`.

<Frame>
  <img alt="Flow editor showing Blind Mode On and Off Nodes around sensitive input" />
</Frame>

Next, choose whether you want to mask or exclude sensitive data:

<AccordionGroup>
  <Accordion title="Mask Data in Analytics">
    To mask sensitive data in the Conversations explorer:

    1. In the **Blind Mode On** Node editor, activate the **Mask Analytics** option.
    2. Go to the Conversations explorer to explore the conversation transcript:
       * For inputs – the message right before the **Blind Mode On** Node will be masked (replaced with `****`). The message right before the **Blind Mode Off** Node will remain unmasked.
       * For outputs – the messages between the **Blind Mode On** and **Blind Mode Off** Nodes are masked.
  </Accordion>

  <Accordion title="Mask Data in Logs">
    To mask sensitive data on the [Logs](/ai/agents/test/logs) page:

    1. In the **Blind Mode On** Node editor, activate the **Mask Logging** option.

    On the **Logs** page, the following data between the **Blind Mode On** and **Blind Mode Off** Nodes will be masked (replaced with `****`): `userId`, `input`, and `output`.
  </Accordion>

  <Accordion title="Exclude Data from Intent Trainer Records">
    To prevent storing user inputs in the Intent Trainer:

    1. In the **Blind Mode On** Node editor, activate the **Disable Intent Trainer Records** option.

    Any user input between the **Blind Mode On** and **Blind Mode Off** Nodes won't be stored in the Intent Trainer.
  </Accordion>

  <Accordion title="Exclude Parts of a Conversation from Logs">
    To stop recording part of conversation:

    1. In the Blind Mode On Node editor, activate the **Disable Conversations** option.
    2. Go to the **Logs** page to view the logs:
       * For inputs – recording will stop one message before the **Blind Mode On** Node and resume one message before the **Blind Mode Off** Node.
       * For outputs – only the messages between the **Blind Mode On** and **Blind Mode Off** Nodes will be excluded from logging.

    <Note>
      When both **Mask Analytics** and **Disable Conversations** options are activated, the **Disable Conversations** option takes precedence.
    </Note>
  </Accordion>
</AccordionGroup>

## Best Practices for Using Blind Mode Nodes

By applying these steps, you can ensure that sensitive user information is protected and that your AI Agent adheres to privacy standards:

* Always add a second Blind Mode Node with the options turned off to signal when data masking is no longer needed and normal data handling can resume.
* For entire Flow masking, enable data masking at the [Endpoint](/ai/agents/deploy/endpoints/data-protection-and-analytics) level.


# Complete Task
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/analytics/complete-task



<a href="/release-notes/4.90"><Badge>Updated in 4.90</Badge></a>

<Frame>
  <img alt="Complete Task Node configuration panel" />
</Frame>

## Description

The Complete Task Node provides an easy way
to track the completion of a specific event once the Node is triggered
and can be used only with [Tasks](/ai/agents/analyze/goals-and-tasks/tasks).

To use this Node, add it below the Node that completes the action you want to track.
Once the user has passed this action and triggered the Complete Task Node,
the completed task is stored in the [Contact Profile](/ai/agents/analyze/contact-profiles) and [Analytics database](/ai/agents/analyze/collecting-data).

Complete Tasks have a variety of potential use cases and help understand how often certain events are achieved.
The completed tasks are recorded in the Contact Profiles of each end user.
The data is also available in Insights (on reports and explorers) and through the OData endpoint.
The [top tasks](/insights/reports/overview#top-tasks) are listed in the Overview report
and allow you to see in how many sessions each task was completed.
If you don't intend to analyze Contact Profiles and prefer a more focused analysis using the [OData Endpoint's Steps collection](/ai/agents/analyze/odata#steps), then [Analytics Steps](/ai/agents/analyze/analytics-steps) are more suitable.

The difference between a Complete Task and an Analytics Step lies in their purpose and usage within a system. Complete Tasks determine whether a user has completed a specific objective or task. In contrast, Analytics Steps collect data about individual events or actions for analytical purposes. Complete Tasks are associated with a user's profile, whereas Analytics Steps focus on tracking and storing data information.

## Parameters

| Parameter | Type          | Description                                                               |
| --------- | ------------- | ------------------------------------------------------------------------- |
| Task      | CognigyScript | The name of the task that should be completed when the Node is triggered. |

## Use Cases

<AccordionGroup>
  <Accordion title="Order Placement">
    When a user completes the order placement process successfully,
    you can use the Complete Task Node to mark the "Order Placement" task as completed.

    This task helps you track the number of successful order placements by customers through the chat or voice call.
  </Accordion>

  <Accordion title="Account Registration">
    When a user has to create an account, you can use the Complete Task Node to indicate that the "Account Registration" task has been achieved.

    This task helps you to track the number of successful account registrations through the chat or voice call.
  </Accordion>

  <Accordion title="Appointment Booking">
    When a user books an appointment, you can use the Complete Task Node to signify the completion of the appointment booking process.

    This task helps you track the number of successful appointments scheduled through the chat or voice call.
  </Accordion>

  <Accordion title="Information Retrieval">
    When a user seeks specific information from the chatbot, such as product details or FAQ answers, you can use the Complete Task Node to mark the "Information Retrieval" task as completed.

    This task allows you to track successful information retrieval interactions through the chat or voice call.
  </Accordion>
</AccordionGroup>

## Example

The example shows how to measure the AI Agent's effectiveness in a customer call center.

<Accordion title="Measuring AI Agent Effectiveness">
  1. In the Flow, add a Question Node with the `Have we resolved your problem?` question and provide buttons as options:
     * B1: Yes, my problem is resolved
     * B2: No, I want to speak to a human agent
  2. If the user selects B1, connect it to a Complete Task Node. This Node marks the `Successful AI Agent Conversation` task as completed, indicating that the AI Agent has effectively assisted the customer.
  3. If the customer selects B2, handle this separately by triggering a Handover to Human Agent Node. In this case, you don't need to use the Complete Task Node.
  4. If the customer doesn't reply or types another question directly, consider it as a neutral conversation. You don't need to track it as a completed or unsuccessful task.
</Accordion>

## Alternatives

As an alternative to this Node, you can use the [Track Goal](/ai/agents/develop/node-reference/analytics/track-goal) Node.
In this case, you can define multiple steps within one [goal](/ai/agents/analyze/goals-and-tasks/goals) and track important metrics such as time and money.


# Deactivate Profile
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/analytics/deactivate-profile



<a href="/release-notes/2026.18"><Badge>Updated in 2026.18</Badge></a>

<Frame>
  <img alt="Deactivate Profile Node configuration panel" />
</Frame>

## Description

This Node deactivates a [specific Contact Profile](/ai/agents/analyze/contact-profiles#working-with-contact-profiles). By default, after the Contact Profile is deactivated, its data is kept and no new data is collected.

You can configure this Node to delete all stored Contact Profile data or mask its personal data with the *Mask and Keep Analytics* feature. With this feature, you can keep only analytics data without any personal traceability. This approach is useful for tracking your AI Agents' performance while complying with data protection standards, such as the General Data Protection Regulation (GDPR).

<Danger>
  If you delete the Contact Profile with this Node, all the Contact Profile data is deleted and is no longer available.
</Danger>

## Restrictions

* The Mask and Keep Analytics feature is available only for Cognigy 2026.14.0 and later versions.

## Parameters

| Parameter      | Type    | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         |
| -------------- | ------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Delete Data    | Boolean | Deactivates the Contact Profile and deletes stored data, including all analytics data and transcripts, from the system. This parameter is deactivated by default. The Contact Profile isn't deleted, and you can reactivate the Contact Profile using the Activate Profile Node or the Data collection inactive toggle.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
| Analytics data | List    | This parameter appears when you activate **Delete Data**. Sets what data is deleted:<ul><li>**Delete analytics records** — deletes the Contact Profile data, including analytics and personal data. This option is selected by default. After the data is deleted, you can't recover it.</li><li>**Keep analytics records, mask personal data** — applies mask and keep analytics. With this option, data from the Contact Profile is processed as follows:<ul><li>Clears only free-text fields, such as input text and transcripts.</li><li>Replaces contact identifiers, such as contact and user IDs, with hashes.</li><li> Keeps analytics data, such as session counts and Intents.</li></ul>Masking is processed asynchronously and may take up to 2 hours to take effect. Masked sessions are excluded from the [Conversations](/insights/explorers/conversations) explorer. However, analytics data related to the masked session is still included in [Insights reports](/insights/overview#reports) and can be accessed through the [OData endpoint](/ai/agents/analyze/odata#data-protection).</li></ul> |

### Scope of Masking

When the **Mask and keep analytics data** parameter is activated, the following fields are affected:

| Collection                    | Hashed fields | Cleared fields                                                                                                                                                 |
| ----------------------------- | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Analytics                     | `contactId`   | `ip`, `inputText`, `userLanguageText`, `ratingComment`, `foundSlots`, `foundSlotDetails`, `custom1`–`custom10`, `inputData`, `slots`, `geoInfo`, `slotMatches` |
| Sessions                      | `userId`      | —                                                                                                                                                              |
| Conversations                 | `contactId`   | `inputText`, `previousInputText`, `ratingComment`, `inputData`, `previousInputData`, `inputAttachments`, `previousInputAttachments`                            |
| Goals, Live Agent Escalations | `contactId`   | —                                                                                                                                                              |
| Steps                         | `userId`      | —                                                                                                                                                              |
| Voice Gateway Events          | `userId`      | `recognizerTranscript`, `recognizerDigits`                                                                                                                     |

## Use Case

Consider placing this Node below an [If Node](/ai/agents/develop/node-reference/logic/if) to trigger it when the contact explicitly requests not to collect their data.

## More Information

* [Activate Profile](/ai/agents/develop/node-reference/analytics/activate-profile)
* [Update Profile](/ai/agents/develop/node-reference/analytics/update-profile)
* [Delete Profile](/ai/agents/develop/node-reference/analytics/delete-profile)
* [Merge Profile](/ai/agents/develop/node-reference/analytics/merge-profile)


# Delete Profile
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/analytics/delete-profile



<a href="/release-notes/2026.14"><Badge>Updated in 2026.14</Badge></a>

<Frame>
  <img alt="Delete Profile Node configuration panel" />
</Frame>

## Description

This Node is used to delete a [specific Contact Profile](/ai/agents/analyze/contact-profiles#working-with-contact-profiles).

You configure this Node to keep analytics records in masked form with the *Mask and Keep analytics* feature instead of deleting all related analytics data. This setting is useful for keeping analytics on your AI Agent interactions while avoiding any personal traceability. Use this Node to comply with data protection standards, such as the General Data Protection Regulation (GDPR), when users request to delete their data.

<Danger>
  If you delete the Contact Profile with this Node, all the Contact Profile data is deleted and is no longer available.
</Danger>

## Restrictions

* Mask and keep analytics is available only for Cognigy 2026.14.0 and later versions.

## Parameters

| Parameter                    | Type    | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
| ---------------------------- | ------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Mask and keep analytics data | Boolean | Applies mask and keep analytics. With this option, data from the Contact Profile is processed as follows:<ul><li>Replaces contact identifiers, such as contact and user IDs, with hashes.</li><li>Keeps analytics data, such as session counts and Intents, instead of deleting them.</li><li>Clears free-text fields, such as input text and transcripts.</li></ul>This parameter is deactivated by default. Masking is processed asynchronously and may take up to 2 hours to take effect. Masked sessions are excluded from the [Conversations](/insights/explorers/conversations) explorer. However, analytics data related to the masked session is still included in [Insights reports](/insights/overview#reports) and can be accessed through the [OData endpoint](/ai/agents/analyze/odata#data-protection). |

### Scope of Masking

When the **Mask and keep analytics data** parameter is activated, the following fields are affected:

| Collection                    | Hashed fields | Cleared fields                                                                                                                                                 |
| ----------------------------- | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Analytics                     | `contactId`   | `ip`, `inputText`, `userLanguageText`, `ratingComment`, `foundSlots`, `foundSlotDetails`, `custom1`–`custom10`, `inputData`, `slots`, `geoInfo`, `slotMatches` |
| Sessions                      | `userId`      | —                                                                                                                                                              |
| Conversations                 | `contactId`   | `inputText`, `previousInputText`, `ratingComment`, `inputData`, `previousInputData`, `inputAttachments`, `previousInputAttachments`                            |
| Goals, Live Agent Escalations | `contactId`   | —                                                                                                                                                              |
| Steps                         | `userId`      | —                                                                                                                                                              |
| Voice Gateway Events          | `userId`      | `recognizerTranscript`, `recognizerDigits`                                                                                                                     |

## Use Case

Consider placing this Node below an [If Node](/ai/agents/develop/node-reference/logic/if) to trigger it when the contact explicitly requests not to collect their data.

## More Information

* [Activate Profile](/ai/agents/develop/node-reference/analytics/activate-profile)
* [Update Profile](/ai/agents/develop/node-reference/analytics/update-profile)
* [Deactivate Profile](/ai/agents/develop/node-reference/analytics/deactivate-profile)
* [Merge Profile](/ai/agents/develop/node-reference/analytics/merge-profile)


# Merge Profile
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/analytics/merge-profile



<Frame>
  <img alt="Merge Profile Node configuration panel" />
</Frame>

## Description

The Merge Profile Node merges the Contact Profile of the current interaction with another Contact Profile. This is useful for omnichannel setups, where an end user can interact with your AI Agent through different platforms, but they need to be recognized as the same end user across them. After the merging, the Contact Profile merged into the current Contact Profile is deactivated.

## Parameters

| Parameter | Type          | Description                                                                                                                 |
| --------- | ------------- | --------------------------------------------------------------------------------------------------------------------------- |
| ContactId | CognigyScript | The ID of the Contact Profile you want to merge with the current Contact Profile of the user interacting with the AI Agent. |

## Example

A user is logged in on your website and using the Webchat widget. They have a Contact Profile where the user ID corresponds to their ID on your website. You can link their website ID, for example, to their account on Amazon. With this approach, they have the same Contact Profile when interacting with your Flow through an Alexa device. To link their data to the same Contact Profile, you only need to obtain their Amazon user ID through account linking and provide that ID to the Merge Profile Node. After this, the contact has one profile activated when they engage with your Flow through both the Webchat widget and their Alexa device.


# Analytics
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/analytics/overview



Analytics Nodes let you take advantage of Cognigy's comprehensive marketing functionality. Use and update [Contact Profiles](/ai/agents/analyze/contact-profiles) within your Flow for personalized, adaptive Flows that connect with your audience.

<CardGroup>
  <Card title="Activate Profile" href="/ai/agents/develop/node-reference/analytics/activate-profile">
    The Activate Profile Node activates the [Contact Profile](/ai/agents/analyze/contact-profiles).
  </Card>

  <Card title="Add Memory" href="/ai/agents/develop/node-reference/analytics/add-memory">
    The Add Memory Node stores any text, user input or other details in the Contact Profile.
  </Card>

  <Card title="Blind Mode (deprecated)" href="/ai/agents/develop/node-reference/analytics/blind-mode">
    The Blind Mode disables logging completely or selectively, for example, to exclude sensitive data from AI Agents' logs.
  </Card>

  <Card title="Complete Task" href="/ai/agents/develop/node-reference/analytics/complete-task">
    The Complete Task Node tracks the completion specifically of a [Task](/ai/agents/analyze/goals-and-tasks/tasks) once the Node is triggered.
  </Card>

  <Card title="Deactivate Profile" href="/ai/agents/develop/node-reference/analytics/deactivate-profile">
    The Deactivate Profile Node deactivates the [Contact Profile](/ai/agents/analyze/contact-profiles).
  </Card>

  <Card title="Delete Profile" href="/ai/agents/develop/node-reference/analytics/delete-profile">
    The Delete Profile Node deletes the currently active [Contact Profile](/ai/agents/analyze/contact-profiles).
  </Card>

  <Card title="Merge Profile" href="/ai/agents/develop/node-reference/analytics/merge-profile">
    The Merge Profile Node merges the current session to a [Contact Profile](/ai/agents/analyze/contact-profiles). You can configure the Node with an identifier that you can use as a user ID.
  </Card>

  <Card title="Overwrite Analytics" href="/ai/agents/develop/node-reference/analytics/overwrite-analytics">
    With the Overwrite Analytics Node, you can selectively overwrite the Analytics data that before storing it in [Insights](/insights/overview).
  </Card>

  <Card title="Request Rating" href="/ai/agents/develop/node-reference/analytics/request-rating">
    The Request Rating Node requests a rating from the user.
  </Card>

  <Card title="Track Goal" href="/ai/agents/develop/node-reference/analytics/track-goal">
    The Track Goal Node lets you monitor the progress specifically of [Goals](/ai/agents/analyze/goals-and-tasks/goals) within a Flow.
  </Card>

  <Card title="Set Rating" href="/ai/agents/develop/node-reference/analytics/set-rating">
    The Set Rating Node extends the analytical capabilities of Cognigy together with the [Insights](/insights/overview) application.
  </Card>

  <Card title="Update Profile" href="/ai/agents/develop/node-reference/analytics/update-profile">
    The Update Profile Node updates [Contact Profile](/ai/agents/analyze/contact-profiles) information.
  </Card>
</CardGroup>


# Overwrite Analytics
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/analytics/overwrite-analytics



<Frame>
  <img alt="Overwrite Analytics Node configuration panel" />
</Frame>

## Description

<Warning>
  As of Cognigy 4.91.0, the Understood Messages concept is deprecated. Use the [Goals and Tasks](https://docs.cognigy.com/ai/analyze/goals-and-tasks/overview/) concepts to evaluate whether your AI Agent understood a particular phrase. In Cognigy 2025.19, the charts related to this concept were removed from Insights reports.

  As of Cognigy 4.91.0, the `understood` field in the Cognigy OData endpoint is deprecated. The `understood` field has been removed from the Cognigy OData endpoint. Use the [Goals collection](https://docs.cognigy.com/ai/analyze/odata#goals) in the Cognigy OData endpoint.

  Refer to the list of all upcoming removals on the [Deprecations and Removals](https://docs.cognigy.com/release-notes/deprecations-and-removals/) page.
</Warning>

This Node allows you to selectively overwrite default values, such as Intents, text, understood messages, and more,
as well as manage the analytics data stored in the [Analytics](/ai/agents/analyze/collecting-data) database.
You can view this data on the [Insights reports](/insights/overview) or retrieve it via the [Cognigy OData Endpoint](/ai/agents/analyze/odata).
You can also use this Node to add custom values.

Empty fields don't overwrite the initial values.
If you add this Node and leave some fields empty, such as the **Intent** default field,
the initial value stored in the Analytics database will remain unchanged.

## Parameters

<AccordionGroup>
  <Accordion title="Custom Fields">
    Custom fields allow you to store any data as a string value.

    Custom fields can be useful for enriching the information about a user's conversation by adding more details.
    For example, if user inputs contain words such as `excellent` or `great`, you can add a custom field with the value `Positive feedback` to make it easier to analyze positive feedback through analytics records.

    | Parameter       | Type          | Description                                                                            |
    | --------------- | ------------- | -------------------------------------------------------------------------------------- |
    | Custom Value 1  | CognigyScript | Store a custom field that can be set to any string value. The default value is `null`. |
    | Custom Value 2  | CognigyScript | Store a custom field that can be set to any string value. The default value is `null`. |
    | Custom Value 3  | CognigyScript | Store a custom field that can be set to any string value. The default value is `null`. |
    | Custom Value 4  | CognigyScript | Store a custom field that can be set to any string value. The default value is `null`. |
    | Custom Value 5  | CognigyScript | Store a custom field that can be set to any string value. The default value is `null`. |
    | Custom Value 6  | CognigyScript | Store a custom field that can be set to any string value. The default value is `null`. |
    | Custom Value 7  | CognigyScript | Store a custom field that can be set to any string value. The default value is `null`. |
    | Custom Value 8  | CognigyScript | Store a custom field that can be set to any string value. The default value is `null`. |
    | Custom Value 9  | CognigyScript | Store a custom field that can be set to any string value. The default value is `null`. |
    | Custom Value 10 | CognigyScript | Store a custom field that can be set to any string value. The default value is `null`. |
  </Accordion>

  <Accordion title="Default Fields">
    Default fields allow you to change existing data or add new data to the default parameters,
    such as Intents, States, and inputs.
    If a default field is left empty, the previous value will be used in the analytics record.

    The table includes the most commonly used parameters for overwriting.
    For example, you can overwrite the Intent name.
    Consider the case where a user says, `I need to return an item`.
    The AI Agent recognizes this message as the `Return request` Intent.
    For more precise tracking,
    you can configure the AI Agent to overwrite this Intent so that it is stored in the Analytics database as `Product return`.

    | Parameter            | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                              |
    | -------------------- | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Intent               | CognigyScript | Store the intent name.                                                                                                                                                                                                                                                                                                                                                                                                                                   |
    | IntentScore          | Number        | Store the intent score.                                                                                                                                                                                                                                                                                                                                                                                                                                  |
    | inputText            | CognigyScript | Store the input text.                                                                                                                                                                                                                                                                                                                                                                                                                                    |
    | Input Data           | JSON          | Store additional input data.                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | State                | CognigyScript | Store the State information.                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | Slots                | JSON          | Store the Slot information.                                                                                                                                                                                                                                                                                                                                                                                                                              |
    | Completed Goals      | Text          | Store a list of completed goals.                                                                                                                                                                                                                                                                                                                                                                                                                         |
    | Understood           | Dropdown      | Select one of the following options:<ul><li>**Don't Overwrite** — the current status of the message (whether it is `understood` or `misunderstood`) will not be overridden.</li><li>**True** — the message will be counted as `understood`.</li><li>**False** — the message will not be counted as `misunderstood`.</li><li>**Don't Count** — the message will not be included in the calculation of `understood` or `misunderstood` messages.</li></ul> |
    | Handover Escalations | CognigyScript | Store the number of handover escalations.                                                                                                                                                                                                                                                                                                                                                                                                                |
  </Accordion>
</AccordionGroup>

## Examples

<AccordionGroup>
  <Accordion title="Anonymize Personal Data for Privacy in OData">
    You can anonymize specific text in the field to ensure privacy while maintaining context in OData.
    For example, instead of `My name is Alex, how do I do something?` you can replace it with `My name is <Name>, how do I do something?`.
  </Accordion>

  <Accordion title="Clean Up NLU Metrics with Understood">
    To clean up NLU metrics, you can set the `Understood` field to `null` or `true` in specific circumstances.
    This approach is useful in voice AI Agents for handling Voice Gateway payloads, which are automatically set to `false`.
  </Accordion>
</AccordionGroup>

## Alternatives

Alternatively, you can update analytics records using the [PATCH /v2.0/analytics](https://api-trial.cognigy.ai/openapi#patch-/v2.0/analytics) request.


# Request Rating
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/analytics/request-rating



<Frame>
  <img alt="Request Rating Node" />
</Frame>

## Description

This Node prompts end users to give feedback or rate their conversation experience using a rating widget in a chat.
The Node is typically placed at the end of a Flow to collect final feedback from conversations with either AI or human agents.

In contrast to the [Set Rating Node](/ai/agents/develop/node-reference/analytics/set-rating), the Request Rating Node doesn't rate a conversation automatically.

## Restrictions

* The Webchat v2 Endpoint doesn't support all parameters in this Node. You can still use it to [implement rating for Webchat v2](#implement-rating-for-webchat-v2).

## Parameters

| Parameter                            | Description                                                                                                                                                                                                                                                                    |
| ------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| Rating Screen Title                  | Sets the text for the screen title of the rating widget. The default text is `Conversation Rating`.                                                                                                                                                                            |
| Rating Title Text                    | Sets the text for the title of the rating widget. The default text is `Please rate your chat experience`. For Webchat v2, overrides the Rating Title Text parameter in the [Webchat v2 Endpoint settings](/webchat/v2/configuration#conversation-rating-settings).             |
| Rating Comment Placeholder           | Sets the placeholder text for the comment field in the rating widget. The default text is `Type something here...`. For Webchat v2, overrides the Rating Comment Text parameter in the [Webchat v2 Endpoint settings](/webchat/v2/configuration#conversation-rating-settings). |
| Rating Submit Button Text            | Sets the text that appears on the button users click to submit their chat ratings and feedback. The default text is `Send feedback`. Applicable only to Webchat v3.                                                                                                            |
| Rating Submitted Event Banner Text   | Sets the text that appears in a banner after a user successfully submits their rating and feedback. The default text is `Your feedback was submitted`. Applicable only to Webchat v3.                                                                                          |
| Rating Submitted Chat Status Message | Sets the status message that appears in the chat after a user successfully submits their rating and feedback. The default text is `Feedback submitted`. Applicable only to Webchat v3.                                                                                         |

### Implement Rating for Webchat v2

For Webchat v2, add this Node to your Flow and configure the rating widget directly in the [Conversation Rating Settings section](/webchat/v2/configuration#conversation-rating-settings) in the Webchat v2 Endpoint.

## Track Ratings in Insights

To track how the rating changed during a user session,
use [Insights](https://docs.cognigy.com/insights/overview/).
On the **Conversations Explorer** page,
examine the [transcript details](https://docs.cognigy.com/insights/explorers/conversations/#session-transcript-details)
to see the final rating for the session.
Also, explore the [Overview](https://docs.cognigy.com/insights/reports/overview/) and [Engagement](https://docs.cognigy.com/insights/reports/engagement/) reports to assess trends in positive ratings for your Agent.

## Use Cases

<Accordion title="Comparison of Rating Nodes">
  The table shows use cases for Set Rating and Request Rating Nodes. These cases are divided into steps.

  <colgroup>
    <col />

    <col />
  </colgroup>

  | Set Rating Node                                                                                                                                                                                 | Request Rating Node                                                                                    |
  | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------ |
  | 1. A customer interacts with the AI Agent to inquire about product specifications.                                                                                                              | 1. A customer completes a troubleshooting session with the AI Agent.                                   |
  | 2. The AI Agent successfully provides the required product information.                                                                                                                         | 2. The AI Agent prompts the customer to provide feedback on their experience.                          |
  | 3. The Set Rating Node assigns a predefined rating score (`1`) to the customer's session based on the positive interaction. The customer doesn't see any rating prompt during the conversation. | 3. The customer rates their experience with the AI Agent by using the Thumbs-up and Thumbs-down icons. |
  | 4. The chosen rating is recorded and stored for analysis.                                                                                                                                       | 4. The chosen rating is recorded and stored for analysis.                                              |
</Accordion>

## More information

* [Set Rating](/ai/agents/develop/node-reference/analytics/set-rating)
* [Insights](/insights/overview)


# Set Rating
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/analytics/set-rating



<Frame>
  <img alt="Set Rating Node configuration panel" />
</Frame>

## Description

This Node prefills a rating prompt with a specific value or updates it based on other actions or user inputs. Multiple Set Rating Nodes can be added to collect feedback at specific steps.

You can combine this Node with the [Request Rating Node](/ai/agents/develop/node-reference/analytics/request-rating).
For example,
add a Set Rating Node
to prefill a rating prompt with a default value based on the user's previous interactions with the AI Agent. Then,
include the Request Rating Node to enable the user to confirm or change the rating value.

In contrast to the [Request Rating Node](/ai/agents/develop/node-reference/analytics/request-rating), the Set Rating Node doesn't display a rating widget to the end user.

## Parameters

| Parameter      | Description                                                                                 |
| -------------- | ------------------------------------------------------------------------------------------- |
| Rating Score   | The rating for the conversation is `-1` for a negative rating or `1` for a positive rating. |
| Rating Comment | Additional information about the rating.                                                    |

## Track Ratings in Insights

To track how the rating changed during a user session,
use [Insights](https://docs.cognigy.com/insights/overview/).
On the **Conversations Explorer** page,
examine the [transcript details](https://docs.cognigy.com/insights/explorers/conversations/#session-transcript-details)
to see the final rating for the session.
Also, explore the [Overview](https://docs.cognigy.com/insights/reports/overview/) and [Engagement](https://docs.cognigy.com/insights/reports/engagement/) reports to assess trends in positive ratings for your Agent.

## Use Cases

<Accordion title="Comparison of Rating Nodes">
  The table shows use cases for Set Rating and Request Rating Nodes. These cases are divided into steps.

  <colgroup>
    <col />

    <col />
  </colgroup>

  | Set Rating Node                                                                                                                                                                                 | Request Rating Node                                                                                    |
  | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------ |
  | 1. A customer interacts with the AI Agent to inquire about product specifications.                                                                                                              | 1. A customer completes a troubleshooting session with the AI Agent.                                   |
  | 2. The AI Agent successfully provides the required product information.                                                                                                                         | 2. The AI Agent prompts the customer to provide feedback on their experience.                          |
  | 3. The Set Rating Node assigns a predefined rating score (`1`) to the customer's session based on the positive interaction. The customer doesn't see any rating prompt during the conversation. | 3. The customer rates their experience with the AI Agent by using the Thumbs-up and Thumbs-down icons. |
  | 4. The chosen rating is recorded and stored for analysis.                                                                                                                                       | 4. The chosen rating is recorded and stored for analysis.                                              |
</Accordion>

## More information

* [Request Rating](/ai/agents/develop/node-reference/analytics/request-rating)
* [Insights](/insights/overview)


# Track Goal
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/analytics/track-goal



<Frame>
  <img alt="Track Goal Node configuration panel" />
</Frame>

## Description

The Track Goal Node lets you monitor the progress of a specific goal within a Flow
and can be used only with [Goals](/ai/agents/analyze/goals-and-tasks/goals).

By adding this Node, you can track key steps defined in your goals, such as time or money saved,
and then analyze the results in the Insights report or via the OData endpoint.

For example, if your goal is to reduce call handling time, the Track Goal Node will track each interaction step, helping you measure whether your Flow is achieving the desired reductions in response times.

## Parameters

| Parameter    | Description                                                                                                                                                                                                                                                                                                                                                        |
| ------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| Select Goal  | Select the goal you previously created in **Build > Goals**. This is the goal you want to track progress on, such as `Reduce Call Handling Time` or `Improve Customer Onboarding`.                                                                                                                                                                                 |
| Select Steps | Select one or more steps that you have configured within your chosen goal. For example, if your goal includes multiple stages, such as `Initial Contact` and `Query Resolved`, you can track each step's progress separately. <br /> Set the Initial step at the beginning of the process you want to track in your Flow. Otherwise, the goal will not be tracked. |

## Example

Let's assume you have set a goal called `Reduce Customer Support Request Time`.
The steps within this goal are the following:

1. Customer Request Received. This step represents when the customer initiates contact with support.
2. Request Resolved by Chatbot. This step tracks when the AI Agent successfully resolves a customer query.
3. Escalation to Human Agent. This step tracks when the AI Agent escalates the request to a human agent if needed.

## More Information

* [Goals](/ai/agents/analyze/goals-and-tasks/goals)
* [Goals Report](/insights/reports/goals)
* [Complete Task Node](/ai/agents/develop/node-reference/analytics/complete-task)


# Update Profile
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/analytics/update-profile



<Frame>
  <img alt="Update Profile Node configuration panel" />
</Frame>

## Description

This Node updates users' Contact Profiles.
The available Profile Schema fields correspond to all defined fields in the Project's [Profile Schema](/ai/agents/analyze/contact-profiles#profile-schema).
If an invalid field is used, no action is taken.

## Parameters

| Parameter            | Type          | Description                                                                                                                                                                                                                                                                                        |
| -------------------- | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Profile Schema Field | Selector      | Identify the field in the user's Contact Profile that needs to be updated. The predefined Profile Schema fields are described in [Contact Profiles](/ai/agents/analyze/contact-profiles); you can also create a [custom field](/ai/agents/analyze/contact-profiles##customize-the-profile-schema). |
| Value                | CognigyScript | The value of the field.                                                                                                                                                                                                                                                                            |

## More Information

* [Activate Profile](/ai/agents/develop/node-reference/analytics/activate-profile)
* [Deactivate Profile](/ai/agents/develop/node-reference/analytics/deactivate-profile)
* [Delete Profile](/ai/agents/develop/node-reference/analytics/delete-profile)
* [Merge Profile](/ai/agents/develop/node-reference/analytics/merge-profile)


# Code
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/basic/code



<a href="/release-notes/4.68"><Badge>Updated in 4.68</Badge></a>

<Frame>
  <img alt="Code Node configuration panel" />
</Frame>

## Description

The Code Node enables users to execute JavaScript, TypeScript, and [CognigyScript](/ai/platform-features/cognigyscript) in the Flow. This Node allows for deeper customization of AI Agents, for example, by calling API requests, using npm modules, and manipulating text and data.

For more information, refer to [Code Node reference](/ai/for-developers/code/overview).


# Debug Message
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/basic/debug-message



<a href="/release-notes/4.84"><Badge>Updated in 4.84</Badge></a>

<Frame>
  <img alt="Debug Message Node configuration panel" />
</Frame>

## Description

This Node sends a debug message to the Interaction Panel.
Debug messages help
you track the flow of data and diagnose issues
by providing real-time information about what's happening during the Flow execution.
To view the debug message, enable [debug mode](/ai/agents/test/interaction-panel/chat).

## Parameters

| Parameter   | Type          | Description                                                                                                                                                                                                                                                                                                                                                                     |
| ----------- | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Debug Level | Select        | The Debug message level. Select one of the following options:<ul><li>**Info** — info messages will be displayed with a blue background and contain general information. For example: `User successfully logged in`.</li><li>**Error** — error messages will be displayed with a red background to highlight issues. For example: `Failed to connect to the database`.</li></ul> |
| Header      | CognigyScript | The header text of the debug message. This parameter is optional.<br />Examples:<ul><li>`User Action`</li><li>`System Error`</li><li>`Database Connection`</li></ul>                                                                                                                                                                                                            |
| Message     | CognigyScript | The message text of the debug message. This parameter is mandatory.<br />Examples:<ul><li>`User input received: {{input.text}}`</li><li>`Invalid phone number entered: {{profile.email}}`</li></ul>                                                                                                                                                                             |

## Examples

<AccordionGroup>
  <Accordion title="Info Message">
    To show that the message relates to user's actions, such as completing a task or making a selection, you can use the Debug Message Node with the header `User Action`.
    It can be helpful when a user fills out an [Adaptive Card](/ai/agents/develop/node-reference/basic/say) form or clicks the [Submit](/ai/agents/develop/node-reference/xApp/set-html-xApp-state) button in xApps to confirm an action.

    ```txt theme={null}
    Header: "User Action"
    Message: "User successfully completed the registration process"
    ```
  </Accordion>

  <Accordion title="Error Message">
    To show that the message relates to an error related to database operations,
    use the Debug Message Node with the header `Database Error`.
    It can be useful for tracking errors that occur during database requests. For example, it can help identify issues when using [MongoDB](/ai/agents/develop/node-reference/service/mongo-db) Nodes.

    ```txt theme={null}
    Header: "Database Error"
    Message: "Failed to retrieve user data"
    ```
  </Accordion>
</AccordionGroup>

## Alternatives

To debug a Flow when your AI Agent experiences issues on the production side, add a [Log Message](/ai/agents/develop/node-reference/other-nodes/log-message) Node.


# Optional Question
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/basic/optional-question



<Frame>
  <img alt="Optional Question Node configuration panel" />
</Frame>

## Description

An Optional Question Node requests specific information from the user. The Node can be configured to expect different types of input. As soon as the selected type of data is detected in the user input, the Optional Question Node is answered.

<Accordion title="Comparison with the Question Node">
  The Optional Question Node is non-blocking. This means that the user can continue their conversation with the Flow if they haven't provided an answer to the Optional Question. Before regular Flow execution, the Cognigy checks if the user input answers the Optional Question to find a result. If not, the Flow execution follows a different path.
  [Watch this Episode](https://support.cognigy.com/hc/en-us/articles/360020609199-Cognigy-Sessions-Question-Nodes) of Cognigy Sessions for a technical deep dive.
</Accordion>

## Parameters

<AccordionGroup>
  <Accordion title="Basic Settings">
    | Type        | Required user input to answer question                                                                                                                                                                            |
    | ----------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Text        | Any text input.                                                                                                                                                                                                   |
    | Yes / No    | A positive or negative response.                                                                                                                                                                                  |
    | Intent      | One of the trained [Intents](/ai/platform-features/nlu/intents/overview) must be identified from the user's response.                                                                                             |
    | Slot        | A [System Slot](/ai/platform-features/nlu/slots/system-defined) or [Lexicon Slot](/ai/platform-features/nlu/slots/user-defined/lexicon) must be detected within the user's response. The slot is defined by name. |
    | Date        | Any date (system-defined).                                                                                                                                                                                        |
    | Number      | Any number (system-defined).                                                                                                                                                                                      |
    | Temperature | Any temperature (system-defined).                                                                                                                                                                                 |
    | Age         | Any age (system-defined).                                                                                                                                                                                         |
    | Duration    | Any time duration (system-defined).                                                                                                                                                                               |
    | Email       | Any email address (system-defined).                                                                                                                                                                               |
    | Money       | Any monetary value (system-defined).                                                                                                                                                                              |
    | URL         | Any reference/address to a resource on the Internet, for example, `http://example.com`.                                                                                                                           |
    | Percentage  | Any percentage (system-defined).                                                                                                                                                                                  |
    | Regex       | Any custom data format defined by a regex expression must be detected in the user's response.                                                                                                                     |
    | Data        | Any data (input.data) input.                                                                                                                                                                                      |

    <Note>
      All data formats supported by the Cognigy NLU for system slot mapping are listed on the [Slot Mapping](/ai/platform-features/nlu/slots/system-defined) page.
    </Note>
  </Accordion>

  <Accordion title="Channels and Output types">
    The Optional Question Node output types carry the same functionality as the [Say Node](/ai/agents/develop/node-reference/basic/say).

    <Check>
      If you select the Date as a Question Type, the Question Node automatically renders a datepicker if the channel supports it. Refer to [Datepicker](/ai/agents/develop/node-reference/data/datepicker) for more information.
    </Check>
  </Accordion>

  <Accordion title="Advance">
    | Parameter                        | Description                                                                                                                                                                                                                                                                                                                                                   |
    | -------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Store detailed Results           | When enabled, stores a more detailed JSON object under the result property of the input. This is useful in case more information is needed.                                                                                                                                                                                                                   |
    | Stop Flow execution after Answer | When enabled, Flow execution stops after the "On answer" case and wait for another input.                                                                                                                                                                                                                                                                     |
    | Forget Question Threshold        | The difference with a regular Question Node is that the Optional Question Node doesn't need to be answered. If the Optional Question Node is triggered and no answer is provided, the fact that the question was "asked" is nonetheless kept in memory. With this setting, you can decide for how many inputs the Optional Question should be kept in memory. |
    | Parse CognigyScript              | Lets you decide on which user input the CognigyScript after Answer should be executed.                                                                                                                                                                                                                                                                        |
  </Accordion>

  <Accordion title="Result Storage">
    If **Store Result in Context** is enabled, the Question Result is also stored in the [Context](/ai/agents/develop/ai-agent-memory/context) object.
  </Accordion>

  <Accordion title="Question Information in Input">
    When an Optional Question is "active", meaning the AI Agent waits for an answer, information about the Optional Question is added to the Input object.

    ```JSON theme={null}
    "activeOptionalQuestion": {
        "nodeId": "cb81db3a-6230-49e7-8ce4-12f877abe0cb",
        "type": "yesNo",
        "lastExecutedAt": 1,
        "forgetQuestionThreshold": 3
    }
    ```
  </Accordion>
</AccordionGroup>

## AI-Enhanced Output

To use AI-enhanced output rephrasing, read the [Generative AI](/ai/agents/develop/gen-ai-and-llms/generative-ai) article.


# Basic Nodes
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/basic/overview



Basic Nodes are the fundamental Nodes you will need to create a Flow.

<CardGroup>
  <Card title="Say" href="/ai/agents/develop/node-reference/basic/say">
    Use the Say Node to send a message to the user on their current channel.
  </Card>

  <Card title="Question" href="/ai/agents/develop/node-reference/basic/question">
    Use the Question Node to ask the user a question.
  </Card>

  <Card title="Optional Question" href="/ai/agents/develop/node-reference/basic/optional-question">
    The Optional Question Node works similar to the Question Node, except that the conversation can continue even if the question is not answered.
  </Card>

  <Card title="Debug Message" href="/ai/agents/develop/node-reference/basic/debug-message">
    The Debug Message Node sends a debug message to the Interaction Panel.
  </Card>

  <Card title="Code Node" href="/ai/agents/develop/node-reference/basic/code">
    The Code Node enables the execution of JavaScript, TypeScript, and CognigyScript in the Flow.
  </Card>
</CardGroup>


# Question
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/basic/question



<a href="/release-notes/2026.18"><Badge>Updated in 2026.18</Badge></a>

<Frame>
  <img alt="Question Node configuration panel" />
</Frame>

## Description

<Warning>
  As of Cognigy 2026.18.0, the **Temperature** parameter in the **LLM Entity Extraction Options** is deprecated. The removal is planned for April 2027. For models that don't support this parameter, activate the **Ignore deprecated parameters** toggle to exclude it from the request to the LLM provider. Otherwise, the request to the LLM provider might fail.

  For example, OpenAI models such as `gpt-5`, `gpt-5.1`, and `gpt-5.2` don't support these parameters. Similar models from other LLM providers might also not support these parameters. Check your LLM provider's documentation for model-specific support.
</Warning>

A Question Node is used to ask a question that requests specific information from the user. When the Node is triggered, the [Entrypoint](/ai/agents/develop/projects-and-flows/editor) shifts to this Node so that the conversation continues only after the user answers. Also, a new [Input](/ai/agents/develop/ai-agent-memory/input) object is generated.

When a user input is received, it's scored based on natural language understanding (NLU). If an attached Flow has an Intent that scores higher than Intents in the current Flow, the [attached Flow](/ai/platform-features/nlu/attachments) is executed. The Intent scoring occurs before validation of the Question Node is completed.

After the AI Agent asks a question and the user answers, the answer is validated according to its type. If it passes, the answer is valid and stored, and the conversation continues.

<Accordion title="Question Nodes and Intent Execution">
  Question Nodes, by default, are triggered repeatedly until a valid answer is provided. To avoid this behavior, you can use an Optional Question or change the Intent Execution setting.

  <iframe title="YouTube video player" />
</Accordion>

## Parameters

<AccordionGroup>
  <Accordion title="Question Types">
    Question Nodes have a selection of types that determine the validation used before a conversation continues.

    | Type                                 | Expected user input to answer question                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  | Example                                                                                                                                                                |
    | ------------------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Text                                 | Any text input.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         |                                                                                                                                                                        |
    | Yes / No                             | A positive or negative response.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |                                                                                                                                                                        |
    | Intent                               | One of the trained [Intents](/ai/platform-features/nlu/intents/overview) must be identified from the user's response.                                                                                                                                                                                                                                                                                                                                                                                                   |                                                                                                                                                                        |
    | Slot                                 | A [System Slot](/ai/platform-features/nlu/slots/system-defined) or [Lexicon Slot](/ai/platform-features/nlu/slots/user-defined/lexicon) must be detected within the user's response. The slot is defined by name.                                                                                                                                                                                                                                                                                                       |                                                                                                                                                                        |
    | Date                                 | Any date (system-defined).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |                                                                                                                                                                        |
    | Number                               | Any number (system-defined).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |                                                                                                                                                                        |
    | Temperature                          | Any temperature (system-defined).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |                                                                                                                                                                        |
    | Age                                  | Any age (system-defined).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |                                                                                                                                                                        |
    | Duration                             | Any time duration (system-defined).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |                                                                                                                                                                        |
    | Email                                | Any email address (system-defined).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |                                                                                                                                                                        |
    | Money                                | Any amount of money (system-defined). The input needs to include a number and a currency: <ul><li>**Number** — if the number has decimals, they need to be separated by the respective separator of the conversation's language. For example, separate decimals with a period for English and with a comma for German.</li><li>**Currency** — it is accepted as a symbol or written out, and can be written before or after the number.</li></ul>                                                                       | `1,300 dollars`, `111.21 USD` (English), `USD 75`, `43 $`, `$ 11`, `300 euros`, `300 Euro`, `150,00 EUR` (German), `150.00 EUR` (English), `EUR 28`, `1900 €`, `€ 200` |
    | URL                                  | Any reference/address to a resource on the Internet, for example, `http://example.com`.                                                                                                                                                                                                                                                                                                                                                                                                                                 |                                                                                                                                                                        |
    | Percentage                           | Any percentage (system-defined).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |                                                                                                                                                                        |
    | Regex                                | Any custom data format defined by a regex expression must be detected in the user's response. The regular expression must start with `/` and end with `/g`. For example, `/^1\d{7}$/g`.                                                                                                                                                                                                                                                                                                                                 |                                                                                                                                                                        |
    | Data                                 | Any data (input.data) input.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |                                                                                                                                                                        |
    | xApp                                 | Any xApp input.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         |                                                                                                                                                                        |
    | Custom                               | Any input.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |                                                                                                                                                                        |
    | Pattern: License Plate (DE)          | A pattern for the [German vehicle registration plate](https://en.wikipedia.org/wiki/Vehicle_registration_plates_of_Germany). This license plate is a unique alphanumeric identification tag displayed on a vehicle. It consists of letters, numbers, and sometimes special characters, for example, `ö`, `ä`, or `ü`. License plates serve as a means of identifying and registering vehicles, providing important information such as vehicle ownership, registration details, and compliance with legal requirements. | `M-345`, `x1Y2Z3`, `D 12345C`                                                                                                                                          |
    | Pattern: IBAN                        | A pattern for the International Bank Account Number (IBAN).                                                                                                                                                                                                                                                                                                                                                                                                                                                             | `DE12345678901234567890`                                                                                                                                               |
    | Pattern: Bank Identifier Code (BIC)  | A pattern for the Bank Identifier Code (BIC).                                                                                                                                                                                                                                                                                                                                                                                                                                                                           | `DEUTDEFF500`                                                                                                                                                          |
    | Pattern: Social Security Number (US) | A pattern for the US Social Security Number.                                                                                                                                                                                                                                                                                                                                                                                                                                                                            | `123-45-6789`                                                                                                                                                          |
    | Pattern: IP Address (IPv4)           | A pattern for the IPv4 address.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         | `192.168.1.1`                                                                                                                                                          |
    | Pattern: Phone Number                | A pattern for the phone number.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         | `+49 0000000000`, `+490000000000`, `(555) 000-000`                                                                                                                     |
    | Pattern: Credit Card                 | A pattern for the bank card.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            | `4111111111111111`                                                                                                                                                     |
    | LLM-extracted Entity                 | Utilizes a chosen LLM to extract entities, such as product codes, booking codes, and customer IDs, from a given string. Go to the LLM Entity Extraction Options.                                                                                                                                                                                                                                                                                                                                                        |                                                                                                                                                                        |

    <Note>
      All data formats supported by the Cognigy NLU for system slot mapping are listed on the [Slot Mapping](/ai/platform-features/nlu/slots/system-defined) page.
    </Note>
  </Accordion>

  <Accordion title="Channels and Output Types">
    Question Node output types carry the same functionality as the [Say Node](/ai/agents/develop/node-reference/basic/say).

    <Check>
      If you select the Date as a Question Type, the Question Node automatically renders a datepicker if the channel supports it. Refer to [Datepicker](/ai/agents/develop/node-reference/data/datepicker) for more information.
    </Check>
  </Accordion>

  <Accordion title="LLM Entity Extraction Options">
    This section appears if you've selected the **LLM-extracted Entity** question type.

    Before using this Question type,
    set the LLM provider in the [Settings](/ai/agents/develop/gen-ai-and-llms/generative-ai).
    You can configure the Node to either use the default model defined in the Settings or choose a specific configured LLM.

    | Parameter              | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
    | ---------------------- | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Large Language Model   | Select        | Select a model or use the default one.                                                                                                                                                                                                                                                                                                                                                                                                                                     |
    | Entity Name            | CognigyScript | The name of the entity to extract. For example, `customerID`.                                                                                                                                                                                                                                                                                                                                                                                                              |
    | Entity Description     | CognigyScript | A sentence which describes the entity. For example, `An alphanumeric string of 6 characters, such as ABC123 or 32G5FD`.                                                                                                                                                                                                                                                                                                                                                    |
    | Example Input          | Text          | Examples of text inputs. For example, `My ID is AB54EE, is that ok?`, `That would be ah bee see double 4 three`, `I guess it's 49 A B 8 K`.                                                                                                                                                                                                                                                                                                                                |
    | Extracted Entity       | CognigyScript | Examples of extracted entities. For example, `AB54EE`, `ABC443`, `49AB8K`.                                                                                                                                                                                                                                                                                                                                                                                                 |
    | Additional Validation  | CognigyScript | User input must meet this extra validation criteria, in addition to the built-in field validation, for example, Email, to be considered valid.                                                                                                                                                                                                                                                                                                                             |
    | Custom Model Options   | JSON          | Sets custom parameters for the LLM configuration, such as `temperature`, `top_k`, or `seed`. These parameters override the ones set in this Node. For the full list of parameters that your model supports, refer to the LLM provider's API documentation, for example, [OpenAI](https://platform.openai.com/docs/api-reference/chat/create) or [Azure OpenAI](https://learn.microsoft.com/en-us/azure/ai-services/openai/). Example: `{ "seed": 42, "temperature": 0.2 }` |
    | Custom Request Options | JSON          | Sets custom parameters for the request sent to the LLM provider, such as HTTP headers. These parameters override the ones set in this Node. For more information, refer to the LLM provider's API documentation. Example: `{ "headers": { "Authorization": "Bearer mytoken" } }`                                                                                                                                                                                           |

    Alternatively, you can add input examples in the **Use JSON Editor** code field. For example:

    ```json theme={null}
    {
      "My ID is AB54EE, is that ok?": "AB54EE",
      "That would be ah bee see double 4 three": "ABC443",
      "I guess it's 49 A B 8 K": "49AB8K"
    }
    ```
  </Accordion>

  <Accordion title="Advanced">
    | Parameter                    | Type      | Description                                                                                                                                                                                                                                          |
    | ---------------------------- | --------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Ignore deprecated parameters | Toggle    | Excludes the deprecated **Temperature** parameter from the request to the LLM provider. Activate this toggle if your model doesn't support this parameter. If you activate this toggle for a model that supports it, the model uses a default value. |
    | Temperature                  | Indicator | The appropriate sampling temperature for the model. Higher values mean the model takes more risks.                                                                                                                                                   |
    | Timeout                      | Number    | The maximum amount of milliseconds to wait for a response from the LLM provider.                                                                                                                                                                     |
  </Accordion>

  <Accordion title="Reprompt Options">
    Reprompt messages are automatically triggered if the question is not answered correctly,
    such as when the expected type of input is not provided or a validation does not return `true`.

    **Reprompt Methods**

    <Tabs>
      <Tab title="Simple Text">
        Outputs a simple text message to the user.

        | Parameter        | Type          | Description                                                                                                |
        | ---------------- | ------------- | ---------------------------------------------------------------------------------------------------------- |
        | Reprompt Message | CognigyScript | The message to output if the given answer is invalid. For example, `Not sure I understood this correctly`. |
      </Tab>

      <Tab title="Channel Message">
        Outputs a comprehensive, channel-specific message to the user.
        This message includes rich-media, for example, images or audio.
        You can define distinct messages for each channel, such as Webchat, Voice Gateway, WhatsApp, and others.

        | Parameter        | Type    | Description                                                                                                                                                                                                                                                |
        | ---------------- | ------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
        | Reprompt Message | Channel | The channel for which you want to configure the repromt message. You can select multiple channels.                                                                                                                                                         |
        | Output Type      | Select  | The format of the content that delivers through the selected channel. Depending on the selected channel, a list of possible output types suitable for that channel is presented. All channels support the text type, but support for other types may vary. |
      </Tab>

      <Tab title="LLM Prompt">
        Runs an LLM prompt and outputs the result in the format of a text message to the user.
        Before using this Node, configure a model in the [LLM](/ai/agents/develop/gen-ai-and-llms/providers/all-providers) settings.

        | Parameter                 | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                    |
        | ------------------------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
        | Large Language Model      | Select        | The channel for which you want to configure the repromt message. You can select multiple channels.                                                                                                                                                                                                                                                                                                                                             |
        | LLM Instructions          | CognigyScript | The input or command, such as a prompt or system message, that is given to the Language Model to generate an output.                                                                                                                                                                                                                                                                                                                           |
        | Transcript Turns          | Slider        | The number of conversation turns to include in the LLM chat completion request. By default, the value is `3`.                                                                                                                                                                                                                                                                                                                                  |
        | Stream Reprompt to Output | Toggle        | Streams the result directly into the output. This means that the model provides prompts directly into the conversation chat, as soon as a Stream Buffer Flush Token is matched, and you don't need to use the LLM Prompt Result token and Say Node. Note that streaming may not be supported by all [Cognigy LLM Prompt Node](/ai/agents/develop/gen-ai-and-llms/providers/all-providers) providers, such as Google<sup>[1](#footnote1)</sup>. |
      </Tab>

      <Tab title="Execute Flow and Return">
        Executes a specific Flow and returns to the question afterward.

        | Parameter      | Type   | Description                                                                                                                                          |
        | -------------- | ------ | ---------------------------------------------------------------------------------------------------------------------------------------------------- |
        | Select Flow    | Select | The Flow that is executed as the reprompt message.                                                                                                   |
        | Select Node    | Select | The Node from which the Flow should be executed as the reprompt message.                                                                             |
        | Parse Intents  | Toggle | This setting determines whether the system should parse Intents separately in the executed Flow. By default, this setting is enabled.                |
        | Parse Slots    | Toggle | This setting determines whether the system should parse Slots separately in the executed Flow. By default, this setting is enabled.                  |
        | Absorb Context | Toggle | This setting determines whether the system should absorb the executed Flow's Context into the current Context. By default, this setting is disabled. |

        **Repeat Question**

        Repeats a question if a user gives an invalid answer. By default, this setting is enabled.

        **Reprompt Condition**

        Optionally, you can add a [CognigyScript](/ai/platform-features/cognigyscript) condition to determine whether a reprompt message is shown.

        **Skip on Intent**

        Skips the reprompt if an Intent was found in the input.
      </Tab>
    </Tabs>
  </Accordion>

  <Accordion title="Result Storage">
    Question results are always stored in `input.result`.

    If **Store Result in Context** is enabled, the Question Result is also stored in the [Context](/ai/agents/develop/ai-agent-memory/context) object.

    If **Store Result to Contact Profile** is enabled, the Question Result is also stored in the [Profile](/ai/agents/develop/ai-agent-memory/profile) object.
  </Accordion>

  <Accordion title="Escalation - Intents">
    Allows the conversation to break out of the Question Node if a specified Intent was found.

    | Parameter         | Type     | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
    | ----------------- | -------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Escalation Action | Selector | Defines what happens when a question step escalates. Select one of the following options: <ul><li> **No Escalation** – no action is taken. This option is selected by default.</li> <li>**Output Message** – sends a message and behaves the same as a Say Node.</li><li>**Skip Question** – skips the question and stores a value under the `result` in the Input object.</li><li> **Go To Node** – jumps to a specific Node and continues from there and behaves as a Go To Node.</li><li>**Execute Flow and Return** – executes a Flow Node and then returns to the question step and behaves as an Execute Flow Node.</li><li>**Handover to Human Agent** – hands the conversation over to a human agent and behaves the same as a Handover to Human Agent Node.</li></ul> |
    | Send Queue Event  | Toggle   | Sends an event to the Flow when the handover is queued.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
    | Send Active Event | Toggle   | Sends an event to the Flow when the handover becomes active.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |

    <Accordion title="Escalation Action: Output Message">
      | Parameter                     | Type   | Description                                                                                                                                                                   |
      | ----------------------------- | ------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
      | Valid Intents                 | List   | A list of Intents that can trigger the escalation. Press **Enter** to add each Intent. Only Intents included in this list will be considered for escalation.                  |
      | Intent Score Threshold        | Number | The minimum confidence score an Intent must have to trigger the escalation. If the detected Intent's score is below this threshold, the escalation will not occur.            |
      | Escalation Message            | Text   | The message the AI Agent sends when the wrong answer count is reached. Supports CognigyScript and [Tokens](/ai/platform-features/tokens).                                     |
      | Output Type                   | Toggle | Defines how the escalation message is sent. For example, **Text** will render the message as plain text in the channel, supporting emojis, HTML formatting, or CognigyScript. |
      | Linear                        | Toggle | Controls whether multiple escalation messages are shown in a specific order (linearly) or randomly. If enabled, messages appear in the defined sequence.                      |
      | Loop                          | Toggle | Works with **Linear**. If **Loop** is enabled, the sequence starts over after the last message. If disabled, the last message repeats after reaching the end.                 |
      | Data                          | JSON   | Additional data to send with the escalation message. For example, `{ "type": "escalation" }`.                                                                                 |
      | Prevent Reprompt Message Once | Toggle | If enabled, the escalation message will only be triggered once per conversation, preventing repeated escalation messages.                                                     |
    </Accordion>

    <Accordion title="Escalation Action: Skip Question">
      | Parameter              | Type   | Description                                                                                                                                                        |
      | ---------------------- | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
      | Valid Intents          | List   | A list of Intents that can trigger the escalation. Press **Enter** to add each Intent. Only Intents included in this list will be considered for escalation.       |
      | Intent Score Threshold | Number | The minimum confidence score an Intent must have to trigger the escalation. If the detected Intent's score is below this threshold, the escalation will not occur. |
    </Accordion>

    <Accordion title="Escalation Action: Go To Node">
      | Parameter              | Type     | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
      | ---------------------- | -------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
      | Valid Intents          | List     | A list of Intents that can trigger the escalation. Press **Enter** to add each Intent. Only Intents included in this list will be considered for escalation.                                                                                                                                                                                                                                                                                                                                                                               |
      | Intent Score Threshold | Number   | The minimum confidence score an Intent must have to trigger the escalation. If the detected Intent's score is below this threshold, the escalation will not occur.                                                                                                                                                                                                                                                                                                                                                                         |
      | Select Flow            | Flow ID  | Select the Flow to switch to when the escalation triggers. Use **Manual Input** to enter the Flow ID directly, or click the target Flow icon to select.                                                                                                                                                                                                                                                                                                                                                                                    |
      | Select Node            | Node ID  | Select the Node to switch to when the escalation triggers. The Node must be an Entrypoint. Use **Manual Input** to enter the Node ID directly.                                                                                                                                                                                                                                                                                                                                                                                             |
      | GoTo Execution Mode    | Selector | Determines the behavior when switching to the selected Node/Flow: <ul> <li>**Go to Node and continue** – the conversation immediately continues in the target Node or Flow using the current input. You can override the text or data using **Optional Injected Text** or **Optional Injected Data**.</li><li>**Go to Node and Wait for input** – the conversation switches to the target Node/Flow but pauses, waiting for the user to provide input before continuing. Useful when you want the escalation to prompt the user.</li></ul> |
      | Optional Injected Text | Text     | When the escalation triggers, you can overwrite the text sent with this field. If left empty, the default text from **Escalation Message** is used.                                                                                                                                                                                                                                                                                                                                                                                        |
      | Optional Injected Data | JSON     | When the escalation triggers, you can overwrite the data sent with this field. If left empty, the default data is used.                                                                                                                                                                                                                                                                                                                                                                                                                    |
    </Accordion>

    <Accordion title="Escalation Action: Execute Flow and Return">
      | Parameter              | Type    | Description                                                                                                                                                        |
      | ---------------------- | ------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
      | Valid Intents          | List    | A list of Intents that can trigger the escalation. Press **Enter** to add each Intent. Only Intents included in this list will be considered for escalation.       |
      | Intent Score Threshold | Number  | The minimum confidence score an Intent must have to trigger the escalation. If the detected Intent's score is below this threshold, the escalation will not occur. |
      | Select Flow            | Flow ID | Select the Flow to switch to when the escalation triggers. Use **Manual Input** to enter the Flow ID directly.                                                     |
      | Select Node            | Node ID | Select the Node to switch to when the escalation triggers. The Node must be an Entrypoint. Use **Manual Input** to enter the Node ID directly.                     |
    </Accordion>

    <Accordion title="Escalation Action: Handover to Human Agent">
      | Parameter                                   | Type          | Description                                                                                                                                                                                                                                                                             |
      | ------------------------------------------- | ------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
      | Valid Intents                               | List          | A list of Intents that can trigger the escalation. Press **Enter** to add each Intent. Only Intents included in this list will be considered for escalation.                                                                                                                            |
      | Intent Score Threshold                      | Number        | The minimum confidence score an Intent must have to trigger the escalation. If the detected Intent's score is below this threshold, the escalation will not occur.                                                                                                                      |
      | Handover Accepted Message                   | CognigyScript | The message that is displayed to the user when the handover request is accepted. It confirms to the user that a human agent will take over the conversation, providing reassurance that their issue is being handled.                                                                   |
      | Repeat Handover Accepted Message            | Toggle        | Repeatedly outputs the Handover Accepted Message if the user sends additional messages while waiting in the queue. This parameter keeps the user informed and reassured during the wait. The parameter is activated by default.                                                         |
      | Cancel Intent                               | CognigyScript | Specifies the Intent that a user can trigger to cancel a handover request. This parameter allows users to regain control of the conversation if they change their mind or no longer want to wait for a human agent.                                                                     |
      | Cancel Button Text                          | CognigyScript | Provides the text for a quick-reply button that users can click to cancel the handover. This parameter is an alternative to triggering the cancel Intent manually, offering a more robust option to stop the handover process.                                                          |
      | Flow Continuation                           | Selector      | Determines where the conversation resumes after the handover is completed. Select one of the following options: <ul><li>**At current AgentInject Entrypoint** – Start again where the handover began.</li><li>**Below this Node** – Continue with the next step in the chat.</li></ul>  |
      | Send resolve event to Virtual Agent         | Toggle        | Sends an event to the AI Agent when the human agent resolves the conversation. This parameter allows the AI Agent to respond appropriately, update conversation status, or trigger additional automated processes after the human interaction. The parameter is deactivated by default. |
      | Handover Provider                           | Selector      | The provider that will handle the handover. The list contains the handover providers created in **Deploy > Handover Providers**.                                                                                                                                                        |
      | Additional Category IDs                     | CognigyScript | The IDs for the RingCentral Engage handover provider to allow you to categorize or route conversations according to custom categories. This parameter ensures that the handover is directed to the appropriate team or workflow.                                                        |
      | Live Agent Inbox Id                         | CognigyScript | The inbox ID used to route the conversation to the human agent. This ID must match the inbox configured in the Live Agent system to ensure messages are delivered correctly.                                                                                                            |
      | Live Agent – Allow reply as user            | Toggle        | Lets human agents respond on behalf of the end user in the chat interface. This parameter is useful for scenarios where a human agent needs to simulate user actions or correct information during the conversation. The parameter is deactivated by default.                           |
      | Chatwoot Inbox Id                           | CognigyScript | The ID for routing the conversation to a Chatwoot inbox. This ensures proper delivery and handling within Chatwoot's system.                                                                                                                                                            |
      | Salesforce Send Transcript As First Message | Toggle        | Sends the conversation transcript to Salesforce as the first message when handover begins. If the transcript is long, it is split into multiple messages. This parameter helps human agents have context at the start of the interaction. The parameter is deactivated by default.      |
      | Salesforce Prechat Entities                 | JSON          | Refers to structured key-value data used for routing, automation, or passing contextual information to Salesforce flows. These are often used internally by Salesforce to map chat data to variables in Omni-Channel or automation logic.                                               |
      | Salesforce Prechat Details                  | JSON          | Refers to user-facing or pre-filled field data in Salesforce chat forms, for example, name, email, issue type. This data is used to pre-populate fields the agent sees when the conversation starts.                                                                                    |
      | NiCE CXone Send Transcript As First Message | Toggle        | Sends the conversation transcript to NiCE CXone as the first message when handover begins. This parameter helps human agents have context at the start of the interaction. The parameter is activated by default.                                                                       |
      | Language                                    | CognigyScript | Defines the languages for the conversation. This ensures the user is connected to agents who can communicate effectively in the chosen language.                                                                                                                                        |
      | Skills                                      | CognigyScript | Specifies the skills required to handle the conversation. Agents are filtered based on these skills, ensuring users are connected to qualified personnel for their issue.                                                                                                               |
      | Priority                                    | CognigyScript | Sets the priority of the conversation. Higher-priority conversations may be routed faster or treated with urgency in the contact center.                                                                                                                                                |
      | Custom Attributes                           | JSON          | Adds additional key-value metadata to provide context or support custom routing and reporting. These attributes can include any relevant information about the user, conversation, or business context.                                                                                 |
      | 8x8 Channel Id                              | CognigyScript | Specifies the ID of the 8x8 chat channel used for handover. Links the conversation to the correct channel.                                                                                                                                                                              |
      | 8x8 Queue Id                                | CognigyScript | Identifies the queue in 8x8 where the conversation should be routed.                                                                                                                                                                                                                    |
      | 8x8 JSON Properties                         | JSON          | Allows sending additional custom JSON properties to 8x8 for configuration, routing, or passing extra metadata.                                                                                                                                                                          |
      | Only escalate once                          | Toggle        | Ensures that a conversation is handed over only once. Prevents repeated escalations for the same issue to avoid confusion or redundant agent interactions.                                                                                                                              |
      | Enable User Connects Message                | Toggle        | Notifies human agents when a user reconnects to the chat. This parameter ensures agents are aware of returning users and can continue the conversation seamlessly.                                                                                                                      |
      | Enable User Disconnects Message             | Toggle        | Notifies human agents when a user leaves or closes the chat. This parameter helps maintain awareness of user engagement and session status.                                                                                                                                             |
      | Display Agent Details                       | Toggle        | Displays the human agent's name and avatar from Genesys or NiCE CXone in the chat for the end user. This parameter is disabled by default.                                                                                                                                              |
      | Get Queue Position                          | Toggle        | Retrieves the end user's current queue position. Supported by Cognigy Live Agent and Salesforce MIAW. The parameter is deactivated by default.                                                                                                                                          |
      | Get Estimated Wait Time                     | Toggle        | Retrieves the estimated wait time for the end user. End users then receive wait time updates while they wait to be connected to a human agent. Supported by Cognigy Live Agent, Salesforce MIAW, and Genesys Cloud Open Messaging. The parameter is deactivated by default.             |
      | Update Interval                             | CognigyScript | Sets the interval, in seconds, for how often to receive updates on the queue position or estimated wait time. This parameter appears if **Get Queue Position** or **Get Estimated Wait Time** is selected.                                                                              |
      | Alternative Update                          | Toggle        | Sends an alternative message to the end user when the maximum queue position or the maximum estimated wait time is reached. The parameter is deactivated by default.                                                                                                                    |
      | Maximum Queue Position                      | CognigyScript | Sets the threshold for the maximum queue position. If this threshold is reached, the alternative text is sent. This parameter appears if **Alternative Update** is selected.                                                                                                            |
      | Maximum Estimated Wait Time                 | CognigyScript | Defines the maximum estimated wait time in milliseconds. If this duration is exceeded, the alternative text is sent. This parameter appears if **Alternative Update** is selected.                                                                                                      |
      | Alternative Text                            | CognigyScript | The message sent to the user when the maximum queue position or wait time threshold is reached. This ensures users are informed and reduces frustration while waiting for a human agent.                                                                                                |
    </Accordion>
  </Accordion>

  <Accordion title="Escalation - Wrong Answers">
    Allows the conversation to break out of the Question Node after a number of incorrect answers were provided.

    | Parameter         | Type     | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
    | ----------------- | -------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Escalation Action | Selector | Defines what happens when a question step escalates. Select one of the following options: <ul><li> **No Escalation** – no action is taken. This option is selected by default.</li> <li>**Output Message** – sends a message and behaves the same as a Say Node.</li><li>**Skip Question** – skips the question and stores a value under the `result` in the Input object.</li><li> **Go To Node** – jumps to a specific Node and continues from there and behaves as a Go To Node.</li><li>**Execute Flow and Return** – executes a Flow Node and then returns to the question step and behaves as an Execute Flow Node.</li><li>**Handover to Human Agent** – hands the conversation over to a human agent and behaves the same as a Handover to Human Agent Node.</li></ul> |
    | Send Queue Event  | Toggle   | Sends an event to the Flow when the handover is queued.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
    | Send Active Event | Toggle   | Sends an event to the Flow when the handover becomes active.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |

    <Accordion title="Escalation Action: Output Message">
      | Parameter                      | Type   | Description                                                                                                                                                                                |
      | ------------------------------ | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
      | Wrong Answer Count Trigger     | Number | The number of incorrect responses a user can give before the escalation message is triggered. For example, if set to `1`, the escalation message will appear after the first wrong answer. |
      | Escalation Message             | Text   | The message the AI Agent sends when the wrong answer count is reached. Supports CognigyScript and [Tokens](/ai/platform-features/tokens).                                                  |
      | Output Type                    | Toggle | Defines how the escalation message is sent. For example, **Text** will render the message as plain text in the channel, supporting emojis, HTML formatting, or CognigyScript.              |
      | Linear                         | Toggle | Controls whether multiple escalation messages are shown in a specific order (linearly) or randomly. If enabled, messages appear in the defined sequence.                                   |
      | Loop                           | Toggle | Works with **Linear**. If **Loop** is enabled, the sequence starts over after the last message. If disabled, the last message repeats after reaching the end.                              |
      | Data                           | JSON   | Additional data to send with the escalation message. For example, `{ "type": "escalation" }`.                                                                                              |
      | Prevent Reprompt on Escalation | Toggle | If enabled, the escalation message will only be triggered once per conversation, preventing repeated escalation messages.                                                                  |
      | Only escalate once             | Toggle | Ensures that a conversation is handed over only once. Prevents repeated escalations for the same issue to avoid confusion or redundant agent interactions.                                 |
    </Accordion>

    <Accordion title="Escalation Action: Skip Question">
      | Parameter                  | Type   | Description                                                                                                                                                                                |
      | -------------------------- | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
      | Wrong Answer Count Trigger | Number | The number of incorrect responses a user can give before the escalation message is triggered. For example, if set to `1`, the escalation message will appear after the first wrong answer. |
      | Only escalate once         | Toggle | Ensures that a conversation is handed over only once. Prevents repeated escalations for the same issue to avoid confusion or redundant agent interactions.                                 |
    </Accordion>

    <Accordion title="Escalation Action: Go To Node">
      | Parameter                  | Type     | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
      | -------------------------- | -------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
      | Wrong Answer Count Trigger | Number   | Set the number of incorrect responses a user can give before the escalation action is triggered. For example, if set to `1`, the escalation will occur after the first wrong answer.                                                                                                                                                                                                                                                                                                                                                       |
      | Select Flow                | Flow ID  | Select the Flow to switch to when the escalation triggers. Use **Manual Input** to enter the Flow ID directly, or click the target Flow icon to select.                                                                                                                                                                                                                                                                                                                                                                                    |
      | Select Node                | Node ID  | Select the Node to switch to when the escalation triggers. The Node must be an Entrypoint. Use **Manual Input** to enter the Node ID directly.                                                                                                                                                                                                                                                                                                                                                                                             |
      | GoTo Execution Mode        | Selector | Determines the behavior when switching to the selected Node/Flow: <ul> <li>**Go to Node and continue** – the conversation immediately continues in the target Node or Flow using the current input. You can override the text or data using **Optional Injected Text** or **Optional Injected Data**.</li><li>**Go to Node and Wait for input** – the conversation switches to the target Node/Flow but pauses, waiting for the user to provide input before continuing. Useful when you want the escalation to prompt the user.</li></ul> |
      | Optional Injected Text     | Text     | When the escalation triggers, you can overwrite the text sent with this field. If left empty, the default text from **Escalation Message** is used.                                                                                                                                                                                                                                                                                                                                                                                        |
      | Optional Injected Data     | JSON     | When the escalation triggers, you can overwrite the data sent with this field. If left empty, the default data is used.                                                                                                                                                                                                                                                                                                                                                                                                                    |
      | Only escalate once         | Toggle   | If enabled, the escalation message is triggered only once per conversation, preventing repeated escalation messages.                                                                                                                                                                                                                                                                                                                                                                                                                       |
    </Accordion>

    <Accordion title="Escalation Action: Execute Flow and Return">
      | Parameter                  | Type    | Description                                                                                                                                                                      |
      | -------------------------- | ------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
      | Wrong Answer Count Trigger | Number  | The number of incorrect responses a user can give before the escalation action is triggered. For example, if set to `1`, the escalation will occur after the first wrong answer. |
      | Select Flow                | Flow ID | Select the Flow to switch to when the escalation triggers. Use **Manual Input** to enter the Flow ID directly.                                                                   |
      | Select Node                | Node ID | Select the Node to switch to when the escalation triggers. The Node must be an Entrypoint. Use **Manual Input** to enter the Node ID directly.                                   |
      | Only escalate once         | Toggle  | If enabled, the escalation action is triggered only once per conversation, preventing repeated escalation triggers.                                                              |
    </Accordion>

    <Accordion title="Escalation Action: Handover to Human Agent">
      | Parameter                                   | Type          | Description                                                                                                                                                                                                                                                                             |
      | ------------------------------------------- | ------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
      | Wrong Answer Count Trigger                  | Slider        | Defines the number of consecutive incorrect answers the AI Agent can give before automatically triggering a handover. This helps ensure users aren't stuck with repeated wrong responses and are directed to a human agent when necessary.                                              |
      | Handover Accepted Message                   | CognigyScript | The message that is displayed to the user when the handover request is accepted. It confirms to the user that a human agent will take over the conversation, providing reassurance that their issue is being handled.                                                                   |
      | Repeat Handover Accepted Message            | Toggle        | Repeatedly outputs the Handover Accepted Message if the user sends additional messages while waiting in the queue. This parameter the user informed and reassured during the wait. The parameter is activated by default.                                                               |
      | Cancel Intent                               | CognigyScript | Specifies the Intent that a user can trigger to cancel a handover request. This parameter allows users to regain control of the conversation if they change their mind or no longer want to wait for a human agent.                                                                     |
      | Cancel Button Text                          | CognigyScript | Provides the text for a quick-reply button that users can click to cancel the handover. This parameter is an alternative to triggering the cancel Intent manually, offering a more robust option to stop the handover process.                                                          |
      | Flow Continuation                           | Selector      | Determines where the conversation resumes after the handover is completed. Select one of the following options: <ul><li>**At current AgentInject Entrypoint** – Start again where the handover began.</li><li>**Below this Node** – Continue with the next step in the chat.</li></ul>  |
      | Send resolve event to Virtual Agent         | Toggle        | Sends an event to the AI Agent when the human agent resolves the conversation. This parameter allows the AI Agent to respond appropriately, update conversation status, or trigger additional automated processes after the human interaction. The parameter is deactivated by default. |
      | Handover Provider                           | Selector      | The provider that will handle the handover. The list contains the handover providers created in **Deploy > Handover Providers**.                                                                                                                                                        |
      | Additional Category IDs                     | CognigyScript | The IDs for the RingCentral Engage handover provider to allow you to categorize or route conversations according to custom categories. This parameter ensures that the handover is directed to the appropriate team or workflow.                                                        |
      | Live Agent Inbox Id                         | CognigyScript | The inbox ID used to route the conversation to the human agent. This ID must match the inbox configured in the Live Agent system to ensure messages are delivered correctly.                                                                                                            |
      | Live Agent – Allow reply as user            | Toggle        | Lets human agents respond on behalf of the end user in the chat interface. This parameter is useful for scenarios where a human agent needs to simulate user actions or correct information during the conversation. The parameter is deactivated by default.                           |
      | Chatwoot Inbox Id                           | CognigyScript | The ID for routing the conversation to a Chatwoot inbox. This ensures proper delivery and handling within Chatwoot's system.                                                                                                                                                            |
      | Salesforce Send Transcript As First Message | Toggle        | Sends the conversation transcript to Salesforce as the first message when handover begins. If the transcript is long, it is split into multiple messages. This parameter helps human agents have context at the start of the interaction. The parameter is deactivated by default.      |
      | Salesforce Prechat Entities                 | JSON          | Refers to structured key-value data used for routing, automation, or passing contextual information to Salesforce flows. These are often used internally by Salesforce to map chat data to variables in Omni-Channel or automation logic.                                               |
      | Salesforce Prechat Details                  | JSON          | Refers to user-facing or pre-filled field data in Salesforce chat forms, for example, name, email, issue type. This data is used to pre-populate fields the agent sees when the conversation starts.                                                                                    |
      | NiCE CXone Send Transcript As First Message | Toggle        | Sends the conversation transcript to NiCE CXone as the first message when handover begins. This parameter helps human agents have context at the start of the interaction. The parameter is activated by default.                                                                       |
      | Language                                    | CognigyScript | Defines the languages for the conversation. This ensures the user is connected to agents who can communicate effectively in the chosen language.                                                                                                                                        |
      | Skills                                      | CognigyScript | Specifies the skills required to handle the conversation. Agents are filtered based on these skills, ensuring users are connected to qualified personnel for their issue.                                                                                                               |
      | Priority                                    | CognigyScript | Sets the priority of the conversation. Higher-priority conversations may be routed faster or treated with urgency in the contact center.                                                                                                                                                |
      | Custom Attributes                           | JSON          | Adds additional key-value metadata to provide context or support custom routing and reporting. These attributes can include any relevant information about the user, conversation, or business context.                                                                                 |
      | 8x8 Channel Id                              | CognigyScript | Specifies the ID of the 8x8 chat channel used for handover. Links the conversation to the correct channel.                                                                                                                                                                              |
      | 8x8 Queue Id                                | CognigyScript | Identifies the queue in 8x8 where the conversation should be routed.                                                                                                                                                                                                                    |
      | 8x8 JSON Properties                         | JSON          | Allows sending additional custom JSON properties to 8x8 for configuration, routing, or passing extra metadata.                                                                                                                                                                          |
      | Only escalate once                          | Toggle        | Ensures that a conversation is handed over only once. Prevents repeated escalations for the same issue to avoid confusion or redundant agent interactions.                                                                                                                              |
      | Enable User Connects Message                | Toggle        | Notifies human agents when a user reconnects to the chat. This parameter ensures agents are aware of returning users and can continue the conversation seamlessly.                                                                                                                      |
      | Enable User Disconnects Message             | Toggle        | Notifies human agents when a user leaves or closes the chat. This parameter helps maintain awareness of user engagement and session status.                                                                                                                                             |
      | Display Agent Details                       | Toggle        | Displays the human agent's name and avatar from Genesys or NiCE CXone in the chat for the end user. This parameter is disabled by default.                                                                                                                                              |
      | Get Queue Position                          | Toggle        | Retrieves the end user's current queue position. Supported by Cognigy Live Agent and Salesforce MIAW. The parameter is deactivated by default.                                                                                                                                          |
      | Get Estimated Wait Time                     | Toggle        | Retrieves the estimated wait time for the end user. End users then receive wait time updates while they wait to be connected to a human agent. Supported by Cognigy Live Agent, Salesforce MIAW, and Genesys Cloud Open Messaging. The parameter is deactivated by default.             |
      | Update Interval                             | CognigyScript | Sets the interval, in seconds, for how often to receive updates on the queue position or estimated wait time. This parameter appears if **Get Queue Position** or **Get Estimated Wait Time** is selected.                                                                              |
      | Alternative Update                          | Toggle        | Sends an alternative message to the end user when the maximum queue position or the maximum estimated wait time is reached. The parameter is deactivated by default.                                                                                                                    |
      | Maximum Queue Position                      | CognigyScript | Sets the threshold for the maximum queue position. If this threshold is reached, the alternative text is sent. This parameter appears if **Alternative Update** is selected.                                                                                                            |
      | Maximum Estimated Wait Time                 | CognigyScript | Defines the maximum estimated wait time in milliseconds. If this duration is exceeded, the alternative text is sent. This parameter appears if **Alternative Update** is selected.                                                                                                      |
      | Alternative Text                            | CognigyScript | The message sent to the user when the maximum queue position or wait time threshold is reached. This ensures users are informed and reduces frustration while waiting for a human agent.                                                                                                |
    </Accordion>
  </Accordion>

  <Accordion title="Reconfirmation Options">
    Allows for answers to be reconfirmed before continuing.
    This is especially useful when using voice agents and reconfirming what the agent understood
    (for example, in Number questions when the user said "*my number is three double five triple nine five six eight*").
    The answer given to the reconfirmation question has
    to be a yes/no style answer and follows the same rules as a Yes/No Question.

    Reconfirmation Questions can contain a specific token `ANSWER`, which is replaced with a short form version of the given answer (for example, "3 EUR" in a Money question). The short form answer is taken from `input.activeQuestion.tentativeShortFormAnswer`;

    Reconfirmation Questions can have a specific re-prompt set, which is output before the question if the answer to the question is not of yes/no style.
  </Accordion>

  <Accordion title="Advanced">
    **Store detailed Results**

    This setting, when enabled, stores a more detailed JSON object under the result property of the input. This is useful in case more information is needed.

    **Skip if Answer in Input**

    When enabled, this setting skips the Question if the answer is already provided in the input text.

    **Additional Validation**

    A [CognigyScript](/ai/platform-features/cognigyscript) condition that validates the answer. The answer is considered valid only if the condition returns `true`. For example, if you enter `input.slots.EMAIL[0].endsWith("cognigy.com")` in this parameter for an email question type, only email addresses ending with `cognigy.com` pass the validation.

    **Result Location**

    By default, when the Question Node recognizes the user's answer, the Node stores the answer under `result` in the Input object. The Result Location parameter lets you replace the recognized answer with a value from another JSON path, for example, `input.bookingReference`. The Result Location parameter only works if the user's answer is recognized and if the path you set is valid. If the path is invalid, the original answer isn't recognized, and the user is reprompted.

    <Note>
      If you use both the Result Location parameter and the Additional Validation parameter at the same time, the Additional Validation parameter checks only the user answer, not the value in the JSON path set in the Result Location parameter.
    </Note>

    **Forget Question Threshold**

    This setting determines how long a user can have been "away" from the Node after the question was initially asked. With the default setting `1` this means that the question has to be answered on the next user input. If a user input comes back to the question at a later stage, it is treated as if the question was hit for the first time and the question is asked.
  </Accordion>

  <Accordion title="Handover to Human Agent">
    | Action                      | Description                                                                                                                                                                             |
    | --------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Handover Output Destination | Defines how messages are delivered after the handover occurs: **User and Agent**, **Agent Only**, or **User Only**. This option is available only for the Live Agent handover provider. |
  </Accordion>

  <Accordion title="AI-Enhanced Output">
    To use AI-enhanced output rephrasing, read the [Generative AI](/ai/agents/develop/gen-ai-and-llms/generative-ai) article.
  </Accordion>

  <Accordion title="Answer Preprocessing">
    You can use various functions of the [Text Cleaner](/ai/platform-features/nlu/text-cleaner) class to preprocess the answer to a question before it is evaluated. This can be helpful, for example, when requesting a name using a `text` type question or when asking for a part number using a `slot` question.

    In addition to the Text Cleaner functions, users have the option to rerun NLU after the cleaning process. This approach allows for tasks such as re-detecting slots or properly filling any remaining slots.
  </Accordion>
</AccordionGroup>

### Exclude from Transcript

Excludes the Node output from the conversation transcript. The output remains visible to the end user but isn't stored in the [`transcript`](/ai/agents/develop/ai-agent-memory/input#nested-objects) object or shared with the LLM provider.

You can use this parameter to:

* Hide sensitive or irrelevant data, such as legal disclaimers, so the model doesn't see or repeat them.
* Prevent the model from copying patterns (called in-context learning) you didn't want it to learn.

<Accordion title="Example">
  By default, the model repeats the question style it learned from the AI Agent's earlier question, even though the end user asked for an answer, not a question.

  ```txt theme={null}
  AI Agent: What is your favorite color? (included in the conversation transcript)
  End User: Blue.
  Later:
  End User: Tell me your favorite food.
  AI Agent: What is your favorite food?
  ```

  By excluding the AI Agent's earlier question from the transcript, the same conversation looks like this:

  ```txt theme={null}
  AI Agent: What is your favorite color? (included in the conversation transcript)
  End User: Blue.
  Later:
  End User: Tell me your favorite food.
  AI Agent: I enjoy pizza.
  ```

  Use this parameter to maintain confidentiality, for example, prevent sensitive data from reaching the LLM, or to display messages such as legal disclaimers or system notes that shouldn't affect the AI Agent's behavior.
</Accordion>

## Question Information in Input

When a question is active, meaning that the AI Agent is waiting for the answer, information regarding the question is added to the Input object.

```JSON theme={null}
"activeQuestion": {
    "nodeId": "18b158bf-71a3-4d4f-a31f-812b1810f8af",
    "type": "yesNo",
    "lastExecutedAt": 2,
    "forgetQuestionThreshold": 1,
    "repromptCount": 1,
    "escalationCount": 0
}
```

This information can be used to trigger specific actions on escalation or to jump back to the Question Node after an escalation.

<Accordion title="Slot Fillers">
  Questions can be combined with **Slot Fillers** to create a "Missing Pattern". This mechanism keeps asking the user for the missing information in a very natural way, until all questions have been answered.
</Accordion>

***

<sup>1</sup>: Note that not all LLM models support streaming.


# Say
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/basic/say



<a href="/release-notes/2025.20"><Badge>Updated in 2025.20</Badge></a>

<Frame>
  <img alt="Say Node configuration panel" />
</Frame>

## Description

The Say Node sends a message to the user. This message can have different formats, ranging from a simple text to rich media such as galleries, adaptive cards, videos, and audio.

Additionally, you can enhance the existing output with [Generative AI](#ai-enhanced-output) and control who can receive messages when the conversation is handed over to a contact center.

## Parameters

### Channels

Depending on the current channel, additional rich media formats are available. Add a new channel output by clicking the <img alt="add-channel" /> icon and selecting the channel that corresponds to the [Endpoint](/ai/agents/deploy/endpoints/overview) that is deployed.

Refer to the [available channels](/ai/agents/develop/nodes/channels).

#### Default Cognigy Channel

The default Cognigy channel allows for the configuration of different output types.
Not all [Endpoints](/ai/agents/deploy/endpoints/overview) can correctly convert your content to the desired output type.
Before configuration,
[check the compatibility](/ai/agents/deploy/endpoints/content-conversion) of the output type with the Endpoint.

##### Output Type

<AccordionGroup>
  <Accordion title="Text">
    The **Text** output type renders text and emojis, if supported by the channel. The text field also supports [CognigyScript](/ai/platform-features/cognigyscript) and [Tokens](/ai/platform-features/tokens) that can be added by clicking <img alt="token" /> at the end of each field.

    | Parameters | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
    | ---------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Text       | CognigyScript | The message you want the AI Agent to send. If you want to customize the text by adding italic or bold styles or including links, only HTML is supported.<br /><br />You can add multiple text messages — one message per field. Enter the first message and press <kbd>Enter</kbd> to add additional messages.<br /><br />When multiple text messages are added, the delivery order is controlled by the **Linear** and **Loop** parameters available in the **Options** section. Only one message is delivered with each Node execution. If you want the AI Agent to send two or more messages at once, add a new Say Node. |

    **Options**

    When you send text responses, you can control how the text is delivered and attach additional data to the message.

    | Parameter | Type   | Description                                                                                                                                                                                                                                            |
    | --------- | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Linear    | Toggle | Controls whether the text responses are shown in a specific order (linearly) or randomly. If enabled, the text options are presented one after the other in the defined sequence.                                                                      |
    | Loop      | Toggle | Works in conjunction with the **Linear** option. If **Linear** and **Loop** are enabled, the sequence starts over from the beginning after reaching the end. If **Loop** is disabled, the last text option will keep repeating after reaching the end. |
    | Data      | JSON   | Allows you to include additional data that you want to send along with the message to the client. For example, `{ "type": "motivational" }`.                                                                                                           |

    When configuring the **Linear** and **Loop** parameters, the message delivery behavior differs:

    | Behavior             | Configuration              | Example Conversation                                                                                                                                                                                                                                                                                                                                                                                             |
    | -------------------- | -------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Random Behavior      | Linear: Off<br />Loop: Off | User: `Give me some motivation.`<br />AI Agent: `You are doing great!`<br />User: `Another one, please.`<br />AI Agent: `Believe in yourself!`<br />User: `More motivation.`<br />AI Agent: `Keep pushing forward!`<br /><br />Responses are in random order each time.                                                                                                                                          |
    | Linear + Non-looping | Linear: On<br />Loop: Off  | User: `Give me some motivation.`<br />AI Agent: `Keep pushing forward!`<br />User: `Another one, please.`<br />AI Agent: `Believe in yourself!`<br />User: `More motivation.`<br />AI Agent: `You are doing great!`<br />User: `One more, please.`<br />AI Agent: `You are doing great!` (repeats last message)<br /><br />Responses follow a fixed sequence and repeat the last message after reaching the end. |
    | Linear + Looping     | Linear: On<br />Loop: On   | User: `Give me some motivation.`<br />AI Agent: `Keep pushing forward!`<br />User: `Another one, please.`<br />AI Agent: `Believe in yourself!`<br />User: `More motivation.`<br />AI Agent: `You are doing great!`<br />User: `Next!`<br />AI Agent: `Keep pushing forward!` (starts sequence over)<br /><br />Responses cycle through in a fixed order and restart from the beginning after reaching the end.  |
  </Accordion>

  <Accordion title="Text with Quick Replies">
    The **Text with Quick Replies** output type can be used to show the user a number of configurable quick replies.
    Quick replies are pre-defined answers that are rendered as input chips.

    | Parameters                | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
    | ------------------------- | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Text                      | CognigyScript | The message you want the AI Agent to send. If you want to customize the text by adding italic or bold styles or including links, note that Markdown isn't supported; only HTML is allowed.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
    | Add Quick Reply           | Button        | Adds a new Empty Quick Reply button. You can add multiple buttons.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | Empty Quick Reply         | Button        | Adds a new Quick Reply button.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
    | Button Title              | CognigyScript | Appears only when the Empty Quick Reply parameter is selected. The title of the button.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
    | Select Button Type        | Dropdown      | You can select the following options:<ul><li>**Postback Value** — sends a specified value to the start of the Flow, simulating user input as if the user had manually typed it. This behavior is typical for Quick Replies and is used to trigger specific responses or actions based on predefined values.</li><li>**Phone Number** — opens the phone application on the device when the button is clicked, allowing users to initiate a call directly from the chat interface.</li><li>**Trigger Intent** — allows you to manually trigger an Intent by using the `cIntent:` prefix, followed by the name of the desired Intent. When this option is used, the regular Intent mapping is bypassed, and the specified Intent is activated directly. For more information, read [Trigger Intent](/ai/platform-features/nlu/intents/trigger-intent).</li><li>**Open xApp** — opens an xApp page. Note that, for the xApp session link to be available, you need to add the [xApp: Init Session](/ai/agents/develop/node-reference/xApp/init-xApp-session) and either the [xApp: Show HTML](/ai/agents/develop/node-reference/xApp/set-html-xApp-state) or [xApp: Show Adaptive Card Nodes](/ai/agents/develop/node-reference/xApp/set-AdaptiveCard-xApp-state) before the Say Node.</li></ul> |
    | Postback Value            | CognigyScript | Appears if the Post back value in Select Button Type is selected. The value that is sent to the start of the Flow, simulating user input as if the user had manually typed it.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
    | Phone number              | CognigyScript | Appears if the Phone number value in Select Button Type is selected. The number must be formatted as a valid phone number and shouldn't contain spaces or special characters.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
    | Intent name               | CognigyScript | Appears if the Trigger Intent value in Select Button Type is selected. The Intent name that should be triggered.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | Image URL                 | CognigyScript | The URL for the image. Note that if you upload a URL to a storage service, such as Google Cloud Storage or AWS, the URL should be publicly accessible.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
    | Image Alternate Text      | CognigyScript | The alternative text for accessibility, which describes the image for users who cannot see it.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
    | Condition: Cognigy Script | CognigyScript | Allows control over when buttons are shown or hidden based on conditions written in CognigyScript. If you're displaying buttons in a card and only want to show one when there's data, use this condition `{{context.array.length > 0}}`. It checks if the array has items (length is greater than 0). If true, the button will appear; if false, it stays hidden.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | Textual Description       | CognigyScript | The text that appears if the buttons aren't rendered in the chat.<br />This text can also serve as context, meaning it is included in the transcript or prompt, allowing AI Agents to provide more accurate responses. If you activate the **Include Rich Media Context** parameter in [Get Transcript](/ai/agents/develop/node-reference/service/get-transcript), [AI Agent](/ai/agents/develop/node-reference/ai/ai-agent), or [LLM Prompt](/ai/agents/develop/node-reference/service/llm-prompt) Nodes, the context will be taken only from the **Textual Description** parameter. If the **Textual Description** parameter is empty, the context will instead be taken from the button titles in Text with Quick Replies.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
  </Accordion>

  <Accordion title="Gallery">
    The **Gallery** output type consists of powerful visual widgets that are ideal for displaying a list of options with images. They are typically used to showcase a variety of products or other items that can be browsed.

    A gallery can be configured with multiple cards. Each card contains an image, a title, and a subtitle, and can be configured with optional buttons.

    **Card**

    | Parameters                | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | ------------------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Add Card                  | Button        | Adds a new card for the Gallery type. You can add multiple cards.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
    | Add Image                 | Button        | Adds an image for the selected card.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
    | Image URL                 | Button        | Appears when the Add Image parameter is selected. The URL for the image. Note that if you upload a URL to a storage service, such as Google Cloud Storage or AWS, the URL should be publicly accessible.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |
    | Image Alternate Text      | CognigyScript | Appears when the Add Image parameter is selected. The alternative text for accessibility, which describes the image for users who cannot see it.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
    | Title                     | CognigyScript | Appears when the Add Card parameter is selected. The title of the card.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
    | Subtitle                  | CognigyScript | Appears when the Add Card parameter is selected. The subtitle of the card.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | Condition: Cognigy Script | CognigyScript | Allows you to control when cards are shown or hidden based on conditions written in CognigyScript. If you're displaying cards in a Gallery and only want to show a card when there's data, use this condition: `{{context.array.length > 0}}`. It checks if the array has items (length is greater than 0). If true, the card will appear; if false, it stays hidden.                                                                                                                                                                                                                                                                                                                                                                                 |
    | Textual Description       | CognigyScript | The text that will appear if the cards aren't rendered in the chat. <br /><br />This text can also serve as context, meaning it is included in the transcript or prompt, allowing AI Agents to provide more accurate responses. If you activate the **Include Rich Media Context** parameter in [Get Transcript](/ai/agents/develop/node-reference/service/get-transcript), [AI Agent](/ai/agents/develop/node-reference/ai/ai-agent), or [LLM Prompt](/ai/agents/develop/node-reference/service/llm-prompt) Nodes, the context will be taken only from the **Textual Description** parameter. If the **Textual Description** parameter is empty, the context will instead be taken from the image alt text, button titles, and subtitles in Gallery. |

    **Button**

    | Parameters                | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
    | ------------------------- | ------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Add Button                | Button        | Adds a new Empty Title button. You can add multiple buttons.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
    | Empty Title               | Button        | Adds a new button to the selected card.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | Button Title              | Button        | Appears only when the Empty Title parameter is selected. The title of the button.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
    | Select Button Type        | Dropdown      | You can select the following options:<ul><li>**Postback Value** — sends a specified value to the start of the Flow, simulating user input as if the user had manually typed it. This behavior is typical for Quick Replies and is used to trigger specific responses or actions based on predefined values.</li><li>**URL** — opens the web page corresponding to the URL provided in the field.</li><li>**Phone Number** — opens the phone application on the device when the button is clicked, allowing users to initiate a call directly from the chat interface.</li><li>**Trigger Intent** — allows you to manually trigger an Intent by using the `cIntent:` prefix, followed by the name of the desired Intent. When this option is used, the regular Intent mapping is bypassed, and the specified Intent is activated directly. For more information, read [Trigger Intent](/ai/platform-features/nlu/intents/trigger-intent).</li><li>**Open xApp** — opens an xApp page. Note that, for the xApp session link to be available, you need to add the [xApp: Init Session](/ai/agents/develop/node-reference/xApp/init-xApp-session) and either the [xApp: Show HTML](/ai/agents/develop/node-reference/xApp/set-html-xApp-state) or [xApp: Show Adaptive Card Nodes](/ai/agents/develop/node-reference/xApp/set-AdaptiveCard-xApp-state) before the Say Node.</li></ul> |
    | Postback Value            | CognigyScript | Appears if the Post back value in Select Button Type is selected. The value that is sent to the start of the Flow, simulating user input as if the user had manually typed it.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
    | URL                       | CognigyScript | Appears if the URL value in Select Button Type is selected. The URL must start with `http://` or `https://`, be a valid, publicly accessible address, and contain no spaces or unsafe characters.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
    | URL Target                | CognigyScript | Appears if the URL value is selected in the Select Button Type. You can select one of the following options:<ul><li>**Open URL in a new tab** — the selected URL will open in a new browser tab.</li><li>**Open URL in the same tab** — the selected URL will replace the current page in the same browser tab.</li></ul>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         |
    | Phone number              | CognigyScript | Appears if the Phone number value in Select Button Type is selected. The number must be formatted as a valid phone number and shouldcontain spaces or special characters.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         |
    | Intent name               | CognigyScript | Appears if the Trigger Intent value in Select Button Type is selected. The Intent name that should be triggered.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
    | Condition: Cognigy Script | CognigyScript | Allows control over when buttons are shown or hidden based on conditions written in CognigyScript. If you're displaying buttons in a card and only want to show one when there's data, use this condition `{{context.array.length > 0}}`. It checks if the array has items (length is greater than 0). If true, the button will appear; if false, it stays hidden.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
  </Accordion>

  <Accordion title="Text with Buttons">
    In all Endpoints except Webchat v3, the **Text with Buttons** output type is similar to Text with Quick Replies but differs in presentation. Instead of quick replies, the **Text with Buttons** output type displays a vertical list with button options. The configuration for this output type is also similar to that of **Text with Quick Replies**. In Webchat v3, the [Text with Buttons](/webchat/v3/features#text-with-buttons) and [Text with Quick Replies](/webchat/v3/features#quick-replies) share the same presentation style.

    **Text**

    | Parameters          | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | ------------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Text                | CognigyScript | The message you want the AI Agent to send. If you want to customize the text by adding italic or bold styles or including links, note that Markdown isn't supported; only HTML is allowed.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |
    | Textual Description | CognigyScript | The text that appears if the buttons aren't rendered in the chat.<br />This text can also serve as context, meaning it is included in the transcript or prompt, allowing AI Agents to provide more accurate responses. If you activate the **Include Rich Media Context** parameter in [Get Transcript](/ai/agents/develop/node-reference/service/get-transcript), [AI Agent](/ai/agents/develop/node-reference/ai/ai-agent), or [LLM Prompt](/ai/agents/develop/node-reference/service/llm-prompt) Nodes, the context will be taken only from the **Textual Description** parameter. If the **Textual Description** parameter is empty, the context will instead be taken from the button titles in Text with Buttons. |

    **Button**

    | Parameters                | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
    | ------------------------- | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Add Button                | Button        | Adds a new Empty Title button. You can add up to 15 buttons.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
    | Empty Title               | Button        | Adds a new button to the selected card.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
    | Button Title              | Button        | Appears only when the Empty Title parameter is selected. The title of the button.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | Select Button Type        | Dropdown      | You can select the following options:<ul><li>**Postback Value** — sends a specified value to the start of the Flow, simulating user input as if the user had manually typed it. This behavior is typical for Quick Replies and is used to trigger specific responses or actions based on predefined values.</li><li>**Phone Number** — opens the phone application on the device when the button is clicked, allowing users to initiate a call directly from the chat interface.</li><li>**Trigger Intent** — allows you to manually trigger an Intent by using the `cIntent:` prefix, followed by the name of the desired Intent. When this option is used, the regular Intent mapping is bypassed, and the specified Intent is activated directly. For more information, read [Trigger Intent](/ai/platform-features/nlu/intents/trigger-intent).</li><li>**Open xApp** — opens an xApp page. Note that, for the xApp session link to be available, you need to add the [xApp: Init Session](/ai/agents/develop/node-reference/xApp/init-xApp-session) and either the [xApp: Show HTML](/ai/agents/develop/node-reference/xApp/set-html-xApp-state) or [xApp: Show Adaptive Card Nodes](/ai/agents/develop/node-reference/xApp/set-AdaptiveCard-xApp-state) before the Say Node.</li></ul> |
    | Postback Value            | CognigyScript | Appears if the Post back value in Select Button Type is selected. The value that is sent to the start of the Flow, simulating user input as if the user had manually typed it.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
    | URL                       | CognigyScript | Appears if the URL value in Select Button Type is selected. The URL must start with `http://` or `https://`, be a valid, publicly accessible address, and contain no spaces or unsafe characters.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | URL Target                | CognigyScript | Appears if the URL value is selected in the Select Button Type. You can select one of the following options:<ul><li>**Open URL in a new tab** — the selected URL will open in a new browser tab.</li><li>**Open URL in the same tab** — the selected URL will replace the current page in the same browser tab.</li></ul>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
    | Phone number              | CognigyScript | Appears if the Phone number value in Select Button Type is selected. The number must be formatted as a valid phone number and shouldcontain spaces or special characters.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
    | Intent name               | CognigyScript | Appears if the Trigger Intent value in Select Button Type is selected. The Intent name that should be triggered.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | Condition: Cognigy Script | CognigyScript | Allows control over when buttons are shown or hidden based on conditions written in CognigyScript. If you're displaying buttons in a card and only want to show one when there's data, use this condition `{{context.array.length > 0}}`. It checks if the array has items (length is greater than 0). If true, the button will appear; if false, it stays hidden.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
  </Accordion>

  <Accordion title="List">
    The **List** output type allows a customized list of items to be displayed, with many configuration options such as a header image, buttons, images, and more.

    The first list item can optionally be converted into a header that includes the list title, subtitle, and button. Each additional list item can have a title, subtitle, image, and button. A button can also be added at the bottom of the list.

    **Item**

    | Parameters           | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
    | -------------------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Add item             | Button        | Adds a new Empty Title button to the selected card. You can add up to 10 items.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
    | Empty Title          | Button        | Adds a new entry to the list.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
    | Title                | CognigyScript | Appears when you select **Empty Title**.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | Subtitle             | CognigyScript | Appears when the Add Card parameter is selected. The subtitle of the card.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | Image URL            | Button        | Appears when the Add Image parameter is selected. The URL for the image. Note that if you upload a URL to a storage service, such as Google Cloud Storage or AWS, the URL should be publicly accessible.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | Image Alternate Text | CognigyScript | Appears when the Add Image parameter is selected. The alternative text for accessibility, which describes the image for users who cannot see it.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
    | Default Action URL   | CognigyScript | Appears when the Empty Title parameter is selected. It is a link that opens when the user clicks anywhere on the item. It's used when you want the entire item to be clickable, directing the user to a specific website or resource.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
    | Condition            | CognigyScript | Allows you to control when items are shown or hidden based on conditions written in CognigyScript. If you're displaying items in a list and only want to show a list when there's data, use this condition: `{{context.array.length > 0}}`. It checks if the array has items (length is greater than 0). If true, the item will appear; if false, it stays hidden.                                                                                                                                                                                                                                                                                                                                                                                   |
    | Textual Description  | CognigyScript | The text that appears if the item buttons aren't rendered in the chat.<br /><br />This text can also serve as context, meaning it is included in the transcript or prompt, allowing AI Agents to provide more accurate responses. If you activate the **Include Rich Media Context** parameter in [Get Transcript](/ai/agents/develop/node-reference/service/get-transcript), [AI Agent](/ai/agents/develop/node-reference/ai/ai-agent), or [LLM Prompt](/ai/agents/develop/node-reference/service/llm-prompt) Nodes, the context will be taken only from the **Textual Description** parameter. If the **Textual Description** parameter is empty, the context will instead be taken from the image alt text, button titles, and subtitles in List. |

    **Item Button**

    Each item can have a separate button.
    To configure it, hover the mouse over the Empty Title and click **+** icon.
    You can add only one button per item.

    | Parameters                | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
    | ------------------------- | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Button Title              | Button        | Appears only when the Empty Title parameter is selected. The title of the button.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | Select Button Type        | Dropdown      | You can select the following options:<ul><li>**Postback Value** — sends a specified value to the start of the Flow, simulating user input as if the user had manually typed it. This behavior is typical for Quick Replies and is used to trigger specific responses or actions based on predefined values.</li><li>**Phone Number** — opens the phone application on the device when the button is clicked, allowing users to initiate a call directly from the chat interface.</li><li>**Trigger Intent** — allows you to manually trigger an Intent by using the `cIntent:` prefix, followed by the name of the desired Intent. When this option is used, the regular Intent mapping is bypassed, and the specified Intent is activated directly. For more information, read [Trigger Intent](/ai/platform-features/nlu/intents/trigger-intent).</li><li>**Open xApp** — opens an xApp page. Note that, for the xApp session link to be available, you need to add the [xApp: Init Session](/ai/agents/develop/node-reference/xApp/init-xApp-session) and either the [xApp: Show HTML](/ai/agents/develop/node-reference/xApp/set-html-xApp-state) or [xApp: Show Adaptive Card Nodes](/ai/agents/develop/node-reference/xApp/set-AdaptiveCard-xApp-state) before the Say Node.</li></ul> |
    | Postback Value            | CognigyScript | Appears if the Post back value in Select Button Type is selected. The value that is sent to the start of the Flow, simulating user input as if the user had manually typed it.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
    | URL                       | CognigyScript | Appears if the URL value in Select Button Type is selected. The URL must start with `http://` or `https://`, be a valid, publicly accessible address, and contain no spaces or unsafe characters.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | URL Target                | CognigyScript | Appears if the URL value is selected in the Select Button Type. You can select one of the following options:<ul><li>**Open URL in a new tab** — the selected URL will open in a new browser tab.</li><li>**Open URL in the same tab** — the selected URL will replace the current page in the same browser tab.</li></ul>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
    | Phone number              | CognigyScript | Appears if the Phone number value in Select Button Type is selected. The number must be formatted as a valid phone number and shouldn't contain spaces or special characters.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
    | Intent name               | CognigyScript | Appears if the Trigger Intent value in Select Button Type is selected. The Intent name that should be triggered.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | Condition: Cognigy Script | CognigyScript | Allows control over when buttons are shown or hidden based on conditions written in CognigyScript. If you're displaying buttons in a card and only want to show one when there's data, use this condition `{{context.array.length > 0}}`. It checks if the array has items (length is greater than 0). If true, the button will appear; if false, it stays hidden.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |

    **Button**

    You can add only one button at the bottom of the list.

    | Parameters                | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
    | ------------------------- | ------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Add Button                | Button        | Adds a new Empty Title button.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
    | Empty Title               | Button        | Opens a configuration menu for the button. Once you configure **Button Title**, the button is added to the bottom of the list. You can also configure **Select Button Type** and **Condition: Cognigy Script**.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
    | Button Title              | Button        | Appears only when the Empty Title parameter is selected. The title of the button.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
    | Select Button Type        | Dropdown      | You can select the following options:<ul><li>**Postback Value** — sends a specified value to the start of the Flow, simulating user input as if the user had manually typed it. This behavior is typical for Quick Replies and is used to trigger specific responses or actions based on predefined values.</li><li>**URL** — opens the web page corresponding to the URL provided in the field.</li><li>**Phone Number** — opens the phone application on the device when the button is clicked, allowing users to initiate a call directly from the chat interface.</li><li>**Trigger Intent** — allows you to manually trigger an Intent by using the `cIntent:` prefix, followed by the name of the desired Intent. When this option is used, the regular Intent mapping is bypassed, and the specified Intent is activated directly. For more information, read [Trigger Intent](/ai/platform-features/nlu/intents/trigger-intent).</li><li>**Open xApp** — opens an xApp page. Note that, for the xApp session link to be available, you need to add the [xApp: Init Session](/ai/agents/develop/node-reference/xApp/init-xApp-session) and either the [xApp: Show HTML](/ai/agents/develop/node-reference/xApp/set-html-xApp-state) or [xApp: Show Adaptive Card Nodes](/ai/agents/develop/node-reference/xApp/set-AdaptiveCard-xApp-state) before the Say Node.</li></ul> |
    | Postback Value            | CognigyScript | Appears if the Post back value in Select Button Type is selected. The value that is sent to the start of the Flow, simulating user input as if the user had manually typed it.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
    | URL                       | CognigyScript | Appears if the URL value in Select Button Type is selected. The URL must start with `http://` or `https://`, be a valid, publicly accessible address, and contain no spaces or unsafe characters.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
    | URL Target                | CognigyScript | Appears if the URL value is selected in the Select Button Type. You can select one of the following options: <ul><li>**Open URL in a new tab** — the selected URL will open in a new browser tab.</li><li>**Open URL in the same tab** — the selected URL will replace the current page in the same browser tab.</li></ul>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
    | Phone number              | CognigyScript | Appears if the Phone number value in Select Button Type is selected. The number must be formatted as a valid phone number and shouldn't contain spaces or special characters.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
    | Intent name               | CognigyScript | Appears if the Trigger Intent value in Select Button Type is selected. The Intent name that should be triggered.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
    | Condition: Cognigy Script | CognigyScript | Allows control over when buttons are shown or hidden based on conditions written in CognigyScript. If you're displaying buttons in a card and only want to show one when there's data, use this condition `{{context.array.length > 0}}`. It checks if the array has items (length is greater than 0). If true, the button will appear; if false, it stays hidden.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
  </Accordion>

  <Accordion title="Audio">
    The **Audio** output type plays audio when supported by the channel. It is configured by providing a URL to the audio file and includes controls to play, pause, and stop the audio.

    | Parameters           | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
    | -------------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Audio URL            | CognigyScript | The URL of the track you want to play. Note that the URL should be publicly accessible.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | Audio Alternate Text | CognigyScript | The audio transcript you can add alongside the audio, allowing users to download it. Only applicable to Webchat v3.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
    | Textual Description  | CognigyScript | The text that appears if the audio content isn't rendered in the chat.<br /><br />This text can also serve as context, meaning it is included in the transcript or prompt, allowing AI Agents to provide more accurate responses. If you activate the **Include Rich Media Context** parameter in [Get Transcript](/ai/agents/develop/node-reference/service/get-transcript), [AI Agent](/ai/agents/develop/node-reference/ai/ai-agent), or [LLM Prompt](/ai/agents/develop/node-reference/service/llm-prompt) Nodes, the context will be taken only from the **Textual Description** parameter. If the **Textual Description** parameter is empty, the context will instead be taken from the alt text in Audio. |
  </Accordion>

  <Accordion title="Image">
    The **Image** output type displays an image.

    | Parameters           | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
    | -------------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Image URL            | Button        | Appears when the Add Image parameter is selected. The URL for the image. Note that if you upload a URL to a storage service, such as Google Cloud Storage or AWS, the URL should be publicly accessible.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
    | Image Alternate Text | CognigyScript | Appears when the Add Image parameter is selected. The alternative text for accessibility, which describes the image for users who cannot see it.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
    | Textual Description  | CognigyScript | The text that appears if the image content isn't rendered in the chat.<br /><br />This text can also serve as context, meaning it is included in the transcript or prompt, allowing AI Agents to provide more accurate responses. If you activate the **Include Rich Media Context** parameter in [Get Transcript](/ai/agents/develop/node-reference/service/get-transcript), [AI Agent](/ai/agents/develop/node-reference/ai/ai-agent), or [LLM Prompt](/ai/agents/develop/node-reference/service/llm-prompt) Nodes, the context will be taken only from the **Textual Description** parameter. If the **Textual Description** parameter is empty, the context will instead be taken from the alt text in Image. |
  </Accordion>

  <Accordion title="Video">
    The **Video** output type allows you to add a video clip.
    It takes a URL as an input parameter and automatically starts playing the video if supported by the Endpoint.

    | Parameters           | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
    | -------------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Video URL            | Button        | The URL of the track you want to play. Note that the URL should be publicly accessible.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | Video Alternate Text | CognigyScript | The video transcript you can add alongside the audio to allow users to download it. This field is applicable only to Webchat v3.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
    | Video Captions URL   | CognigyScript | A link to a file containing captions for the video. This URL provides text-based descriptions of the audio content, which helps with accessibility and understanding the video. The captions should be in `.vtt` format.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
    | Textual Description  | CognigyScript | The text that appears if the video content isn't rendered in the chat.<br /><br />This text can also serve as context, meaning it is included in the transcript or prompt, allowing AI Agents to provide more accurate responses. If you activate the **Include Rich Media Context** parameter in [Get Transcript](/ai/agents/develop/node-reference/service/get-transcript), [AI Agent](/ai/agents/develop/node-reference/ai/ai-agent), or [LLM Prompt](/ai/agents/develop/node-reference/service/llm-prompt) Nodes, the context will be taken only from the **Textual Description** parameter. If the **Textual Description** parameter is empty, the context will instead be taken from the alt text in Video. |
  </Accordion>

  <Accordion title="Adaptive Card">
    The **Adaptive Card** output type allows you
    to add [Microsoft Adaptive Cards](https://learn.microsoft.com/en-us/adaptive-cards/).
    They offer customization options, support for rich media (images, video, and audio),
    ease of use with a simple JSON schema,
    and the ability to create dynamic content for users to match their specific needs and branding.

    | Parameters             | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
    | ---------------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Adaptive Card Designer | JSON          | The JSON schema that defines the structure and content of the adaptive card. You can create an adaptive card using the [Adaptive Card Designer](https://adaptivecards.io/designer/) and then copy and paste the JSON code into the editor. Customize the existing JSON as needed, and if the JSON is valid, the adaptive card will be rendered below the editor.<br /><br />Cognigy supports a limited number of Adaptive Card versions. Using the latest versions may cause compatibility issues. Refer to [supported versions](/ai/agents/deploy/endpoints/content-conversion) for best results.                                                                                                                            |
    | Textual Description    | CognigyScript | The text that appears if the adaptive card content isn't rendered in the chat.<br /><br />This text can also serve as context, meaning it is included in the transcript or prompt, allowing AI Agents to provide more accurate responses. If you activate the **Include Rich Media Context** parameter in [Get Transcript](/ai/agents/develop/node-reference/service/get-transcript), [AI Agent](/ai/agents/develop/node-reference/ai/ai-agent), or [LLM Prompt](/ai/agents/develop/node-reference/service/llm-prompt) Nodes, the context will be taken only from the **Textual Description** parameter. If the **Textual Description** parameter is empty, the context will instead be taken from the JSON in Adaptive Card. |

    <Accordion title="Adaptive Card JSON example">
      ```json theme={null}
      {
        "$schema": "http://adaptivecards.io/schemas/adaptive-card.json",
        "type": "AdaptiveCard",
        "version": "1.0",
        "body": [
          {
            "type": "TextBlock",
            "size": "Medium",
            "weight": "Bolder",
            "text": "Publish Adaptive Card Schema"
          },
          {
            "type": "ColumnSet",
            "columns": [
              {
                "type": "Column",
                "items": [
                  {
                    "type": "Image",
                    "style": "Person",
                    "url": "https://pbs.twimg.com/profile_images/3647943215/d7f12830b3c17a5a9e4afcc370e3a37e_400x400.jpeg",
                    "size": "Small"
                  }
                ],
                "width": "auto"
              },
              {
                "type": "Column",
                "items": [
                  {
                    "type": "TextBlock",
                    "weight": "Bolder",
                    "text": "Matt Hidinger",
                    "wrap": true
                  },
                  {
                    "type": "TextBlock",
                    "spacing": "None",
                    "text": "Created Tue, Feb 14, 2017",
                    "isSubtle": true,
                    "wrap": true
                  }
                ],
                "width": "stretch"
              }
            ]
          },
          {
            "type": "TextBlock",
            "text": "Publish Adaptive Card Schema easily.",
            "wrap": true
          },
          {
            "type": "Action.OpenUrl",
            "title": "View",
            "url": "https://adaptivecards.io"
          }
        ]
      }
      ```
    </Accordion>

    **Create an Adaptive Card with Generative AI**

    You can also use Generative AI to create a new adaptive card or improve an existing one. Before using it, ensure that you are connected to one of the [LLM Providers](/ai/agents/develop/gen-ai-and-llms/generative-ai#prerequisites).

    To use this feature, follow these steps:

    1. Select the **Adaptive Card** output type.
    2. On the right side of the **Output type** list, click <img alt="generate with AI" />.

    1) In the **Generate Node Output** section, instruct the Generative AI model how to improve the current adaptive card. For example, `Create a form with a customer name field and a date input field`.
    2) Click **Generate**. The adaptive card will be generated.
    3) Iteratively improve the resulting adaptive card by giving further instructions in the **Generate Node Output** section. For example, `Add a flight number field`.
    4) Click **Generate**. The existing adaptive card will be updated.

    To navigate between your inputs, use <img alt="previous input" /> and <img alt="next input" />.

    To replace the current Adaptive Card with a new one, click <img alt="remove adaptive card" />.

    <Frame>
      <img alt="Say Node Adaptive Card output type with Generative AI prompt" />
    </Frame>

    <Accordion title="Generative AI Adaptive Card JSON example">
      ```json theme={null}
      {
        "$schema": "http://adaptivecards.io/schemas/adaptive-card.json",
        "type": "AdaptiveCard",
        "version": "1.0",
        "body": [
          {
            "type": "TextBlock",
            "text": "Customer Form"
          },
          {
            "type": "Input.Text",
            "id": "customerName",
            "placeholder": "Enter customer name"
          },
          {
            "type": "Input.Date",
            "id": "dateInput",
            "placeholder": "Enter date"
          },
          {
            "type": "Input.Text",
            "id": "flightNumber",
            "placeholder": "Enter flight number"
          }
        ]
      }
      ```
    </Accordion>
  </Accordion>
</AccordionGroup>

### AI-Enhanced Output

To use AI-enhanced output rephrasing, read the [Generative AI](/ai/agents/develop/gen-ai-and-llms/generative-ai) article.

### Handover Settings

When using a handover to a contact center, you can choose who receives the message from the AI Agent:

* **User and Agent** — by default, both the end user and the human agent will receive the message.
* **User only** — the end user will receive the message.
* **Agent only** — the responsible human agent will receive the message.

### Advanced

| Parameters              | Type   | Description                                                                                                                                                                                                                                                                                                                                                                                                             |
| ----------------------- | ------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Exclude from Transcript | Toggle | Excludes the Node output from the conversation transcript. This parameter is useful when confidentiality is necessary, such as preventing unnecessary data from being sent to the LLM provider. Also, you can use this parameter to send messages that shouldn't be interpreted by the AI Agent, including legal disclaimers, sensitive information, or other specific instructions irrelevant to the ongoing dialogue. |

## More Information

* [Overview](/ai/agents/develop/node-reference/overview)
* [Channels](/ai/agents/develop/nodes/channels)


# CXone
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/cxone/overview

CXone Nodes integrate Flows with NiCE CXone capabilities.

<CardGroup>
  <Card title="Retrieve Knowledge" href="/ai/agents/develop/node-reference/cxone/retrieve-knowledge">
    The Retrieve Knowledge Node provides information from NiCE CXone Knowledge Hub to your AI Agents.
  </Card>

  <Card title="Send Status to CXone Copilot" href="/ai/agents/develop/node-reference/cxone/send-copilot-status">
    The Send Status to CXone Copilot Node sends workflow completion status to NiCE CXone Copilot.
  </Card>
</CardGroup>


# Retrieve Knowledge
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/cxone/retrieve-knowledge



<a href="/release-notes/2026.8"><Badge>Added in 2026.8 (beta)</Badge></a>

<Frame>
  <img alt="CXone Retrieve Knowledge Node" />
</Frame>

## Description

The Retrieve Knowledge Node extracts information from knowledge databases in the [NiCE CXone Knowledge Hub](https://help.nicecxone.com/content/knowledgehub/knowledgehub.htm) and provides this information to your AI Agents. For details, refer to the [NiCE CXone documentation](https://help.nicecxone.com/content/cognigyai/nativenodes/addretrieveknowledgenodeincognigyai.htm).

## Restrictions

* The Retrieve Knowledge Node is supported only in Cognigy installations that are integrated with NiCE CXone.

## More information

* [NiCE CXone Knowledge Hub](https://help.nicecxone.com/content/knowledgehub/knowledgehub.htm)


# Send Status to CXone Copilot Node
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/cxone/send-copilot-status



<a href="/release-notes/2026.9"><Badge>Added in 2026.9 (beta)</Badge></a>

<Frame>
  <img alt="Send Status to CXone Copilot Node" />
</Frame>

## Description

The Send Status to CXone Copilot Node allows you to send a completion status to [NiCE CXone Copilot for Agents](https://help.nicecxone.com/content/agent/cxoneagent/enlightencopilotforagentscxa.htm). This Node allows you to provide human agents using the NiCE CXone Agent Workspace with information about the Flow status, indicating whether it succeeded or failed. For details, refer to the [NiCE CXone documentation](https://help.nicecxone.com/content/cognigyai/nativenodes/sendnode.htm#Add).

## Restrictions

* This Node is supported only in Cognigy installations that are integrated with NiCE CXone.

## More information

* [NiCE CXone Copilot for Agents](https://help.nicecxone.com/content/agent/cxoneagent/enlightencopilotforagentscxa.htm)


# Datepicker
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/data/datepicker



<Frame>
  <img alt="Datepicker Node configuration panel" />
</Frame>

## Description

The Datepicker Node renders a datepicker to the user.

<Accordion title="Automatically Triggered by Question Nodes">
  The Datepicker is automatically triggered whenever a Question Node with the Date type is used. This means that you rarely need to add the Datepicker Node.
</Accordion>

## Parameters

| Parameter        | Description                                                                                                                                                                                                                      |
| ---------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Datepicker Title | This is the title that will be shown in the header of the datepicker widget.                                                                                                                                                     |
| Locale           | A dropdown menu that lets you select a locale for the Datepicker. This automatically outputs the date format in the localized language.                                                                                          |
| Mode             | Mode determines the behavior of the Datepicker. <ul><li>**Single** – a single date can be selected.</li><li>**Multiple** – lets the user select multiple dates.</li><li>**Range** – lets the user select a date range.</li></ul> |

<AccordionGroup>
  <Accordion title="Button Texts">
    | Parameter                   | Description                                                                                                                            |
    | --------------------------- | -------------------------------------------------------------------------------------------------------------------------------------- |
    | Open Datepicker Button Text | This is the text that is shown in the button that is rendered when the Datepicker is triggered. By default, is says: **pick date**.    |
    | Cancel Button Label         | This is the text that is shown on the "cancel" button that is rendered when the Datepicker has been opened. The default is **cancel**. |
    | Submit Button Label         | This is the text that is shown on the "submit" button that is rendered when the Datepicker has been opened. The default is **submit**. |
  </Accordion>

  <Accordion title="Date Settings">
    The Datepicker allows you to configure precisely which dates can be selected.

    | Parameter    | Description                                                                              |
    | ------------ | ---------------------------------------------------------------------------------------- |
    | Default Date | This is the default selected date. If not filled out, it will be the current date.       |
    | Minimum Date | This is the minimum date. If not filled out, it will not have a minimum date.            |
    | Maximum Date | This is the maximum selectable date. If not filled out, it will not have a maximum date. |
  </Accordion>

  <Accordion title="Time settings">
    The Datepicker allows you to configure whether times can be selected. If so, it also allows you to configure the default times in detail.

    | Parameter                | Description                                                                                               |
    | ------------------------ | --------------------------------------------------------------------------------------------------------- |
    | Allow Date & Time        | Determines whether the Datepicker should additionally show selectable time.                               |
    | Use 24-Hour Format       | If selected, the Datepicker will display times in 24-hour format (for example, 20:00h instead of 8:00PM). |
    | Default Hour to show     | Configurable default hour.                                                                                |
    | Default minute to show   | Configurable default minute.                                                                              |
    | Step to increase hours   | The increment step size for hours. The default is 1 hour.                                                 |
    | Step to increase minutes | The incremental increase in minutes, when the minute selection is used. The default is 5 minutes.         |
    | Show seconds in picker   | Enable this setting to show seconds in the Datepicker options.                                            |
  </Accordion>

  <Accordion title="Enable/Disable Dates">
    The Datepicker can be configured to exclude or include specific dates.
    This setting is useful when certain dates should be unavailable, such as for a booking or reservation.
  </Accordion>
</AccordionGroup>

#### Enable specific dates

This setting allows you to enter a range of dates that should be enabled.
If configured, all other dates will automatically be disabled.
In addition, you can define a function that is used to enable dates, for example, only weekdays.

Example of enabling dates by function:

```ts theme={null}
(date: Date): boolean => {
    /* The function takes in a Date object, and should return a boolean value.
    * If the function returns true, the date will be enabled.
    * Sunday = 0, Monday = 1, Tuesday = 2, Wednesday = 3, Thursday = 4, Friday = 5, Saturday = 6	 
    */ 	
    // This will enable Monday to Friday
    return (date.getDay() > 0 && date.getDay() < 6);
};
```

#### Disable specific dates

This setting allows you to enter a range of dates that should be disabled.
If configured, all other dates will automatically be enabled.
In addition, you can define a function that is used to disable dates, for example, every Saturday and Sunday.

Disable dates by function - example:

```ts theme={null}
(date: Date): boolean => {
    /* The function takes in a Date object, and should return a boolean value.
    * If the function returns true, the date will be disabled.
    * Sunday = 0, Monday = 1, Tuesday = 2, Wednesday = 3, Thursday = 4, Friday = 5, Saturday = 6	 
    */ 	
    // This will disable every Sunday and Saturday
    return (date.getDay() === 0 || date.getDay() === 6);
};
```

<Accordion title="Advanced Options">
  The Datepicker comes with four advanced options:

  | Parameter                   | Description                                                            |
  | --------------------------- | ---------------------------------------------------------------------- |
  | Date Format                 | This field can be used to output a specific date format, like `Y-m-d`. |
  | Hide Calendar               | When enabled, the Datepicker will not display a calendar.              |
  | Show week numbers           | When enabled, the week numbers are shown in the calendar.              |
  | Hide Datepicker Quick Reply | This hides the quick reply in the Datepicker.                          |
</Accordion>


# Email Notification
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/data/email-notification



<a href="/release-notes/2026.1"><Badge>Updated in 2026.1</Badge></a>

<Frame>
  <img alt="Email Notification Node configuration panel" />
</Frame>

## Description

The Email Notification Node allows you to send quick and easy emails during sessions or for in-house communications without requiring any provider settings to be set up in the UI. Instead, you can configure the necessary provider settings in the environment variables and secrets. Only one provider can be set up at a time.

## Restrictions

* Your SMTP Server settings need to be adjusted to use the internal Cognigy Email notification.
* Email attachments are restricted (for example, files, images). If required, use [Send SMTP Email Node](/ai/agents/develop/node-reference/data/send-smtp-email).

## Authentication

This Node supports both Basic Auth and OAuth2 authentication methods.

If you have an on-premises installation, you need to add the following parameters to the `values.yaml` file:

<Tabs>
  <Tab title="OAuth2 (recommended)">
    object:

    ```yaml theme={null}
    smtpEmailNotificationCredentials:
     authType: "oauth2"
     oauth2:
       user: "<smtp-server>@domain.com"
       clientId: "your client id"
       clientSecret: "your client secret"
       refreshToken: "your refresh token"
       accessToken: "your access token"
    ```

    variables:

    ```yaml theme={null}
    EMAIL_NOTIFICATION_SMTP_FROM: "<smtp-server>@domain.com"
    EMAIL_NOTIFICATION_SMTP_AUTH_TYPE: "oauth2"
    EMAIL_NOTIFICATION_SMTP_EMAIL_SERVICE_TYPE: "<email service, for example, Outlook365>"
    EMAIL_NOTIFICATION_SMTP_OAUTH2_TOKEN_ACCESS_URL: "<provider token access url, for example, for Microsoft https://login.microsoftonline.com/common/oauth2/v2.0/token, for Google https://accounts.google.com/o/oauth2/token>"
    EMAIL_NOTIFICATION_SMTP_OAUTH2_EXPIRY_DATE_UNIX_TIMESTAMP: "<unix timestamp for the current accessToken, for example, 1681382246000>"
    EMAIL_NOTIFICATION_SMTP_TLS_REQUIRED: "<tls method, for example, 'tls' for default TLS on port 443, 'starttls' to upgrade an insecure connection on the same port, or 'none' to connect without TLS>"
    ```

    Cognigy caches connections by default, and for OAuth2 connections, Cognigy also caches the refresh and access tokens. To update the refresh token or access token, create a [Snapshot](/ai/agents/deploy/snapshots).
  </Tab>

  <Tab title="Basic Auth">
    object:

    ```yaml theme={null}
    smtpEmailNotificationCredentials:
      authType: "basic"
      basic:
        username: "<your email client's username>"
        password: "<your email clientl's password>"
    ```

    variables:

    ```yaml theme={null}
    EMAIL_NOTIFICATION_SMTP_AUTH_TYPE: "basic"
    EMAIL_NOTIFICATION_SMTP_HOST: "<SMTP_HOST>"
    EMAIL_NOTIFICATION_SMTP_PORT: "<SMTP_PORT>"
    EMAIL_NOTIFICATION_SMTP_FROM: "<smtp-server>@domain.com"
    EMAIL_NOTIFICATION_SMTP_TLS_REQUIRED: "<tls method, for example, 'tls' for default TLS on port 443, 'starttls' to upgrade an insecure connection on the same port, or 'none' to connect without TLS>"
    ```
  </Tab>
</Tabs>

## Parameters

| Parameter         | Description                                                           |
| ----------------- | --------------------------------------------------------------------- |
| Email Sender Name | The name displayed as the sender of the email.                        |
| Recipient         | Comma-separated list of email addresses to send the email to.         |
| Subject           | The subject of the email.                                             |
| Content           | The content of the email. The field supports basic rich text editing. |

<Accordion title="Store location of SMTP Request result">
  | Option             | Description                                                                                |
  | ------------------ | ------------------------------------------------------------------------------------------ |
  | Don't store result | Default option. The result isn't stored.                                                   |
  | Input              | The result is stored at a specified location in the **Input Key to store Result** field.   |
  | Context            | The result is stored at a specified location in the **Context Key to store Result** field. |
</Accordion>

#### Stop on Error

If switched on, the Flow execution will halt in case the mail-sending fails.

<Accordion title="Advanced">
  | Parameter                    | Description                                                                                                                                                          |
  | ---------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
  | CC Email Addresses           | The CC recipients of your email.                                                                                                                                     |
  | BCC Email Addresses          | The BCC recipients of your email.                                                                                                                                    |
  | Email Priority               | The priority of the email.                                                                                                                                           |
  | Execute Request asynchronous | If switched on, the Flow execution continues directly after the Send Email Node has been triggered. If turned-off, the Flow execution waits until the email is sent. |
</Accordion>


# Data
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/data/overview



Data Nodes are designed to manage, manipulate, process, or store data within a system.

<CardGroup>
  <Card title="Datepicker" href="/ai/agents/develop/node-reference/data/datepicker">
    The Datepicker Node renders a date selection widget in certain channels.
  </Card>

  <Card title="Send SMTP Email" href="/ai/agents/develop/node-reference/data/send-smtp-email">
    The Send SMTP Email Node sends an email directly from the Flow.
  </Card>

  <Card title="Email Notification" href="/ai/agents/develop/node-reference/data/email-notification">
    The Email Notification Node sends quick and easy emails during sessions or for in-house communications.
  </Card>
</CardGroup>


# Send SMTP Email
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/data/send-smtp-email



<a href="/release-notes/2026.3"><Badge>Updated in 2026.3</Badge></a>

<Frame>
  <img alt="Send SMTP Email Node configuration panel" />
</Frame>

## Description

<Warning>
  As of Cognigy 2026.4.0, the OAuth2 authentication method for the Send SMTP Email Node is deprecated. The full removal of this authentication method is planned for Cognigy 2026.16.0. To avoid disruptions in AI Agents using OAuth2 in this Node, migrate to the [OAuth2 Client Credentials](/ai/agents/develop/node-reference/data/send-smtp-email#oauth2-client-credentials-recommended) authentication method before the removal date.
</Warning>

Use the Send SMTP Email Node to send an email via an SMTP connection to a recipient.

## Parameters

<AccordionGroup>
  <Accordion title="Email Service Configuration">
    Some email providers, such as Gmail, require you to configure your mail account to allow sending mail via SMTP before you can use it.

    | Parameter                    | Description                                                                                                                                                                                                         |
    | ---------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Email Service                | Select the Email Provider of the address you want to send from. If yours is not listed, you can select **Other SMTP Service** to do a manual configuration.                                                         |
    | SMTP Settings                | Depending on which Email Service selected above, you need to either enter your login credentials for that service or for the full SMTP connection information in case you selected the "Other SMTP Service" option. |
    | Recipient TO Email Addresses | Comma-separated list of email addresses to send the email to.                                                                                                                                                       |
    | Email Subject                | The subject of the email.                                                                                                                                                                                           |
    | Email Content                | The content of the email. The field supports basic rich text editing.                                                                                                                                               |
  </Accordion>

  <Accordion title="Authentication">
    You can use one of the following authentication methods:

    <Tabs>
      <Tab title="OAuth2 Client Credentials (recommended)">
        OAuth2 Client Credentials is a token-based authentication method. It uses a client ID and client secret to obtain an access token and authenticate with the email provider. This authentication method provides a higher level of security than Basic Auth and doesn't require refresh tokens.

        To create a connection, follow these steps:

        1. Next to the **OAuth2 Client Credentials Parameters** setting, click **+**.
        2. Add a new connection or use an existing one:
           * Connection name — enter a unique connection name.
           * Token Endpoint URL — enter the email provider URL to retrieve the access token.
           * Client ID — enter the registered client ID of the application.
           * Client Secret — enter the registered client secret of the application.
           * *(Optional)* Scope — enter the scope of the access token, as defined in [RFC 6749](https://datatracker.ietf.org/doc/html/rfc6749#section-4.4.2). Some email providers require the scope to be set. For example, use `https://outlook.office365.com/.default` if you've selected **Office365** for the **Email Service** parameter. Refer to your email provider's documentation to find the correct scope.
           * User — enter the email address from which the email is sent.
        3. Click **Create**.
      </Tab>

      <Tab title="OAuth2 (deprecated)">
        OAuth2 is a token-based authentication method. It uses access tokens generated by the service provider to grant authorization for client applications to access user data. Access tokens are unique identifiers that contain user, client application, and permission information. They are short-lived and can be revoked if necessary, providing a higher level of security than Basic Auth. Email providers such as Google and Microsoft may use only OAuth2 for authentication.

        To create a connection, follow these steps:

        1. Next to the **OAuth2 Parameters** setting, click **+**.
        2. Add a new connection or use an existing one:
           * Connection name — enter a unique connection name.
           * user — add your email address.
           * clientId — add a registered client id of the application.
           * clientSecret — add a registered client secret of the application.
           * accessToken — add an access token. Required only if refreshToken is not available.
           * refreshToken — add a refresh token. If it is provided, the Node tries to generate a new access token if existing one expires or fails.
           * expiryDate — add an expiration Unix timestamp in milliseconds for the current `accessToken`.
           * accessUrl — add an HTTP endpoint for requesting new access tokens. This value defaults to Gmail `https://accounts.google.com/o/oauth2/token`.
           * tlsOption — specify one of the options if you want to connect with TLS:
             * tls — is used by default. TLS is a separate protocol that operates on a different port (usually port 443 for HTTPS).
             * starttls — is a command used to upgrade an existing insecure connection to a secure one within the same port.
             * none — don't connect with TLS.
        3. Click **Create**.

        Cognigy caches connections by default, and for OAuth2 connections, Cognigy also caches the refresh and access tokens. To update the refresh token or access token, create a [snapshot](/ai/agents/deploy/snapshots).
      </Tab>

      <Tab title="Basic Auth">
        Basic Auth is a simple authentication scheme where the client sends the user's credentials (username and password) to the server in the HTTP header. It's used for web applications requiring authentication and adds an Authorization header to the HTTP request. However, it has security limitations as the credentials are sent in clear text, and the server doesn't provide a way to log out.

        To create a connection, follow these steps:

        1. Next to the **Basic Auth Parameters** setting, click **+**.
        2. Add a new connection or use an existing one:
           * Connection name — enter a unique connection name.
           * username — add a login that you obtained from your email provider.
           * password — add a password that you obtained from your email provider.
        3. Click **Create**.
      </Tab>
    </Tabs>
  </Accordion>

  <Accordion title="Attachments">
    By default, the **No Attachment** option is selected.

    | Attachment Type             | Description                                                                                                                                                    |
    | --------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | From URL                    | You can specify a public Web URL or a data URL. The linked File will then be attached to that mail. You can also pick a custom filename for the attached file. |
    | From Text Content           | You can define a text that should be sent as an attachment.                                                                                                    |
    | From Base64 String          | You can define Base64-encoded data that should be attached.                                                                                                    |
    | Custom Content Attachment   | If you need to send text content with a specific MIME-Type, allowing you to define an own MIME-Type.                                                           |
    | From Pregenerated MIME Node | Allows you to send a fully customized attachment body manually.                                                                                                |
  </Accordion>

  <Accordion title="Result Storage & Error Handling">
    | Option             | Description                                                                                |
    | ------------------ | ------------------------------------------------------------------------------------------ |
    | Don't store result | Default option. The result isn't stored.                                                   |
    | Input              | The result is stored at a specified location in the **Input Key to store Result** field.   |
    | Context            | The result is stored at a specified location in the **Context Key to store Result** field. |
  </Accordion>
</AccordionGroup>

#### Stop on Error

If switched on, the flow execution will halt in case the mail-sending fails.

<Accordion title="Advanced">
  | Parameter                        | Description                                                                                                                                                                 |
  | -------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
  | Custom From Email Address        | This email address will be displayed in the "from" field (sender's address).                                                                                                |
  | CC Email Addresses               | The CC recipients of your email.                                                                                                                                            |
  | BCC Email Addresses              | The BCC recipients of your email.                                                                                                                                           |
  | Reply To Address                 | This email address is the target for receiving mails when recipients reply to your email.                                                                                   |
  | Define custom text Email Content | If your content contains non-text content, you can provide a customized text-only version of the mail content here for improved compatibility with limited devices.         |
  | Email Priority                   | You can use this option to set your priority to `high`, `normal` (default), or `low`.                                                                                       |
  | Execute Request asynchronous     | If switched on, the flow execution will **continue directly after** the Send Email Node was triggered. If turned-off, flow execution will **wait until the email is sent**. |
</Accordion>

## Example

This example shows how to configure OAuth2 Client Credentials with Microsoft 365 SMTP using Microsoft Entra ID.

<AccordionGroup>
  <Accordion title="Prerequisites">
    Before starting, ensure the following:

    * Your Microsoft Entra account must have one of the following roles:
      * Application Developer (minimum).
      * Cloud Application Administrator or Global Administrator (recommended).
    * The mailbox that will send emails must:
      * Be hosted in Microsoft 365 or Exchange Online.
      * Have [SMTP AUTH](https://learn.microsoft.com/en-us/exchange/clients-and-mobile-in-exchange-online/authenticated-client-smtp-submission) activated.
  </Accordion>

  <Accordion title="Create a Multi-Tenant Application in Microsoft Entra ID">
    1. Log in to the [Microsoft Entra admin center](https://entra.microsoft.com/).
    2. In the left sidebar, select **Entra ID > App registrations**.
    3. Click **New registration**. On the **Register an application** page, enter a name, for example, `Cognigy SMTP Integration`.
    4. In the **Supported account types** section, select **Accounts in any organizational directory (Any Azure AD directory - Multitenant)**. Follow [Microsoft best practices](https://learn.microsoft.com/en-us/entra/identity-platform/single-and-multi-tenant-apps#best-practices-for-multitenant-apps) for multitenant applications, especially regarding security and permission scope.
    5. Leave **Redirect URI** empty, this field isn't required for OAuth2 Client Credentials. Click **Register**.
    6. After the registration, the **Application Overview** page appears. Copy and securely store:

       * Application (client) ID.
       * Directory (tenant) ID.

       You will need these IDs later for the Send SMTP Email Node.
  </Accordion>

  <Accordion title="Create a Client Secret">
    1. In the created application, go to **Manage > Certificates & secrets**.
    2. Click **New client secret**.
    3. Enter the required information:

    * **Description** – for example, `Cognigy SMTP Secret`.
    * **Expiration** – specify the date according to your company's security policy.

    4. Click **Add**. Copy the appeared secret value and store it securely. You will need this secret later for the Send SMTP Email Node.
  </Accordion>

  <Accordion title="Configure API Permissions">
    1. [Add the SMTP.SendAsApp permission](https://learn.microsoft.com/en-us/exchange/client-developer/legacy-protocols/how-to-authenticate-an-imap-pop-smtp-application-by-using-oauth#add-the-pop-imap-or-smtp-permissions-to-your-microsoft-entra-application) to your Microsoft Entra application. Ensure [admin consent](https://learn.microsoft.com/en-us/exchange/client-developer/legacy-protocols/how-to-authenticate-an-imap-pop-smtp-application-by-using-oauth#get-tenant-admin-consent) is granted where required.
    2. [Register a Service Principal](https://learn.microsoft.com/en-us/exchange/client-developer/legacy-protocols/how-to-authenticate-an-imap-pop-smtp-application-by-using-oauth#register-service-principals-in-exchange) for your application.
  </Accordion>

  <Accordion title="Configure the Send SMTP Email Node">
    1. In the Send SMTP Email Node, select **Microsoft 365** as the email service.
    2. For the **Authentication Method**, select **OAuth2 Client Credentials** and click **+** to create a new connection.
    3. In the connection configuration, fill in the required fields:

    * **Connection Name** — enter a unique name for the connection.
    * **Client ID** — enter the application (Client) ID from Microsoft Entra.
    * **Client Secret** — enter the secret value created under **Certificates & Secrets** from Microsoft Entra.
    * **Scope** — enter `https://outlook.office365.com/.default`
    * **Token Endpoint URL** — enter `https://login.microsoftonline.com/{tenant_id}/oauth2/v2.0/token` and replace `{tenant_id}` with your Directory (tenant) ID from Microsoft Entra.
    * **User** — enter the email address of the sender mailbox. Make sure SMTP AUTH is activated for the user, and the user is granted the `SMTP.SendAsApp` permission.

    4. Click **Create** to save the connection, then fill in the required fields for the email configuration, such as **Recipient TO Email Addresses**, **Email Subject**, and **Email Content**.
  </Accordion>

  <Accordion title="Troubleshooting">
    | Issue                                              | Possible Cause                                                                                                                                 | Solution                                                                                                                                                                                                                                                              |
    | -------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | General errors                                     | Typos in the configuration, expired or missing secrets.                                                                                        | Double-check the scope, the token endpoint, the client secret, and ensure you are using the same mailbox as the OAuth2 credentials.                                                                                                                                   |
    | Invalid login or authentication failed             | **Custom FROM Email Address** is set, OAuth credentials mismatch, disabled SMTP AUTH, or wrong token endpoint or scope.                        | Leave **Custom FROM Email Address** empty, ensure credentials match the sender mailbox, enable SMTP AUTH, and verify the scope (`https://outlook.office365.com/.default`) and the token endpoint (`https://login.microsoftonline.com/{tenant_id}/oauth2/v2.0/token`). |
    | 401 Unauthorized                                   | Your application lacks `SMTP.SendAsApp`, Admin consent isn't granted, the Service Principal is missing, or the `SendAs` permission is missing. | Grant the `SMTP.SendAsApp` permission, provide admin consent, ensure the Service Principal exists in Exchange, and assign the `SendAs` permission for the mailbox.                                                                                                    |
    | 535 Authentication failed                          | The mailbox SMTP AUTH is deactivated.                                                                                                          | Activate SMTP AUTH in **Microsoft 365 Admin Center > Mailbox > SMTP AUTH**.                                                                                                                                                                                           |
    | Multitenant app issues or token acquisition failed | App not multi-tenant, the Service Principal missing, or incorrect client ID, tenant ID, or client secret.                                      | Ensure the app is multitenant, the Service Principal exists, and verify the client ID, tenant ID, and Client Secret used in Cognigy.                                                                                                                                  |
  </Accordion>
</AccordionGroup>


# Add To Context
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/logic/add-to-context



<Frame>
  <img alt="Add To Context Node configuration panel" />
</Frame>

## Description

This Node stores information in the [Context](/ai/agents/develop/ai-agent-memory/context) object. With this approach, you can use the stored value throughout the session on subsequent user inputs.

## Parameters

| Parameter | Type          | Description                                                                                                                                                                                                                                                                                                                                                                        |
| --------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Key       | CognigyScript | The key where to store in the Context object. For example, if you enter `requestedProduct`, the value is assigned to the `context.requestedProduct` key. You can use nested keys, for example, entering `user.profile.name` assigns the value to the `context.user.profile.name` key. This parameter doesn't support dashes or hyphens.                                            |
| Value     | CognigyScript | The value to store in the Context object.                                                                                                                                                                                                                                                                                                                                          |
| Mode      | Select        | <ul><li>**Simple** — assigns the value to the provided key.</li><li>**Array** — appends the value to an array stored under the key. If the key doesn't exist, a new array is created with the provided value as the first element. If the key contains a value that isn't an array, this value is replaced with an array including the value provided in this parameter.</li></ul> |

<Accordion title="Storing an Object">
  If you want to store an object in the Context object, use `.preserveType()` at the end of the CognigyScript expression to keep it as an object. For example, `input.slots.city[0].preserveType()` stores the Slot object for the first found city entity in the Context object.
</Accordion>


# Copy Data To Context
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/logic/copy-data-to-context



<Frame>
  <img alt="Copy Data To Context Node configuration panel" />
</Frame>

## Description

The Copy Data To Context Node copies the contents of the `input.data` object into the [Context](/ai/agents/develop/ai-agent-memory/context) object. The `input.data` object exists only for the current input. The Context object persists for the whole session, so the copy keeps the values available on later inputs.

Use this Node when a channel or an earlier Node places structured data in the Input object. The Node stores every key at once, so you don't need a separate [Add To Context](/ai/agents/develop/node-reference/logic/add-to-context) Node for each key.

The Node handles the data as follows:

* Copies top-level keys from input.data to Context. Nested objects and arrays retain their structure.
* Merges into Context. Keys that are absent from `input.data` remain unchanged.
* Replaces existing values. For example, a new `orderId` value overwrites the stored value.
* Resolves keys that contain a dot as nested paths. For example, the data key `user.id` writes to `context.user.id`.
* Copies nothing when `input.data` is empty. The Flow continues with the next Node.

<Note>
  The `input.data` object contains the data submitted with the input. Common sources include:

  * The `data` field of a [REST Endpoint](/ai/agents/deploy/endpoint-reference/rest) request.
  * The data payload of a [Webchat](/ai/agents/deploy/endpoint-reference/webchat) message.
  * The **Data Payload** parameter for [Webchat start behavior](/webchat/v3/configuration#start-behavior) or [Voice Gateway call events](/ai/agents/deploy/endpoint-reference/voice-gateway#specific-endpoint-settings).
  * Any Node that writes to the [Input](/ai/agents/develop/ai-agent-memory/input) object.
</Note>

## Parameters

This Node has no parameters. The Input object determines which keys the Node writes and which values it stores.

## Examples

### Copy Customer Details at Session Start

To test copying customer details from `input.data` to Context in the Interaction Panel:

1. Go to **Settings > Input**.
2. Activate **Show data input**. The **JSON data payload** field appears under the message field on the **Test** tab.
3. In the **Debug Mode** section, activate **Debug mode** to see a `Copied <key> to context` message for each copied key.
4. Enter the following JSON in the **JSON data payload** field and send the message. You can send the payload without message text.

```json theme={null}
{
  "customerId": "C-88213",
  "cart": { "items": 3, "total": 49.9 },
  "locale": "en-US"
}
```

5. Add a Copy Data To Context Node at the start of the Flow.
6. Go to the **Context** tab and verify that `customerId`, `cart`, and `locale` were copied.
7. You can reference these values later with CognigyScript, for example, `{{context.customerId}}` or `{{context.cart.total}}`.

Context is scoped to the session. Reset the conversation between test runs to avoid values from previous runs.

### Replace Existing Context Values

The Copy Data To Context Node replaces existing Context values when `input.data` contains a key with the same name.

For example, before the Node runs, the Context contains:

```json theme={null}
{
  "customerId": "C-12345",
  "locale": "de-DE",
  "cart": {
    "items": 1,
    "total": 19.9
  }
}
```

The current input contains the following `input.data`:

```json theme={null}
{
  "customerId": "C-88213",
  "cart": {
    "items": 3,
    "total": 49.9
  }
}
```

After the Copy Data To Context Node runs, the Context contains:

```json theme={null}
{
  "customerId": "C-88213",
  "locale": "de-DE",
  "cart": {
    "items": 3,
    "total": 49.9
  }
}
```

The Node replaces `customerId` and `cart` with the values from `input.data`. The `locale` key remains unchanged because it is not included in the current `input.data`.

## Alternatives

* To store a single value under a key that you choose, use the [Add To Context](/ai/agents/develop/node-reference/logic/add-to-context) Node.
* To store a found Slot instead of Input data, use the [Copy Slots to Context](/ai/agents/develop/node-reference/ai/copy-slots-to-context) Node.


# Execute Flow
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/logic/execute-flow



<Frame>
  <img alt="Execute Flow Node" />
</Frame>

## Description

An Execute Flow Node is used to immediately execute another Flow. This means that the Entrypoint will not switch to the new Flow. This operation can be compared to a function call.

After the target Flow finished executing, the current Flow will continue executing.

Cognigy monitors the number of times an Execute Flow Node is triggered without user input. The maximum loop count is set to 4 by default. This value is configurable only for on-premises installations.

When a Flow is executed, it will inherit the current Context from the parent Flow. The executed Flow can then make modifications to this Context, and these changes will be available to the parent Flow once the execution of the child Flow is completed.

## Parameters

| Parameter   | Description                                                                                                                                                                                                                                                                                         |
| ----------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Select Flow | Select the target Flow that is executed whenever the Flow execution triggers the **Execute Flow** Node. Click <img alt="target-flow" /> to open the target Flow.                                                                                                                                    |
| Select Node | A Node must be set as an Entrypoint before it can be selected. To set any Node as an Entrypoint, use the [Set Entrypoint](/ai/agents/develop/nodes/overview#set-nodes-as-entrypoints) setting. Select the target Node that executes whenever the Flow execution triggers the **Execute Flow** Node. |

<Accordion title="Advanced">
  | Parameter      | Type   | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
  | -------------- | ------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
  | Parse Intents  | Toggle | Determines whether Cognigy should parse Intents within the child Flow. This is enabled by default.<br /><br /> If the setting is enabled, Cognigy NLU will reparse the Intents in the target Flow, using either [input.text](/ai/agents/develop/ai-agent-memory/input#accessing-input-properties) or the [Optionally Injected Text](/ai/agents/develop/node-reference/logic/go-to). <br /><br /> This setting will parse the input into the child Flow, and the NLU will then score this text using the available Intents in the child Flow. This will not score based on Intents in the parent Flow. |
  | Parse Slots    | Toggle | Determines whether Cognigy should parse Slots within the child Flow. This is enabled by default.<br /><br /> If the setting is enabled, Cognigy NLU will reparse the Slots in the target Flow, using either [input.text](/ai/agents/develop/ai-agent-memory/input#accessing-input-properties) or the [Optionally Injected Text](/ai/agents/develop/node-reference/logic/go-to). <br /><br /> This setting will parse the input into the child Flow, and the NLU will then score this text using the available Slots in the child Flow. This will not score based on Slots in the parent Flow.         |
  | Absorb Context | Toggle | Determines whether the default Flow Context of the child Flow should be applied to the current Context. <br /><br /> If the child Flow has a value within its default Context, the value will be copied into the current Context as soon as you enter the child Flow.                                                                                                                                                                                                                                                                                                                                 |
</Accordion>

### Input Object Restoration

When you parse Intents and Slots in a child Flow, the Input Object behaves differently depending on the options you select.

If enable both **Parse Intents** and **Parse Slots**, most of the Input Object's properties are restored when you return to the parent Flow. However, the properties `inputId`, `execution`, and any custom properties added to the Input Object are not restored. For example:

1. Parent Flow receives an Input Object with intent **A**.
2. Child Flow parses the input and recognizes intent **B**.
3. Upon returning to the parent Flow, the Input Object will be restored with intent **A** and any slot values passed from the child Flow.

If disable both **Parse Intents** and **Parse Slots**, the Input Object will not be restored when you return to the parent Flow. For example:

1. Parent Flow receives an Input Object with intent **A**.
2. Child Flow parses the input but does not recognize any Intents or Slots.
3. Upon returning to the parent Flow, the Input Object will not be restored and any changes made in the child Flow will not be propagated to the parent Flow.


# Go To
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/logic/go-to



<Frame>
  <img alt="Go To Node configuration panel" />
</Frame>

## Description

A Go To Node permanently switches the conversation to a target Flow. The target Flow is executed immediately when this Node is activated. The conversation doesn't automatically return to the Flow where the Go To Node was triggered.

<Tip>
  To switch back to the Flow where the Go To Node was triggered, add another Go To Node or use the [Execute Flow Node](/ai/agents/develop/node-reference/logic/execute-flow) instead.
</Tip>

<Accordion title="Using Go To and Execute Flow Together">
  A Go To Node overrides the execution of an Execute Flow Node. If an Execute Flow Node switches to a Flow containing a Go To Node, the Go To Node stops the original execution and the session stays in the Go To Node’s target Flow. This action prevents the normal behavior of the Execute Flow Node, which returns to the Flow it is located in.
</Accordion>

## Parameters

| Parameter   | Description                                                                                                                                                                                                                                                                                                                                                                                   |
| ----------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Select Flow | Select the target Flow that is switched to whenever the Flow execution triggers the **Go To** Node. To manually enter the Flow ID, click **Manual Input** and enter the Flow ID. Click <img alt="target-flow" /> to open the target Flow.                                                                                                                                                     |
| Select Node | Select the target Node that is switched to whenever the Flow execution triggers the **Go To** Node. To manually enter the Node ID, click **Manual Input** and enter the Node ID. <br /><br />A Node must be set as an Entrypoint before it can be selected. To set any Node as an Entrypoint, use the [Set Entrypoint](/ai/agents/develop/nodes/overview#set-nodes-as-entrypoints) parameter. |

<AccordionGroup>
  <Accordion title="Advanced">
    | Parameter                | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
    | ------------------------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Execution Mode           | The Execution Mode determines the behavior of the Go To Node. It can go to the target Node and continue, or it can wait for input.                                                                                                                                                                                                                                                                                                                                                                                                                                   |
    | Optionally Injected Text | If the Execution Mode is set to **Go to Node and continue**, you can overwrite the text with which to continue here. If no text is specified, input.text is used.                                                                                                                                                                                                                                                                                                                                                                                                    |
    | Optionally Injected Data | If the Execution Mode is set to **Go to Node and continue**, you can overwrite the data with which to continue here. If no data is specified, input.data is used.                                                                                                                                                                                                                                                                                                                                                                                                    |
    | Parse Intents            | The user decides whether we should **parse intents** within the sub Flow. This is disabled by default.<br /><br />If the **Parse Intents** parameter is enabled, Cognigy NLU reparses the Intents in the target Flow, using either [input.text](/ai/agents/develop/ai-agent-memory/input#accessing-input-properties) or the Optionally Injected Text.<br /><br />This parameter parses the input into the sub Flow, and the NLU then scores this text using the available Intents in the sub Flow. This parameter doesn't score based on Intents in the parent Flow. |
    | Parse Slots              | The user decides whether we should **parse slots** within the sub Flow. This is disabled by default.<br /><br />If the **Parse Slots** parameter is enabled, Cognigy NLU reparses the Slots in the target Flow, using either [input.text](/ai/agents/develop/ai-agent-memory/input#accessing-input-properties) or the Optionally Injected Text.<br /><br />This parameter parses the input into the sub Flow, and the NLU then scores this text using the available Slots in the sub Flow. This parameter doesn't score based on Slots in the parent Flow.           |
    | Absorb Context           | If this parameter is enabled, the default Context of the target Flow is applied to the current Context upon switching.                                                                                                                                                                                                                                                                                                                                                                                                                                               |
  </Accordion>

  <Accordion title="Loops">
    Cognigy checks how often a Go To Node was activated without a user input in-between. The default setting for maximum loops is 4, but it is configurable for on-premises installations.
  </Accordion>
</AccordionGroup>


# If
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/logic/if



<Frame>
  <img alt="If Node configuration panel" />
</Frame>

## Description

This Node executes either one or another part of a [Flow](/ai/agents/develop/projects-and-flows/overview) based on a condition. An If Node has two child Nodes: a *Then Node* and an *Else Node*. The condition is evaluated during the execution of the Flow. If the condition evaluates to `true`, the Flow takes the path of the Then Node; if it evaluates to `false`, the path of the Else Node is executed.

## Parameters

### Condition Field

In the **Condition** field, use [CognigyScript](/ai/platform-features/cognigyscript) without `{{ }}`. For example, `input.slots.city` is all that is required. This evaluates to `true` if an Input Keyphrase with the tag `city` is found.

### Rule Editor

The Rule editor allows you to quickly create simple conditions.

<Frame>
  <img alt="If Node Rule editor with condition settings" />
</Frame>

### Advanced Editor

The Advanced editor allows you to write Conditions that go beyond one comparison.

<Frame>
  <img alt="If Node Advanced editor with CognigyScript conditions" />
</Frame>


# Interval
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/logic/interval



<Frame>
  <img alt="Interval Node configuration panel" />
</Frame>

## Description

An Interval Node executes a part of the Flow only if it has not been executed within a certain time span.

On reaching an Interval Node, only if it has not been executed within the defined time span, its child Nodes are executed, otherwise it is skipped.

At the beginning of a session, all Interval Nodes are active.

## Parameters

| Parameter | Type   | Description            |
| --------- | ------ | ---------------------- |
| Interval  | Number | Time (in milliseconds) |


# Lookup
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/logic/lookup



<a href="/release-notes/4.88"><Badge>Updated in 4.88</Badge></a>

<Frame>
  <img alt="Lookup Node" />
</Frame>

## Description

A Lookup Node creates different Flow paths based on a chosen operand.

The operand, also called lookup type, can be set to Intent, State, Type and Mode, or custom [CognigyScript](/ai/platform-features/cognigyscript). In the child Case Nodes, the specific values for the selected operand are entered. The operand is evaluated during the execution of the Flow, and the result is compared with the values of each case. If there is a match, the Flow path of the matched case is executed. If there is no case match, the Default Node is executed.

## Parameters

<Accordion title="Case Sensitivity">
  Lookup Case Nodes are case-sensitive. Ensure that you use the correct upper or
  lower case spelling in the **Value** field of your Case Nodes.
</Accordion>

### Type

<AccordionGroup>
  <Accordion title="Intent">
    This is the default setting for the Lookup Node, and it allows you to easily look up the Intent that has been triggered.

    By default, the Intent at the most detailed [Intent Hierarchy](/ai/platform-features/nlu/intents/intent-hierarchy/overview) level is matched. In this case, the Intents of all three levels are selectable in the Intent dropdown menu of the Case Nodes.

    However, you can also select a level via the **Intent Level** list if Intents should be matched only on a certain hierarchy level. In this case, only the Intents of the chosen level are selectable in the **Intent** list of the Case Nodes.

    For more information about building Flows with Intents, refer to the [NLU Overview](/ai/platform-features/nlu/overview) page.
  </Accordion>

  <Accordion title="Text">
    By changing the lookup type to **Text**, the Flow selects a case that matches
    the value of `input.text`. The case values must manually be written for this
    type of lookup.
  </Accordion>

  <Accordion title="State">
    By changing the lookup type to State, the Flow selects a case that matches the value of `input.state`. The case values must manually be written for this type of lookup.

    States allow limitations to be placed on a conversation that restrict the valid Intents that the NLU has access to. You can use [Intent Conditions](/ai/platform-features/nlu/intents/conditions) as an alternative for controlling Intent recognition.
  </Accordion>

  <Accordion title="Type">
    By changing the lookup type to **Type**, the Flow selects a case that matches the value of `input.type`. The case values must manually be written for this type of lookup.

    <Note>
      The Type is defined as a classification of the last user message as determined by the NLU. The type is an [Input](/ai/agents/develop/ai-agent-memory/input) variable that can be one of the following values: Statement, Command, Greeting, BGreeting, whQuestion, howQuestion, ynQuestion, pAnswer or nAnswer.
    </Note>
  </Accordion>

  <Accordion title="Mode">
    By changing the lookup type to Mode, the Flow selects a case that matches the value of `input.mode`. The case values must manually be written for this type of lookup.

    <Note>
      The Mode provides information on what was contained in the last user message. The type is an [Input](/ai/agents/develop/ai-agent-memory/input) variable that can be one of the following values: *TextOnly or TextData.* (Data only messages have an implicitly defined text that includes `DATA_ONLY_MESSAGE_` followed by a randomly generated string of 20 characters)
    </Note>
  </Accordion>

  <Accordion title="Handover Status">
    By changing the lookup type to **Handover Status**,
    the Flow selects a case that matches the value of `input.handover.status`.
    The case values must manually be written for a type of lookup.

    You can now receive **Status** and queue update messages from **Genesys Cloud** using the **Handover Status** Node. To learn how to do that, go to [Genesys Status and Bot Messages](/ai/escalate/handover-reference/genesys-cloud-open-messaging). If you wish to use RingCentral Engage instead, go to [RingCentral Engage](/ai/escalate/handover-reference/ring-central-engage) for more information.

    The table below shows handover statuses that you can configure in the Case Node.

    | Status                       | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                | Restrictions                                                                                                                                                                                                                                                                                                                                                                                                                                               |
    | ---------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Completed                    | Conversations are resolved successfully. Agent assistance is provided.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     | -                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
    | Cancelled                    | Conversations are ended by users, indicating termination before resolution.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                | -                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
    | Error                        | Conversations are encountering issues. This issues could be technical, communicative, or related to misinformation. Human agents work to address the issue and guide the conversation back to its intended course.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         | -                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
    | Events                       | Conversations are triggered by specific occurrences significant to the conversation. These could be time-sensitive events, important updates, or prompts that guide the conversation's direction or provide additional context.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            | -                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
    | Active                       | Conversations are actively managed by a human agent.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       | -                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
    | Queue                        | Conversations are on hold, awaiting agent assignment. The end user can receive a message if a Say Node is placed under the Case Node. The Say Node should contain the following text:<br /><br />`Estimated time: {{input.data.request.estimatedWaitTime / (1000 * 60)}} seconds`<br />`Position: {{input.data.request.position}}`<br /><br />Where `Estimated time` is the approximate duration in seconds that the user must wait for a response from a human agent, and `Position` is the conversation's order in the queue. <br /><br /> If you use the [Agent Copilot workspace](/agent-copilot/index) in your contact center, you should add the Handover Node and Lookup Node with the Queue status in your Agent Copilot Flow, along with a Say Node. In the Main Flow, after the Case Node with the Queue status, add the Go To Node that leads to the Agent Copilot Flow and its Lookup Node.                                                                                                                                                                                                                                                                                                    | The status is applicable only to the following handover providers:<ul><li>[RingCentral Engage](/ai/escalate/handover-reference/ring-central-engage) when the [Get Queue Updates](/ai/escalate/handover-reference/ring-central-engage) setting is activated.</li><li>[Cognigy Live Agent](/ai/escalate/handover-features/real-time-queue-notifications).</li><li>[Salesforce MIAW](/ai/escalate/handover-features/real-time-queue-notifications).</li></ul> |
    | Handover Already in Progress | Conversations maintain a handover connection when a user with the same ID opens multiple conversations. In this scenario, when the user opens a new conversation, the first conversation with the human agent becomes inactive. The chat continues in the new conversation with the human agent. No new handover is created, and the human agent has access to the history of both conversations with the user in one chat window.<br /><br />To inform the user that the first chat is inactive, add a Say Node after the Case Node. This message appears if the user revisits the first conversation. If the user creates multiple conversations, each chat displays this message, except for the last one that the user opened.<br /><br /> If you use the [Agent Copilot workspace](/agent-copilot/index) in your contact center, you should add the Handover Node and Lookup Node with the Handover Already in Progress status in your Agent Copilot Flow, along with a Say Node to inform the user that the first chat is inactive. In the Main Flow, after the Case Node with the Handover Already in Progress status, add the Go To Node that leads to the Agent Copilot Flow and its Lookup Node. | The status is applicable only to the [Genesys Cloud Messaging](/ai/escalate/handover-reference/genesys-cloud-open-messaging) provider.                                                                                                                                                                                                                                                                                                                     |
    | `genericHandoverUpdate`      | Conversations include messages from the Genesys Bot. By default, the Genesys Inbound Message flow routes messages to human agents only. However, you can configure your settings so that these messages are sent not only to human agents but also to end users. For more information, refer to the [Send Genesys Bot Messages to End Users](/ai/escalate/handover-reference/genesys-cloud-open-messaging) section.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        | The status is applicable only to the [Genesys Cloud Messaging](/ai/escalate/handover-reference/genesys-cloud-open-messaging) provider.                                                                                                                                                                                                                                                                                                                     |
  </Accordion>

  <Accordion title="CognigyScript">
    <Note>
      If you choose CognigyScript as the operand, use CognigyScript without `{{ }}`. You can also select to parse the CognigyScript as a string or not.
    </Note>

    By changing the lookup type to **CognigyScript**, the Flow selects a case that matches the value of any variable that is entered in the "Operator" field, either by using a *token* or by writing the variable path in *CognigyScript*. This feature makes it possible to change the Flow path based on any variable in the `input`, `context` or `profile`. The case values must manually be written for this type of lookup.

    To learn more, read the [CognigyScript](/ai/platform-features/cognigyscript) page.
  </Accordion>

  <Accordion title="Call Event Status">
    By changing the lookup type to **Call Event Status**, the Flow selects a case that matches the value of `input.data.event`. The case values must manually be written for this type of lookup.

    In the Case Node, you can handle various Voice Gateway events, including Recognized Speech, Recognized DTMF, Call Created, Answering Machine Detection, and more.

    More information about these events you can find in the [Voice Gateway Events](/voice-gateway/references/events/overview) reference.
  </Accordion>
</AccordionGroup>

### Advanced

#### Strict Mode

By default, the Strict Mode setting is inactive, which lets the Lookup Node accept values of any data type from the Input, Context, or Profile objects.
For instance, if you use CognigyScript to retrieve a value in integer format (for example, `2024`), it is considered equal to the string format (for example, `"2024"`)
in the **Value** field of the Case Node.

In Strict Mode, value types must be identical.
For example, the integer `2024` and the string `"2024"` aren't considered equal.
This mode is useful when strict type matching is necessary to ensure data is processed correctly.

When Strict Mode is enabled, the value retrieved from the Input, Context, or Profile objects must match the type of the value specified in the Case Node.
To achieve this result, in the **Value** field of the Case Node, convert the string to the required format using JavaScript.
For example, to convert the string `"2024"` to the integer `2024`, use `{{ Number("2024") }}` or CognigyScript to retrieve the value: `{{ Number(input.currentTime.year) }}`.

## An Alternative to Nested If Nodes

Lookup Nodes are useful when the number of possible options becomes too large for If Nodes.

Rather than using a series of conditional statements like if `input.intent === "orderFood"`, `else if input.intent === "orderDrink"`, `else if input.intent === "askHelp"`, and so on, you can simplify the process by using a Lookup on the Intent with multiple cases (`orderFood`, `orderDrink`, `askHelp`) and a default option.


# Once
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/logic/once



<Frame>
  <img alt="Once Node configuration panel" />
</Frame>

## Description

The Once Node is used to execute a part of the Flow only once during a session. It can be reset using the Reset Context Node.

The Once Node has two child Nodes: an On First Time Node and an Afterwards Node. The first time that a Once Node is activated in a session, the On First Time Node is executed. All future activations of the same Once Node execute the Afterwards Node. It is common to use a Once Node to deliver a single welcome message to a new user, and ensure that the same message is not delivered again throughout the conversation.


# Logic
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/logic/overview



Logic Nodes allow you to customize your Flow's behavior. You can use them to control which parts of the Flow should run based on conditions or integrate with other Flows.

<CardGroup>
  <Card title="Add To Context" href="/ai/agents/develop/node-reference/logic/add-to-context">
    Use the Add To Context Node to store information in the [Context](/ai/agents/develop/ai-agent-memory/context) object and use this information throughout the session.
  </Card>

  <Card title="Copy Data To Context" href="/ai/agents/develop/node-reference/logic/copy-data-to-context">
    The Copy Data To Context Node copies the contents of the Input `data` object into the [Context](/ai/agents/develop/ai-agent-memory/context) object.
  </Card>

  <Card title="Execute Flow" href="/ai/agents/develop/node-reference/logic/execute-flow">
    Use the Execute Flow Node to execute another Flow and return to the initial Flow afterward.
  </Card>

  <Card title="Go To" href="/ai/agents/develop/node-reference/logic/go-to">
    Use the Go To Node to switch between Flows or between Nodes.
  </Card>

  <Card title="If" href="/ai/agents/develop/node-reference/logic/if">
    The If Node evaluates a condition statement and triggers either the Then or Else child Nodes.
  </Card>

  <Card title="Interval" href="/ai/agents/develop/node-reference/logic/interval">
    Use the Interval Node to add a minimal interval before other Nodes are executed.
  </Card>

  <Card title="Lookup" href="/ai/agents/develop/node-reference/logic/lookup">
    The Lookup Node acts as a switch between different options. Based on the input value, the Node executes one of the options.
  </Card>

  <Card title="Once" href="/ai/agents/develop/node-reference/logic/once">
    Once Nodes limit the Flow execution after them to once per session, for example, for greetings or database requests.
  </Card>

  <Card title="Remove From Context" href="/ai/agents/develop/node-reference/logic/remove-from-context">
    The Remove From Context Node removes information from the [Context](/ai/agents/develop/ai-agent-memory/context) object.
  </Card>

  <Card title="Sleep" href="/ai/agents/develop/node-reference/logic/sleep">
    The Sleep Node pauses the Flow execution for a period you define.
  </Card>

  <Card title="Stop and Return" href="/ai/agents/develop/node-reference/logic/stop-and-return">
    The Stop and Return Node stops the Flow execution and returns to the Start Node.
  </Card>

  <Card title="Switch Locale" href="/ai/agents/develop/node-reference/logic/switch-locale">
    The Switch Locale Node changes [Locales](/ai/platform-features/translation-and-localization/localization) during a conversation, for example, based on language detection or user region.
  </Card>

  <Card title="Think" href="/ai/agents/develop/node-reference/logic/think">
    Use the Think Node to inject a text into the Start Node, for example, to restart the Flow, with specific data, text, or Intent.
  </Card>

  <Card title="Wait for Input" href="/ai/agents/develop/node-reference/logic/wait-for-input">
    The Wait for Input Node halts the Flow execution until the user input is received.
  </Card>
</CardGroup>


# Remove From Context
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/logic/remove-from-context



<Frame>
  <img alt="Remove From Context Node configuration panel" />
</Frame>

## Description

This Node removes information from the [Context](/ai/agents/develop/ai-agent-memory/context) object.

## Parameters

| Parameter | Type          | Description                                                                                                                                                                                                                                                                                                                                                 |
| --------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Key       | CognigyScript | The key of the property you want to remove from the Context object. For example, if you enter `requestedProduct`, it removes the `context.requestedProduct` property. You can use nested keys, for example, `user.profile.name` removes the `context.user.profile.name` property from the Context object. This parameter doesn't support dashes or hyphens. |
| Value     | CognigyScript | The value to remove from the array. Required only if you select **Array** in the **Mode** parameter.                                                                                                                                                                                                                                                        |
| Mode      | Select        | <ul><li>**Simple** — removes the key-value pair completely from the Context object.</li><li>**Array** — removes the value entered in the **Value** parameter from the array under the key in the Context object.</li></ul>                                                                                                                                  |


# Sleep
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/logic/sleep



<Frame>
  <img alt="Sleep Node configuration panel" />
</Frame>

## Description

A Sleep Node pauses the execution of a Flow for a certain time.

If your Flow returns multiple messages at once with a lot of content, you can use a Sleep Node to set a pause between messages, simulating a more natural, humanlike response behavior.

<Warning>
  This Node blocks the execution of Cognigy and prevents the execution of additional messages on the same Cognigy message processing instance (of which there are multiple). This can lead to issues in high-load scenarios.
</Warning>


# Stop and Return
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/logic/stop-and-return



<Frame>
  <img alt="Stop and Return Node configuration panel" />
</Frame>

## Description

A Stop and Return Node immediately stops the Flow execution, returns it to the Start Node in the same Flow, and waits for input.


# Switch Locale
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/logic/switch-locale



<Frame>
  <img alt="Switch Locale Node configuration panel" />
</Frame>

## Description

The Switch Locale Node changes the [Locale](/ai/platform-features/translation-and-localization/localization) of the Flow during a conversation. It is common to use this Node after language detection or based on the region of the user to execute a change to the conversational language or structure based on an assessment of the user's region or spoken language.

## Prerequisites

* A [Locale](/ai/platform-features/translation-and-localization/localization) must be configured in order to populate the options available for selection in the **Locale** list.

## Parameters

<a href="/release-notes/4.80"><Badge>Updated in 4.80</Badge></a>

| Parameter         | Description                                                                                                |
| ----------------- | ---------------------------------------------------------------------------------------------------------- |
| Use CognigyScript | Toggle to use CognigyScript to switch the Locale using the Locale reference ID in the **Locale** field.    |
| Locale            | Select the Locale to switch to or enter the Locale reference ID if the **Use CognigyScript** is activated. |

## Usage

To switch the Locale for the current execution, you need to use a [Think](/ai/agents/develop/node-reference/logic/think) Node after the Switch Locale Node. With this functionality, the current Flow is triggered again, but the new Locale is used.


# Think
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/logic/think



<a href="/release-notes/4.88"><Badge>Updated in 4.88</Badge></a>

<Frame>
  <img alt="Think Node configuration panel" />
</Frame>

## Description

This Node re-executes the Flow within a single iteration. The Node is useful when you want to restart the Flow, either with changed text or data or by explicitly triggering an Intent.

When the Think Node is reached, the Flow stops and restarts at the Start Node.
This behavior can be changed if you use an [Execute Flow Node](/ai/agents/develop/node-reference/logic/execute-flow) in the parent Flow to access a child Flow with a Think Node. When the Think Node is reached, the child Flow stops,
and execution resumes from the beginning of the parent Flow at its Start Node.

## Parameters

| Parameter         | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
| ----------------- | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Think Action      | Select        | A type of action to execute:<ul><li>**Think Text (default)** — lets you rerun the Flow with input text and/or input data, which will then be treated as a simulated message from the end user. This input will be sent through the NLU engine again, identifying Slots and matching Intents. This option is selected by default.</li><li>**Trigger Intent** — lets you rerun the Flow, ensuring a specific Intent is found instead of relying on the NLU engine. This approach is similar to the process outlined in [Trigger Intent](/ai/platform-features/nlu/intents/trigger-intent), where using an input like `cIntent:orderFood` triggers the `orderFood` Intent. You can achieve the same result by selection this option and entering `orderFood` in the **Intent to trigger** field.</li></ul> |
| Text to think     | CognigyScript | This field appears if the **Think Text (default)** action is selected.<br /><br /> The provided text serves as a message to emulate user input. This custom data will be available in the `input.text` object.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
| Data to think     | JSON          | This field appears if the **Think Text (default)** action is selected.<br /><br /> The provided data serves as a message to emulate user input. This custom data will be available in the `input.data` object and can include any additional information that are relevant to the Flow.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
| Intent to trigger | CognigyScript | This field appears if the **Trigger Intent** action is selected.<br /><br /> Cognigy will trigger the Intent using the [`cIntent`](/ai/platform-features/nlu/intents/trigger-intent) method.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |

### Error Handling

| Parameter                      | Type   | Description                                                                                                                                                                                                                                                                                                                                                                                                  |
| ------------------------------ | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| Log to System Logs             | Toggle | Log errors to the system logs. They can be viewed on the [Logs](/ai/agents/test/logs) page of your Project. The parameter is inactive by default.                                                                                                                                                                                                                                                            |
| Select Error Handling Approach | Select | You can select one of the Error Handling options:<br />- **Stop Flow Execution** — terminate the current Flow execution.<br />- **Continue Flow Execution** — allow the Flow to continue executing, bypassing the error and proceeding to the next steps.<br />- **Go to Node** — redirect the workflow to a specific Node in the Flow, which can be useful for error recovery or customized error handling. |
| Error Message (optional)       | Text   | Active when the **Continue Flow Execution** option is selected. Optional message to output if the Search or Extract service fails.                                                                                                                                                                                                                                                                           |
| Select Flow                    | Select | Active when the **Go to Node** option is selected. Select a Flow from the available options.                                                                                                                                                                                                                                                                                                                 |
| Select Node                    | Select | Active when the **Go to Node** option is selected. Select a Node from the available options.                                                                                                                                                                                                                                                                                                                 |

## Alternatives

When you need to process user input or other text in the middle of a Flow to retrieve an NLU result, and you do not want to restart the Flow from the Start Node, use the [Execute Cognigy NLU Node](/ai/agents/develop/node-reference/ai/execute-cognigy-nlu).


# Wait for Input
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/logic/wait-for-input



<Frame>
  <img alt="Wait for Input Node configuration panel" />
</Frame>

## Description

The Wait for Input Node stops the Flow execution and waits for the user input.
There is no timeout; the Flow execution stays paused until the user responds.

## More Information

* [Question Node](/ai/agents/develop/node-reference/basic/question)
* [Optional Question Node](/ai/agents/develop/node-reference/basic/optional-question)


# Knowledge Search
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/other-nodes/knowledge-search



<a href="/release-notes/4.61"><Badge>Updated in 4.61</Badge></a>

<Frame>
  <img alt="Knowledge Search Node configuration panel" />
</Frame>

## Description

<Danger>
  In Cognigy 2026.5.0, the Knowledge Search Node was removed. If you use this Node, it will still work, but you won't be able to create new Knowledge Search Nodes. To ensure a seamless transition, we strongly recommend using the [Search Extract Output](/ai/agents/develop/node-reference/other-nodes/search-extract-output) Node, as it offers the same functionality and additional features.
</Danger>

This Node searches and retrieves content from data that was previously uploaded to the Knowledge AI solution.

Before using this Node, ensure that you have a `text-embedding-ada-002` LLM configured. Refer to the [list of supported providers](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature).

To display the output of the Knowledge Search Node to the user, follow these steps:

1. In the Flow editor, add a Say Node below the Knowledge Search Node.
2. In the **Output Type** field, select **Text**.
3. In the **Text** field, click <img alt="token" /> and select the **KSearch:TopK Texts** Token.
4. Click **Save Node**.

## Parameters

### Knowledge Store

Select a store from the list to allow the Node to retrieve the data.

### Search Settings

| Parameter                     | Type          | Description                                                                                                                                                                                  |
| ----------------------------- | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Top K                         | Number        | Specify how many of the top K knowledge search results should be returned.                                                                                                                   |
| Where to store the result     | Select        | The storage for the search results. Select a storage: <ul><li>Input — the content will be stored in the Input object.</li><li>Context — the content will be stored in the Context.</li></ul> |
| Input Key to store result     | CognigyScript | The location in the input object where the result will be stored. For example, `input.knowledgeSearch`.                                                                                      |
| Input Context to store result | CognigyScript | The location in the context object where the result will be stored. For example, `context.knowledgeSearch`.                                                                                  |

## More Information

* [Search Extract Output](/ai/agents/develop/node-reference/other-nodes/search-extract-output)
* [Knowledge Search Overview](/ai/agents/develop/knowledge-ai/overview)


# LLM Entity Extract
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/other-nodes/llm-entity-extract



<a href="/release-notes/2026.18"><Badge>Updated in 2026.18</Badge></a>

<Frame>
  <img alt="LLM Entity Extract Node configuration panel" />
</Frame>

## Description

<Warning>
  As of Cognigy 2026.18.0, the **Temperature** parameter is deprecated. The removal is planned for April 2027. For models that don't support this parameter, activate the **Ignore deprecated parameters** toggle to exclude it from the request to the LLM provider. Otherwise, the request to the LLM provider might fail. If you plan to use this parameter, migrate it to the [Custom Options section](#custom-options) and activate the **Ignore deprecated parameters** toggle. Otherwise, Custom Options might not override it for every provider and model combination.

  For example, OpenAI models such as `gpt-5`, `gpt-5.1`, and `gpt-5.2` don't support these parameters. Similar models from other LLM providers might also not support these parameters. Check your LLM provider's documentation for model-specific support.
</Warning>

This Node uses a Large Language Model (LLM) to extract entities, such as product codes, booking codes, or customer IDs, from `input.text` of the Input object.

The Node supports both chat and voice use cases by processing text and transcribed speech inputs.

Before using this Node, set the LLM provider in the [Settings](/ai/agents/develop/gen-ai-and-llms/generative-ai).
You can configure the Node to either use the default model defined in the Settings or choose a specific configured LLM.

To view the extracted entity in the Interaction Panel, activate [debug mode](/insights/explorers/conversations#debug-mode).

To output the extracted entity, add a [Say Node](/ai/agents/develop/node-reference/basic/say) below the LLM Entity Extract Node in the Flow editor. In the **Text** field of the Say Node, use the key you specified in the [Storage Options](#storage-options) section, for example, `{{input.extractedEntity}}`.

## Parameters

| Parameter            | Type          | Description                                                                                                                                                                                                                             |
| -------------------- | ------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Large Language Model | List          | Select a model or use the default one.                                                                                                                                                                                                  |
| Entity Name          | CognigyScript | The name of the entity to extract. For example, `customerID`.                                                                                                                                                                           |
| Entity Description   | CognigyScript | A sentence that describes the entity. For example, `An alphanumeric string of 6 characters, such as ABC123 or 32G5FD`.                                                                                                                  |
| Example Input        | Text          | Examples of text inputs. For example, `My ID is AB54EE, is that ok?`, `That would be ah bee see double 4 three`, `I guess it's 49 A B 8 K`. Alternatively, you can click **Show JSON Editor** and add input examples in the code field. |
| Extracted Entity     | CognigyScript | Examples of extracted entities. For example, `AB54EE`, `ABC443`, `49AB8K`.                                                                                                                                                              |

<AccordionGroup>
  <Accordion title="Advanced">
    | Parameter                    | Type      | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
    | ---------------------------- | --------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Ignore deprecated parameters | Toggle    | Excludes the deprecated **Temperature** parameter from the request to the LLM provider. Activate this toggle if your model doesn't support this parameter. If you activate this toggle for a model that supports it, the model uses a default value.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
    | Temperature                  | Indicator | The appropriate sampling temperature for the model. Higher values mean the model will take more risks.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
    | Timeout                      | Number    | The maximum number of milliseconds to wait for a response from the LLM provider.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
    | Response Format              | Select    | Choose the format for the model's output result. You can select one of the following options:<ul><li>**None** — No response format is specified. Use this option if the LLM provider doesn't accept the response format you're using, or if you want to use the provider's default format. This option is selected by default.</li><li>**Text** — the model returns messages in text format.</li><li>**JSON Object** — the model returns messages in JSON format. In contrast to the [LLM Prompt](/ai/agents/develop/node-reference/service/llm-prompt) Node, this Node is already instructed to generate a JSON output when this option is selected Note that not all LLMs support this format, which may cause model calls to fail. For more information, refer to the LLM provider's API documentation.</li></ul> |
  </Accordion>

  <Accordion title="Custom Options">
    These options let you pass parameters that this Node doesn't expose, or override existing configurations.

    | Parameter              | Type | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
    | ---------------------- | ---- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Custom Model Options   | JSON | Sets custom parameters for the LLM configuration, such as `temperature`, `top_k`, or `seed`. These parameters override the ones set in this Node. For the full list of parameters that your model supports, refer to the LLM provider's API documentation, for example, [OpenAI](https://platform.openai.com/docs/api-reference/chat/create) or [Azure OpenAI](https://learn.microsoft.com/en-us/azure/ai-services/openai/). Example: `{ "seed": 42, "temperature": 0.2 }` |
    | Custom Request Options | JSON | Sets custom parameters for the request sent to the LLM provider, such as HTTP headers. These parameters override the ones set in this Node. For more information, refer to the LLM provider's API documentation. Example: `{ "headers": { "Authorization": "Bearer mytoken" } }`                                                                                                                                                                                           |
  </Accordion>

  <Accordion title="Storage Options">
    | Parameter                   | Type          | Description                                                                                                                                                                                     |
    | --------------------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | How to handle the result    | Select        | Determine how to handle the prompt result:<ul><li>**Store in Input** — stores the result in the Input object.</li><li>**Store in Context** — stores the result in the Context object.</li></ul> |
    | Input Key to store Result   | CognigyScript | The parameter appears when **Store in Input** is selected. The result is stored in the `extractedEntity` Input object by default. You can specify another key.                                  |
    | Context Key to store Result | CognigyScript | The parameter appears when **Store in Context** is selected. The result is stored in the `extractedEntity` Context object by default. You can specify another key.                              |
  </Accordion>

  <Accordion title="Debugging Options">
    When using the Interaction Panel, you can trigger two types of debug logs. These logs are only available when using the Interaction Panel and aren't intended for production debugging. You can also combine both log types.

    | Parameter                  | Type   | Description                                                                                                                                                                                                                                                                         |
    | -------------------------- | ------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Show Token Count           | Toggle | Send a debug message containing the input, output, and total token count. The message appears in the Interaction Panel when debug mode is activated. Cognigy uses the GPT-3 tokenizer, so actual token usage may vary depending on the model. The parameter is inactive by default. |
    | Log Request and Completion | Toggle | Send a debug message containing the LLM provider and the subsequent completion. The message appears in the Interaction Panel when Debug Mode is enabled. The parameter is inactive by default.                                                                                      |
  </Accordion>

  <Accordion title="Custom Options">
    These options let you use parameters that aren't included in the LLM Entity Extract Node or overwrite set configurations.

    | Parameter              | Type | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  | Example                                                                      |
    | ---------------------- | ---- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------- |
    | Custom Options         | JSON | Set custom parameters to the LLM configuration. You can set individual parameters as well as entire functions. These parameters customize the behavior of the model, such as adjusting `temperature`, `top_k`, or `presence_penalty`. This field overrides the parameters set in the LLM Entity Extract Node. To view the full list of available parameters for your model, refer to the LLM provider's API documentation, for example, [OpenAI](https://platform.openai.com/docs/api-reference/chat/create) or [Azure OpenAI](https://learn.microsoft.com/en-us/azure/ai-services/openai/). | `{ "model": "gpt-4o", "temperature": 0.7, "max_tokens": 4000 }`              |
    | Custom Request Options | JSON | Set custom parameters to customize the request to the LLM provider. These parameters override the configurations set in this Node. For example, setting a `{ "stream": true }` parameter overrides the **Output result immediately** option. For more information, refer to the LLM provider's API documentation.                                                                                                                                                                                                                                                                            | `{ "stream": true }`, `{ "headers": { "Authorization": "Bearer <token>" } }` |

    **Forcing Model Versions**

    You can force the LLM Entity Extract Node to use a specific model version by including it in the Custom Options.
    This means that the LLM Entity Extract Node will use the specified version of the language model instead of the default or any other available versions. This allows for more control over the behavior of the LLM Entity Extract Node, ensuring it uses the desired model version for generating prompts or responses.

    You can use models from any LLM provider supported by Cognigy, including those not yet directly integrated.
    However, you can only replace a model with another from the same provider.

    Let's consider an example with Anthropic:
    You can force the LLM Entity Extract Node to use the model version `claude-opus-4-1-20250805`, despite the LLM resource defaulting to the `claude-opus-4-5-20251101` model:

    1. [Create an Anthropic LLM resource](/ai/agents/develop/gen-ai-and-llms/providers/anthropic) for Claude, for example, `claude-opus-4-5-20251101`.
    2. Create a Flow and add an LLM Entity Extract Node to it.
    3. In the LLM Entity Extract Node, select the model `claude-opus-4-5-20251101` from the **Large Language Model** list.
    4. In the **Custom Model Options** field, add `{ "model": "claude-opus-4-1-20250805" }` to force the use of the `claude-opus-4-1-20250805` model.
    5. Click **Save Node**.

    The LLM Entity Extract Node now utilizes the `claude-opus-4-1-20250805` model.

    Below, you'll find documentation for supported models:

    * [Aleph Alpha Models](https://docs.aleph-alpha.com/)
    * [Anthropic Models](https://docs.anthropic.com/claude/docs/models-overview)
    * [Azure OpenAI](https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/models)
    * [OpenAI Models](https://platform.openai.com/docs/models)
  </Accordion>
</AccordionGroup>

## Examples

<AccordionGroup>
  <Accordion title="Extract a Booking Code from User Input">
    The user input is stored under `input.text`:

    ```text theme={null}
    Hello, my booking code is XYZ987. Can you confirm my reservation?
    ```

    LLM Entity Extract Node configuration:

    * **Entity Name**: `bookingCode`
    * **Entity Description**: `An alphanumeric booking code of 6 characters, such as XYZ123 or ABC987`
    * **Example Input**:
      * `My booking code is XYZ123.`
      * `It's X Y Z 1 2 3.`
    * **Extracted Entity**:
      * `XYZ123`

    The extracted entity is stored under `input.extractedEntity`:

    ```text theme={null}
    XYZ987
    ```
  </Accordion>

  <Accordion title="Extract a Product Code from User Input">
    The user input is stored under `input.text`:

    ```text theme={null}
    I’d like to order the product with code P45K2Q. Can you add it to my cart?
    ```

    LLM Entity Extract Node configuration:

    * **Entity Name**: `productCode`
    * **Entity Description**: `An alphanumeric product code, such as P12A3B or X45K2Q`
    * **Example Input**:
      * `The product code is P12A3B.`
      * `That would be P 1 2 A 3 B.`
    * **Extracted Entity**:
      * `P12A3B`

    The extracted entity is stored under `input.extractedEntity`:

    ```text theme={null}
    P45K2Q
    ```
  </Accordion>
</AccordionGroup>

## More Information

* [Question Node: LLM-extracted Entity](/ai/agents/develop/node-reference/basic/question#)


# Log Message
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/other-nodes/log-message



<Frame>
  <img alt="Log Message Node configuration panel" />
</Frame>

## Description

This Node writes a message into the [Project Logs](/ai/agents/test/logs).

## Parameters

| Parameter | Type          | Description                                                                                                                                                                                                                                                                                                                                                 |
| --------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Message   | CognigyScript | The message to log. The message should be a string that provides the content of the log entry. For example: `User login attempt successful`.                                                                                                                                                                                                                |
| Log Level | Select        | The log level to use. The level determines the severity and visibility of the log message. Select one of the following options:<ul><li>**Debug** — logs detailed information for debugging purposes.</li><li>**Info** — logs general information about system operations.</li><li>**Error** — logs critical errors that need immediate attention.</li></ul> |

## Alternatives

To log messages, you can also use the [API](/ai/for-developers/code/api-functions) object in a [Code](/ai/for-developers/code/overview) Node.

If you want to send a debug message to the Interaction Panel, use the [Debug Message](/ai/agents/develop/node-reference/basic/debug-message) Node for more precise testing.


# Other Nodes
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/other-nodes/overview



<CardGroup>
  <Card title="Knowledge Search" href="/ai/agents/develop/node-reference/other-nodes/knowledge-search">
    This Node searches and retrieves content from data that was previously uploaded to the Knowledge AI solution.
  </Card>

  <Card title="LLM Entity Extract" href="/ai/agents/develop/node-reference/other-nodes/llm-entity-extract">
    This Node uses a chosen LLM to extract entities, such as product codes, booking codes, and customer IDs, from a given string.
  </Card>

  <Card title="Log Message" href="/ai/agents/develop/node-reference/other-nodes/log-message">
    The Log Message Node writes a message into the [Project Logs](/ai/agents/test/logs).
  </Card>

  <Card title="Search Extract Output" href="/ai/agents/develop/node-reference/other-nodes/search-extract-output">
    This Node searches through data within the Knowledge AI solution, extracts a relevant answer via a Generative AI model, and creates an output.
  </Card>

  <Card title="Set Translation" href="/ai/agents/develop/node-reference/other-nodes/set-translation">
    This Node enables the override of real-time translation settings for the ongoing session. Translation settings are defined in the [Project Settings](/ai/administer/access/project-settings) page.
  </Card>
</CardGroup>


# Search Extract Output
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/other-nodes/search-extract-output



<a href="/release-notes/2026.18"><Badge>Updated in 2026.18</Badge></a>

<Frame>
  <img alt="Search Extract Output Node configuration panel" />
</Frame>

## Description

<Warning>
  As of Cognigy 4.69.0, the `topK.order` property of Knowledge Chunks in the search result generated by [Search Extract Output](/ai/agents/develop/node-reference/other-nodes/search-extract-output) Nodes is deprecated.
</Warning>

<Warning>
  As of Cognigy 2026.18.0, the **Temperature**, **Maximal Tokens**, **Presence Penalty**, **Frequency Penalty**, **Use Stops**, and **Stops** parameters are deprecated. The removal is planned for April 2027. For models that don't support these parameters, activate the **Ignore deprecated parameters** toggle to exclude them from the request to the LLM provider. Otherwise, the request to the LLM provider might fail. If you plan to use these parameters, migrate them to the [Custom Options section](#custom-options) and activate the **Ignore deprecated parameters** toggle. Otherwise, Custom Options might not override them for every provider and model combination.

  For example, OpenAI models such as `gpt-5`, `gpt-5.1`, and `gpt-5.2` don't support these parameters. Similar models from other LLM providers might also not support these parameters. Check your LLM provider's documentation for model-specific support.
</Warning>

This Node searches through data within the Knowledge AI solution, extracts a relevant answer via a generative AI model, and creates an output.

## Prerequisites

* Configure a model that supports [Knowledge Search](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) in the **Knowledge AI Settings** section in **Manage > Settings**.

## Parameters

### Mode

There are three modes that offer different levels of data extraction and presentation:

<Tabs>
  <Tab title="Search & Extract & Output">
    Performs a search, extracts an answer, and outputs the result as text or adaptive card.

    In the Search & Extract & Output mode, the following steps are involved:

    1. The system performs a knowledge search and retrieves relevant information.
    2. The retrieved data is sent to a Language Model (LLM) for further processing.
    3. The LLM extracts key points or answers from the search results, filtering out unnecessary details.
    4. The extracted information is stored in the Input or Context objects.
    5. The system automatically outputs the result from the LLM as either plain text or an adaptive card.

    For this mode, select models from the [list of supported providers](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) that cover the `LLM Prompt Node`, `Answer Extraction`, and `Knowledge Search` cases.
  </Tab>

  <Tab title="Search & Extract">
    Performs a search, extracts an answer, but doesn't perform an automatic output. Our customers can then manually create the output they desire.

    In the Search & Extract mode, the following steps are involved:

    1. The system performs a knowledge search and retrieves relevant information.
    2. The retrieved data is directly sent to the LLM for extraction of key points or answers.
    3. The LLM extracts the relevant information and writes it to the Input or Context objects.

    For this mode, select models from the [list of supported providers](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) that cover both the `LLM Prompt Node`, `Answer Extraction`, and `Knowledge Search` cases.
  </Tab>

  <Tab title="Search Only">
    Performs a knowledge search and retrieves information without extraction or automatic output.

    In the Search only mode, the following steps are involved:

    1. The system performs a knowledge search and retrieves relevant information.
    2. The search response is stored in the Input or Context objects.

    For this mode, select models from the [list of supported providers](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) that cover the `Knowledge Search` case.
  </Tab>
</Tabs>

### Knowledge Store

Select a store from the list to allow the Node to retrieve the data.

<AccordionGroup>
  <Accordion title="Search Settings">
    | Parameter                  | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | -------------------------- | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Context-Aware Search       | Select        | Considering the transcript's context for the search. It allows an AI Agent to answer follow-up questions. You can select one of the following options:<ul><li>**On** — is the default option. When this setting is enabled, it affects tokenization costs, which impact the amount of memory, computational power, and overall expenses needed for running an OpenAI or Azure OpenAI model. To get a better understanding of how the feature works in practice, see [Examples](#examples).</li><li>**Off** — the setting is turned off.</li></ul>                                                                                                                                                                                                                                                       |
    | Transcript Steps           | Slider        | Number of prior transcript steps to use in a context-aware search. This parameter is active when the Context-Aware Search setting is enabled. One transcript step consists of a pair: one user input and one AI Agent output. For example, if you select a quantity of 2, the system considers the 2 most recent pairs of user input and AI Agent output.                                                                                                                                                                                                                                                                                                                                                                                                                                               |
    | Where to store the result  | Select        | The storage for the search results. Select a storage:<ul><li>**default** — the content is stored in the Input object (`input.knowledgeSearch`).</li><li>**Input** — the content is stored in the Input object.</li><li>**Context** — the content is stored in the Context.</li></ul>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
    | String to search for       | CognigyScript | The string to search for in the Knowledge Store. A default value is stored in `input.text`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | Top K                      | Slider        | Specify how many of the knowledge search results are returned. The results are stored under `input.knowledgeSearch.topK`. Any `\n` and `\t` instances in the [Knowledge Chunk text](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk#knowledge-chunk-text) are replaced with spaces in the results.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
    | Source Tags                | CognigyScript | The tags serve to refine the scope of your knowledge search, allowing you to include only the most pertinent sections of the knowledge base and, as a result, improve the accuracy of search outputs. <br /><br /> Before specifying tags, ensure that they were provided during the creation of the Knowledge Sources. Add Tags by specifying each Tag separately and pressing <kbd>Enter</kbd>. The maximum number of tags is 5. <br /><br />When you specify multiple Source Tags, the Search Extract Output Node defaults to an `AND` operator, meaning it only considers Sources that have all the specified Tags. This approach ensures the search results are precise and highly relevant to the end user's query. To change this behavior, go to the **Match Types for Source Tags** parameter. |
    | Match type for Source Tags | Select        | The operator to filter Knowledge Sources by Source Tags. Select one of the following options:<ul><li>**AND** — the default value, requires all tags to match across multiple Knowledge Sources. Consider the following example: there are Knowledge Sources with Tags `S-a`, `S-b`, and `S-c`. When you use the `AND` operator to filter by `S-a` and `S-b`, only Sources with both Tags `S-a` and `S-b` are included in the search results.</li><li>**OR** — requires at least one tag to match across multiple Knowledge Sources. Consider the following example: there are Knowledge Sources with Tags `S-a`, `S-b`, and `S-c`. When you use the `OR` operator to filter by `S-a` or `S-b`, any Source with either Tag `S-a` or `S-b` is included in the search results.</li></ul>                   |

    <Accordion title="Source Tags">
      * You can add CognigyScript within Source Tags.
      * Source Tags are always lower-case, only alpha-numerics (a-z, 0-9), underscores, and hyphens are allowed.
      * When using CognigyScript for Source Tags, it is important to lowercase them. For example, by using the following: `{{input.text.toLowerCase()}}`.
    </Accordion>
  </Accordion>

  <Accordion title="Extract Settings">
    | Parameter                    | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         |
    | ---------------------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Ignore deprecated parameters | Toggle        | Excludes the deprecated **Temperature**, **Maximal Tokens**, **Presence Penalty**, **Frequency Penalty**, **Use Stops**, and **Stops** parameters from the request to the LLM provider. Activate this toggle if your model doesn't support these parameters. If you activate this toggle for a model that supports them, the model uses default values.                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | Prompt                       | CognigyScript | The prompt for generating completions. This prompt is designed for ChatGPT-based models. The behavior of this prompt may vary depending on the model used. You can change the prompt at your discretion. <br /><br /> In the prompt, the following tags are used to retrieve information from the Knowledge Store and Input object to ensure that the LLM receives all the necessary information to generate the most accurate response: <ul><li>`@answerLanguage` – the language that refers to the `input.language` value from the Input object for the current locale.</li><li>`@foundDocuments` – the top-k knowledge results determined from the end user's question.</li><li>`@userInput` – the message that the end user sent to the AI Agent in the conversation.</li></ul> |
    | Temperature                  | Slider        | The appropriate sampling temperature for the model. Higher values mean the model takes more risks.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
    | Maximal Tokens               | Slider        | The maximum number of tokens to generate in the completion.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         |
    | Presence Penalty             | Slider        | Number between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood of talking about new topics.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
    | Frequency Penalty            | Slider        | Number between -2.0 and 2.0. The penalty assigns a lower probability to tokens frequently appearing in the generated text, encouraging the model to generate more diverse and unique content.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
    | Use Stops                    | Toggle        | Whether to use a list of stop words to let Generative AI know where the sentence stops.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | Stops                        | Text          | Up to 4 sequences where the API stops generating further tokens. The returned text doesn't contain the stop sequence.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
    | Timeout                      | Number        | The maximum number of milliseconds to wait for a response from the LLM provider.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
    | Timeout Message              | CognigyScript | The message to output if the timeout is reached.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
  </Accordion>

  <Accordion title="Output Settings">
    | Parameter            | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                       |
    | -------------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | OutputMode           | Select        | This parameter allows you to choose between two options for the output:<ul><li>**Text** — select this option if you want the system to provide a basic text-based response as the output.</li><li>**Adaptive Card** — select this option if you want a visual representation of an Adaptive Card as the output. Adaptive Cards are interactive and display rich content with buttons, images, and more.</li></ul> |
    | Output Fallback Text | CognigyScript | The text to output if no answer was found.                                                                                                                                                                                                                                                                                                                                                                        |
  </Accordion>

  <Accordion title="Error Handling">
    | Parameter                      | Type   | Description                                                                                                                                                                                                                                                                                                                                                                                                              |
    | ------------------------------ | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Select Error Handling Approach | Select | You can select one of the Error Handling options:<ul><li>**Stop Flow Execution** — terminate the current Flow execution.</li><li>**Continue Flow Execution** — allow the Flow to continue executing, bypassing the error and proceeding to the next steps.</li><li>**Go to Node** — redirect the workflow to a specific Node in the Flow, which can be useful for error recovery or customized error handling.</li></ul> |
    | Error Message (optional)       | Text   | Active when the **Continue Flow Execution** option is selected. Optional message to output if the Search or Extract service fails.                                                                                                                                                                                                                                                                                       |
    | Select Flow                    | Select | Active when the **Go to Node** option is selected. Select a Flow from the available options.                                                                                                                                                                                                                                                                                                                             |
    | Select Node                    | Select | Active when the **Go to Node** option is selected. Select a Node from the available options.                                                                                                                                                                                                                                                                                                                             |
  </Accordion>

  <Accordion title="Custom Options">
    These options let you use parameters that aren't included in the Search Extract Output Node or overwrite set configurations.

    | Parameter              | Type | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            | Example                                                                      |
    | ---------------------- | ---- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------- |
    | Custom Options         | JSON | Set custom parameters for the LLM configuration. You can set individual parameters as well as entire functions. These parameters customize the behavior of the model, such as adjusting `temperature`, `top_k`, or `presence_penalty`. This field overrides only the parameters set to generative models in the Search Extract Output Node; the field doesn't support custom options for embedding models used in the knowledge search. To view the full list of available parameters for your model, refer to the LLM provider's API documentation, for example, [OpenAI](https://platform.openai.com/docs/api-reference/chat/create) or [Azure OpenAI](https://learn.microsoft.com/en-us/azure/ai-services/openai/). | `{ "model": "gpt-4o", "temperature": 0.7, "max_tokens": 4000 }`              |
    | Custom Request Options | JSON | Set custom parameters to customize the request to the LLM provider. These parameters override the configurations set in this Node. For example, setting a `{ "stream": true }` parameter overrides the **Output result immediately** option. For more information, refer to the LLM provider's API documentation.                                                                                                                                                                                                                                                                                                                                                                                                      | `{ "stream": true }`, `{ "headers": { "Authorization": "Bearer <token>" } }` |

    **Forcing Model Versions**

    You can force the Search Extract Output Node to use a specific model version by including it in the Custom Options.
    This means that the Search Extract Output Node will use the specified version of the language model instead of the default or any other available version. This allows for more control over the behavior of the Search Extract Output Node, ensuring it uses the desired model version for generating prompts or responses.

    You can use models from any LLM provider supported by Cognigy, including those that aren't yet directly integrated.
    However, you can only replace a model with another from the same provider.

    Let's consider an example with Anthropic:
    You can force the Search Extract Output Node to use the model version `claude-opus-4-1-20250805`, although the LLM resource defaults to the `claude-opus-4-5-20251101` model:

    1. [Create an Anthropic LLM resource](/ai/agents/develop/gen-ai-and-llms/providers/anthropic) for Claude, for example, `claude-opus-4-5-20251101`.
    2. Create a Flow and add a Search Extract Output Node to it.
    3. In the Search Extract Output Node, select the model `claude-opus-4-5-20251101` from the **Large Language Model** list.
    4. In the **Custom Model Options** field, add `{ "model": "claude-opus-4-1-20250805" }` to force the use of the `claude-opus-4-1-20250805` model.
    5. Click **Save Node**.

    The Search Extract Output Node now utilizes the `claude-opus-4-1-20250805` model.

    Below, you'll find documentation for supported models:

    * [Aleph Alpha Models](https://docs.aleph-alpha.com/)
    * [Anthropic Models](https://docs.anthropic.com/claude/docs/models-overview)
    * [Azure OpenAI](https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/models)
    * [OpenAI Models](https://platform.openai.com/docs/models)
  </Accordion>
</AccordionGroup>

### Debugging Options

| Parameter                  | Type   | Description                                                                                                                                                                                                                                                                                      |
| -------------------------- | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| Show Token Count           | Toggle | Send a debug message containing the input, output, and total token count. The message appears in the Interaction Panel when Debug Mode is enabled. Cognigy uses the GPT-3 tokenizer algorithm, so actual token usage may vary depending on the model used. The parameter is inactive by default. |
| Log Request and Completion | Toggle | Send a debug message containing the request sent to the LLM provider and the subsequent completion. The message appears in the Interaction Panel when Debug Mode is enabled. The parameter is inactive by default.                                                                               |

## Examples

### Context-Aware Search

By considering the context,
the search can help users find information within a transcript more quickly,
as it understands the nuances of the content.
When enabling a context-aware search system,
you can configure the number of prior transcript steps to use in the search to suit your specific requirements.
The number of prior transcript steps affects the depth of context considered when retrieving search results.

In the following table, a dialogue between the user and the AI Agent is presented, showing the agent's responses in both the enabled and disabled modes of the Context-Aware Search feature.

| <div>Context-Aware Search is off</div>                                                                                                                                                                                                                                                                                                                                         | <div>Context-Aware Search is on</div>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
| ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| **user:** What animals are allowed on a plane?<br /><br />**AI Agent:** Passengers are allowed to bring a maximum of two animals (dogs or cats) on the plane, either in the cabin or in the cargo hold, as long as they are in approved transport containers.<br /><br />**user:** Give me more details!<br /><br />**AI Agent:** Unfortunately, I can't answer your question. | **user:** What animals are allowed on a plane?<br /><br />**AI Agent:** Passengers are allowed to bring a maximum of two animals (dogs or cats) on the plane, either in the cabin or in the cargo hold, as long as they are in an approved transport container.<br /><br />**user:** Give me more details!<br /><br />**AI Agent:** The approved transport container for animals on a plane must have a maximum size of 118 cm (55 x 40 x 23 cm) or 47 in (22 x 16 x 9 in) and a total weight, including the animal, not exceeding 8 kg. The container should be escape-proof, leak-proof, bite-proof, and lined with absorbent material. Only one transport container per passenger is allowed in the cabin, and the number of animals per container is limited to one, with exceptions for specific cases. |

## More Information

* [Knowledge Search Node](/ai/agents/develop/node-reference/other-nodes/knowledge-search)
* [Knowledge Search Overview](/ai/agents/develop/knowledge-ai/overview)
* [LLM](/ai/agents/develop/gen-ai-and-llms/llms)


# Set Translation
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/other-nodes/set-translation



<a href="/release-notes/2026.11"><Badge>Updated in 2026.11</Badge></a>

<Frame>
  <img alt="Set Translation Node configuration panel" />
</Frame>

## Description

This Node lets you override the real-time translation settings defined in the [Endpoint settings](/ai/agents/deploy/endpoints/real-time-translation-settings) for the ongoing session. The settings defined in this Node persist for all subsequent user inputs, including during a handover to a human agent, until another Set Translation Node is executed or the session ends.

## Restrictions

* Translation with the Set Translation Node is supported only for the output types of the [AI channel](/ai/agents/develop/node-reference/basic/say) in your Nodes.

## Parameters

| Parameter                                  | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
| ------------------------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Translation Enabled                        | Sets how to override the real-time translation settings in the Endpoint settings:<ul><li>When turned on, overrides real-time translation set in the Endpoint settings for the rest of the session.</li><li>When turned off, deactivates real-time translation and defaults the **User Input Language** and **Flow Language** settings to English.</li></ul>                                                                                                                               |
| User Input Language                        | Sets the language identified in the user input. Enter the respective language code supported by [the machine translation provider you configured](/ai/agents/deploy/endpoints/real-time-translation-settings). If an unsupported language code is provided, the user input language defaults to English.                                                                                                                                                                                  |
| Flow Language                              | Sets the language to translate the output to. Enter the respective language code supported by [the machine translation provider you configured](/ai/agents/deploy/endpoints/real-time-translation-settings). If an unsupported language code is provided, the Flow language defaults to English.                                                                                                                                                                                          |
| Don't translate Payloads                   | Avoids translating the payload content when it is returned. This parameter is activated by default.                                                                                                                                                                                                                                                                                                                                                                                       |
| No-Translation Marker                      | Sets the no-translation marker. Any text inside two no-translation markers isn't translated in the AI Agent output. This parameter can be used for trademarks or other business terms that must not be translated. By default, the no-translation marker is `**`. For example, if you write `Willkommen, **Moin**!` in a Say Node of your Flow and the detected language is English, the AI Agent sends `Welcome, Moin!` to the user because `Moin` is inside the no-translation markers. |
| Set user input language on execution count | This parameter works only when the **User Input Language** is set to `Auto-Detect`. Sets the number of user inputs the AI Agent uses to identify the user's language.                                                                                                                                                                                                                                                                                                                     |
| Always remove No-translation Markers       | Removes the no-translation markers from the user-facing output, even when translation isn't enabled.                                                                                                                                                                                                                                                                                                                                                                                      |


# Node Reference
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/overview

Explore the Node reference documentation for configuring Flows in Cognigy, including Basic, Logic, Voice, Analytics, Service, AI, and other Node categories.

This Node reference provides details about the [Nodes'](/ai/agents/develop/nodes/overview) parameters to help you configure your [Flow](/ai/agents/develop/projects-and-flows/overview).

## Node Categories

<CardGroup>
  <Card title="Basic" href="/ai/agents/develop/node-reference/basic/overview">
    Use Basic Nodes for basic conversational Flows, for example, the [Say Node](/ai/agents/develop/node-reference/basic/say) and the [Question Node](/ai/agents/develop/node-reference/basic/question).
  </Card>

  <Card title="Logic" href="/ai/agents/develop/node-reference/logic/overview">
    Use Logic Nodes to control the Flow execution, for example, the [If Node](/ai/agents/develop/node-reference/logic/if) and the [Lookup Node](/ai/agents/develop/node-reference/logic/lookup).
  </Card>

  <Card title="Voice" href="/ai/agents/develop/node-reference/voice/overview">
    Voice Nodes are for developing voice AI Agents.
  </Card>

  <Card title="Analytics" href="/ai/agents/develop/node-reference/analytics/overview">
    Nodes for tracking and analyzing conversation data and taking advantage of Cognigy AI's comprehensive marketing functionality.
  </Card>

  <Card title="CXone" href="/ai/agents/develop/node-reference/cxone/overview">
    Nodes for integrating NiCE CXone into your Flows.
  </Card>

  <Card title="Service" href="/ai/agents/develop/node-reference/service/overview">
    Use Service Nodes to trigger external services, like databases and email servers. Additional services can be added through [Extensions](/ai/for-developers/extensions).
  </Card>

  <Card title="AI" href="/ai/agents/develop/node-reference/ai/overview">
    AI Nodes are a collection of Nodes for Cognigy [NLU](/ai/platform-features/nlu/overview) and [Agentic AI](/ai/agents/overview).
  </Card>

  <Card title="Data" href="/ai/agents/develop/node-reference/data/overview">
    Data Nodes are designed to manage, manipulate, process, or store data.
  </Card>

  <Card title="Agent Copilot" href="/ai/agents/develop/node-reference/agent-copilot/overview">
    Nodes for creating the Agent Copilot workspace.
  </Card>

  <Card title="xApps" href="/ai/agents/develop/node-reference/xApp/overview">
    Nodes for creating xApps.
  </Card>

  <Card title="Other" href="/ai/agents/develop/node-reference/other-nodes/overview">
    Additional specialized Nodes.
  </Card>
</CardGroup>

If you want to create a custom Node, go to the [Extensions](/ai/for-developers/extensions) section.

<Note>
  Nodes can belong to multiple categories for easier search and navigation. In the following sections, each Node is assigned to a single category to avoid duplicated descriptions.
</Note>

## Common Sections to All Node Types

### Settings

All Node types include the **Settings** section, which contains the following parameters:

| Parameter      | Description                                                                                                              |
| -------------- | ------------------------------------------------------------------------------------------------------------------------ |
| Label          | The name of the Node that is displayed in the Flow editor. This is useful to distinguish between Nodes of the same type. |
| Comment        | A comment that is displayed in the Flow editor. This is useful to add additional information about the Node.             |
| Analytics Step | An [analytic step](/ai/agents/analyze/analytics-steps) to track which paths users take in the Flow.                      |

### Mocking

Use this section to control [Mocking](/ai/agents/test/mocking) at the Node level.


# Add Transcript Step
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/service/add-transcript-step



<Frame>
  <img alt="Add Transcript Step Node configuration panel" />
</Frame>

## Description

The Add Transcript Step Node allows you to add messages to the conversation transcript. These messages become part of the LLM context and directly influence AI Agent behavior, tool handling, conversation flow, and contextual awareness.

Use this Node to inject additional information into the transcript without requiring a new user input. The Node is especially helpful in complex AI Agent architectures.

## Restrictions

* This Node isn't compatible with the `@cognigyRecentConversation` and `@cognigyRecentUserInputs` tags in the [LLM Prompt Node](/ai/agents/develop/node-reference/service/llm-prompt).
* Overuse of empty or redundant transcript entries can reduce clarity and make debugging more difficult.
* Entries added with this Node count toward the same 50-turn transcript limit used by the [Get Transcript](/ai/agents/develop/node-reference/service/get-transcript) Node. Once the limit is reached, the oldest entries are automatically removed to make room for new ones.

## Parameters

| Parameter    | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
| ------------ | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Role         | Selector      | Defines the conversation actor for the added transcript entry. The selected role determines how the LLM interprets the message and how strongly it may influence behavior. Select one of the following roles:<ul><li>**User** — represents direct input from the end user, such as questions, requests, or instructions. User messages are primary drivers of the model's responses.</li><li>**Assistant** — represents previous AI-generated responses in the conversation. Assistant messages provide context and continuity from the model's perspective.</li><li>**Agent** — represents internal reasoning or planning steps when the LLM operates in agentic mode. Agent messages can influence structured thinking and tool selection within multi-agent workflows.</li><li>**System** — defines overarching rules, constraints, persona, and behavioral guidelines. System messages have strong influence over the model's behavior and establish global instructions.</li><li>**Tool** — represents outputs returned from external systems or tools invoked by the AI Agent. Tool messages provide factual results, for example, API responses, that the model uses to continue reasoning or generate an answer.</li></ul> Only transcript entries included in the active conversation context are visible to the LLM during the request. |
| Type         | Selector      | Defines the type of transcript entry. The selected type determines how the message is processed and interpreted within the conversation transcript. Available types depend on the selected option in the **Role** parameter:<ul><li>**Input** — used when **User** is selected. Represents messages provided by the end user.</li><li>**Output** — used when **Agent** is selected and available when **Assistant** is selected. Represents natural-language responses generated by the assistant or agent and shown to the end user.</li><li>**Tool Call** — available when **Assistant** is selected. Represents structured requests to invoke a function or external tool. Not shown directly to the end user.</li><li>**Debug Log** — used when **System** is selected. Internal system messages for debugging and diagnostics.</li><li>**Tool Answer** — used when **Tool** is selected. Represents responses returned from external tools or functions.</li></ul>                                                                                                                                                                                                                                                                                                                                                                           |
| Name         | CognigyScript | Appears if **Tool Call** or **Tool Answer** is selected as **Type**. This parameter allows you to specify the name of the tool being called, for example, `getWeather`. The tool name can be used in the LLM context to provide clarity about which tool is being invoked and to help guide the model's reasoning and response generation.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
| ID           | CognigyScript | Appears if **Tool Call** is selected as **Type**. This parameter allows you to specify a unique identifier for the tool call, for example, `getWeather_12345`. The ID can be used to track specific tool calls and their corresponding responses, especially in complex conversations with multiple tool interactions. It can also help ensure that the correct tool answer is associated with the right tool call in the LLM context.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
| Tool Call ID | CognigyScript | Appears if **Tool Answer** is selected as **Type**. This parameter allows you to specify the unique identifier of the corresponding tool call, for example, `getWeather_12345`. The **Tool Call ID** parameter can be used to link the tool answer back to the specific tool call it is responding to, ensuring that the LLM can correctly associate the response with the original request and maintain coherent reasoning in conversations with multiple tool interactions.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
| Content      | CognigyScript | Appears if **Tool Answer** is selected as **Type**. This parameter allows you to specify the content of the tool answer, for example, `Location is New York. Date is 2024-06-01`. The content can provide the LLM with the necessary information returned from the external tool, which can then be used to generate informed responses or continue reasoning based on the results of the tool call. The parameter helps ensure that the LLM has access to the relevant data needed to process the tool answer effectively.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
| Input        | JSON          | Appears if **Tool Call** is selected as **Type**. This parameter allows you to specify the input parameters for the tool call in a structured JSON format, for example, `{"index":0,"id":"call_zxXusUexoviWzfwqjNHLfjP4","type":"function","function":{"name":"unlock_account","arguments":{"email":"elena@cognigy.com"}}}`. The input data can be used by the LLM to understand the context of the tool call and to generate more accurate responses based on the provided parameters. It also helps ensure that the tool receives the necessary information to execute correctly.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
| Header       | CognigyScript | Appears if **Debug Log** is selected as **Type**. This parameter allows you to specify a header for the debug log entry, for example, `API Response Debug`. The header can help categorize and identify different debug messages in the transcript, making it easier to analyze and troubleshoot issues during development. The parameter provides a clear label for the type of information being logged.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
| Message      | CognigyScript | Appears if **Debug Log** is selected as **Type**. This parameter allows you to specify the content of the debug log message, for example, `Received response status: Success`. The message can contain detailed information about the system's internal state, API responses, or other relevant data for diagnosing and resolving issues during the development process. The parameter provides valuable insights into the system's behavior.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
| Meta Data    | JSON          | Appears if **Debug Log** is selected as **Type**. This parameter allows you to specify additional metadata for the debug log entry in a structured JSON format, for example, `{ "timestamp": "2024-06-01T12:00:00Z", "severity": "info" }`. The metadata can provide context about the debug message, such as when it was logged and its severity level, which can help developers prioritize and analyze logs more effectively during troubleshooting.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
| Text         | CognigyScript | Appears if **Input** or **Output** is selected as **Type**. Enter the text depending on the type of the message: <ul><li>**Input** – represents messages from the end user. Intended for questions, requests, instructions, or any data the user provides. These entries serve as the primary input for the LLM and directly influence the AI Agent's responses. The text here becomes part of the LLM context for understanding user intent.</li><li>**Output** – represents messages generated by the AI Agent or human agent.<ul><li>If **Assistant** is selected as the role: enter natural language responses shown to the user. These responses maintain conversation continuity and provide instructions, explanations, or information.</li><li>If **Agent** is selected as the role: enter internal notes, reasoning, or guidance that influence workflow, structured decision-making.</li></ul></li></ul>                                                                                                                                                                                                                                                                                                                                                                                                                                |
| Data         | JSON          | Appears if **Input** or **Output** is selected as **Type**. Optional structured metadata that complements the **Text** field. For example, `{ "type": "motivational" }` can be used to categorize the input or provide context to influence how the system processes the text.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |

## Use Cases

The following use cases illustrate how the Add Transcript Step Node is used:

* **Add External System Results**. Inject API responses such as CRM data, ID\&V results, or order information into the transcript so the AI Agent can use this data for reasoning.
* **Manage AI Agent Transitions**. Add transcript entries to mark milestones, such as completed handovers or transitions between AI Agents.
* **Context Refocusing**. Insert targeted messages to highlight new priorities or provide technical context in long-running conversations.
* **Workflow-Based Context Injection**. Add follow-up instructions to guide subsequent behavior after deterministic processes—for example, GDPR handling.

## Examples

The examples show how different message types from different roles appear in the conversation transcript, how they are processed, and how they influence the conversation. Each role is shown with its corresponding JSON representation in the Input object.
The `"source": "system"` key-value pair indicates that the role was added to the transcript by the system using the Add Transcript Step Node.

<AccordionGroup>
  <Accordion title="User — Input">
    End-user messages are processed to detect intent and trigger Tool Calls or Outputs. Metadata can be used to pass structured parameters through the Flow.

    **Parameters:**

    * `type`: Input
    * `text`: What is the weather like in New York today?
    * `data`: `{ "location": "New York", "date": "2024-06-01" }`

    **Example JSON:**

    ```json theme={null}
    {
      "role": "user",
      "type": "input",
      "source": "system",
      "payload": {
        "text": "What is the weather like in New York today?",
        "data": {
          "location": "New York",
          "date": "2024-06-01"
        }
      },
      "id": "62de038b-b1f0-4ca1-b522-93879046342e",
      "traceId": "endpoint-realtimeClient-85035eb7-7271-48ef-aa2d-e01d2a9600bf",
      "timestamp": 1773131963644
    }
    ```
  </Accordion>

  <Accordion title="Assistant — Output">
    Outputs can be displayed directly to the user or guide conditional logic in the Flow based on metadata types.

    **Parameters:**

    * `type`: Output
    * `text`: Your current account balance is \$1,245.67
    * `data`: `{ "type": "responsive" }`

    **Example JSON:**

    ```json theme={null}
    {
      "role": "assistant",
      "type": "output",
      "source": "system",
      "payload": {
        "text": "Your current account balance is $1,245.67",
        "data": {
          "type": "responsive"
        }
      },
      "id": "c8f1f8dc-7b02-493c-890d-0a292806e213",
      "traceId": "endpoint-realtimeClient-85035eb7-7271-48ef-aa2d-e01d2a9600bf",
      "timestamp": 1773131963645
    }
    ```
  </Accordion>

  <Accordion title="Assistant — Tool Call">
    Calls an external tool to fetch data or perform actions. The `id` property links this call to the Tool Answer for later processing.

    **Parameters:**

    * `type`: Tool Call
    * `name`: getWeather
    * `id`: getWeather-123
    * `input`: `{ "location": "New York", "date": "2024-06-01" }`

    **Example JSON:**

    ```json theme={null}
    {
      "role": "assistant",
      "type": "toolCall",
      "source": "system",
      "payload": {
        "name": "getWeather",
        "id": "getWeather-123",
        "input": {
          "location": "New York",
          "date": "2024-06-01"
        }
      },
      "id": "21a2c839-207b-4c25-ad9b-804c9c3dc109",
      "traceId": "endpoint-realtimeClient-85035eb7-7271-48ef-aa2d-e01d2a9600bf",
      "timestamp": 1773131963645
    }
    ```
  </Accordion>

  <Accordion title="Agent — Output">
    Represents internal reasoning or planning. Influences workflow decisions but is generally not shown to the end user.

    **Parameters:**

    * `type`: Output
    * `text`: Verified customer identity and approved the transaction
    * `data`: `{ "type": "responsive" }`

    **Example JSON:**

    ```json theme={null}
    {
      "role": "agent",
      "type": "output",
      "source": "system",
      "payload": {
        "text": "Verified customer identity and approved the transaction",
        "data": {
          "type": "responsive"
        }
      },
      "id": "82bc4c22-be0a-454f-a8d4-1c1f53006971",
      "traceId": "endpoint-realtimeClient-85035eb7-7271-48ef-aa2d-e01d2a9600bf",
      "timestamp": 1773131963646
    }
    ```
  </Accordion>

  <Accordion title="System — Debug Log">
    For monitoring, logging, and troubleshooting AI Agent execution. Helps developers diagnose errors without exposing them to users.

    **Parameters:**

    * `type`: Debug Log
    * `header`: Validation Error
    * `message`: User ID missing in request
    * `metadata`: `{ "severity": "info" }`

    **Example JSON:**

    ```json theme={null}
    {
      "role": "system",
      "type": "debugLog",
      "source": "system",
      "payload": {
        "header": "Validation Error",
        "message": "User ID missing in request",
        "metadata": {
          "severity": "info"
        }
      },
      "id": "663bd605-367f-44f9-858c-93911e05878c",
      "traceId": "endpoint-realtimeClient-85035eb7-7271-48ef-aa2d-e01d2a9600bf",
      "timestamp": 1773131963647
    }
    ```
  </Accordion>

  <Accordion title="Tool — Tool Answer">
    Contains the result of a tool call. The AI Agent can use this output to craft responses or make further decisions in the workflow.

    **Parameters:**

    * `type`: Tool Answer
    * `toolCallId`: getWeather-123
    * `name`: getWeather
    * `content`: `The weather in New York is +10°C`

    **Example JSON:**

    ```json theme={null}
    {
      "role": "tool",
      "type": "toolAnswer",
      "source": "system",
      "payload": {
        "toolCallId": "getWeather-123",
        "name": "getWeather",
        "content": "The weather in New York is +10°C"
      },
      "id": "45a07d1a-f813-48fa-850f-b6c6febcb5bd",
      "traceId": "endpoint-realtimeClient-85035eb7-7271-48ef-aa2d-e01d2a9600bf",
      "timestamp": 1773131963647
    }
    ```
  </Accordion>
</AccordionGroup>

## More Information

* [Get Transcript](/ai/agents/develop/node-reference/service/get-transcript)
* [Agentic AI](/ai/agents/overview)


# Check Agent Availability
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/service/check-agent-availability



<a href="/release-notes/2026.16"><Badge>Updated in 2026.16.0</Badge></a>

<Frame>
  <img alt="Check Agent Availability Node configuration panel" />
</Frame>

## Description

This Node checks the availability of agents with specific skills.
With this Node, you can manage your Flow,
for example, send notifications to the user that the agent is unavailable at the moment,
or suggest an alternative response.

The Check Agent Availability Node supports the following handover providers:

* Cognigy Live Agent
* Genesys Cloud: Guest Chat and Open Messaging
* Chatwoot

## Parameters

| Parameter         | Type     | Description                                                                                                                                                                                                         |
| ----------------- | -------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Handover Provider | Selector | Select the handover provider from the list and [configure the handover accordingly](/ai/escalate/handover-reference/overview). The list contains the handover providers created in **Deploy > Handover Providers**. |

<AccordionGroup>
  <Accordion title="Live Agent">
    | Parameter           | Type          | Description                                                                   |
    | ------------------- | ------------- | ----------------------------------------------------------------------------- |
    | Skills              | CognigyScript | A list of skills that should be used to filter the available human agents.    |
    | Live Agent Inbox Id | CognigyScript | The ID of your Inbox in Live Agent.                                           |
    | Languages           | CognigyScript | A list of languages that should be used to filter the available human agents. |
  </Accordion>

  <Accordion title="Genesys Cloud">
    These settings are applicable to both [Genesys Cloud Guest Chat](/ai/escalate/handover-reference/genesys-cloud-guest-chat) and [Genesys Cloud Open Messaging](/ai/escalate/handover-reference/genesys-cloud-open-messaging).

    | Parameter       | Type | Description                                                                         |
    | --------------- | ---- | ----------------------------------------------------------------------------------- |
    | Skills          | List | A list of skills that should be used to filter the available human agents.          |
    | Profile Skills  | List | A list of profile skills that should be used to filter the available human agents.  |
    | Language Skills | List | A list of language skills that should be used to filter the available human agents. |
  </Accordion>

  <Accordion title="Chatwoot">
    | Parameter         | Type          | Description                       |
    | ----------------- | ------------- | --------------------------------- |
    | Chatwoot Inbox Id | CognigyScript | The ID of your Inbox in Chatwoot. |
  </Accordion>

  <Accordion title="Storage Options">
    | Parameter                     | Type          | Description                                                                                                                                                                               |
    | ----------------------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Where to store the result     | Select        | The storage for the search results. Select a storage:<ul><li>**Input** — the content is stored in the Input object.</li><li>**Context** — the content is stored in the Context.</li></ul> |
    | Input Key to store result     | CognigyScript | The location in the input object where the result is stored. For example, `input.knowledgeSearch`.                                                                                        |
    | Input Context to store result | CognigyScript | The location in the context object where the result is stored. For example, `context.knowledgeSearch`.                                                                                    |
  </Accordion>

  <Accordion title="Error Handling">
    | Parameter                     | Type   | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
    | ----------------------------- | ------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Error Logging                 | Select | You can select one of the Error Logging options:<ul><li>**Don't log errors** — this parameter doesn't allow errors to be logged. It's suitable for environments where error tracking isn't desired, such as production systems, to maintain clean logs.</li><li>**Log error codes and responses** — this parameter records essential information, including error codes such as HTTP status codes, and the responses received from the server. It's useful for basic troubleshooting and understanding error contexts.</li><li>**Log full errors and request payloads** — this parameter provides comprehensive error logging. It not only captures error codes and responses but also logs the complete error messages or details and the entire request payload sent to the server. Be cautious with this option, as it can expose sensitive data. It's most beneficial for in-depth debugging and diagnostics.</li></ul> |
    | Abort Flow Execution on Error | Toggle | Halts Flow Execution on HTTP status codes greater than 299.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
  </Accordion>
</AccordionGroup>

## Example

To build a Flow using the Check Agent Availability Node, follow these steps:

1. In the Flow editor, add a **Check Agent Availability Node** and enter the required values. For example, for the Live Agent handover provider, you need to specify **Live Agent Inbox Id**, **Skills**, and **Languages** that help you sort available agents.
2. Below the Check Agent Availability Node, add an IF Node and configure the **Condition** fields as follows:
   1. Click <img alt="token" /> and select the **Available Agents** Token.
   2. Select **>**.
   3. Enter **0**.
3. Below your **Then** child Node, add a **Handover to Human Agent** Node. Enter the required values for your handover provider. For example, for the Live Agent handover provider, you need to specify **Live Agent Inbox Id**, **Skills**, and **Languages** that help you route the conversation to available agents.
4. Below the **Else** child Node, add a **Say** Node and specify the following text in the **Text** field: `All of our agents are currently assisting other customers. Please try again later`.

To test your Flow, use [Demo Webchat](/webchat/demo).

<Frame>
  <img alt="Check Agent Availability example Flow with If and Handover Nodes" />
</Frame>

## More Information

* [Agent to Handover Node](/ai/escalate/handover-reference/live-agent)
* [IF Node](/ai/agents/develop/node-reference/logic/if)
* [Say Node](/ai/agents/develop/node-reference/basic/say)


# Close Handover
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/service/close-handover



<Frame>
  <img alt="Close Handover Node configuration panel" />
</Frame>

## Description

This Node closes the handover conversation in the handover provider, such as Live Agent. If turned off, the user is only returned to the AI Agent.

If the Close Handover Node isn't active and the human agent resolves a conversation in the Handover Provider later, the Node is ignored and doesn't trigger a Flow.

You can use this Node after the [Set Handover Inactivity](/ai/agents/develop/node-reference/service/set-handover-inactivity) Node.

## More Information

* [User Inactivity Detection and Handling](/ai/escalate/handover-features/user-inactivity-detection)
* [Set Handover Inactivity](/ai/agents/develop/node-reference/service/set-handover-inactivity)
* [Handover Providers](/ai/escalate/handover-reference/overview)
* [Agent Handover](/ai/agents/develop/node-reference/service/handover-to-agent)


# Get Token
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/service/get-token

The Get Token Node triggers Microsoft SSO authentication for an Azure Bot Services OAuth connection.

<a href="/release-notes/2026.17"><Badge>Updated in 2026.17</Badge></a>

<Frame>
  <img alt="Get Token Node configuration panel" />
</Frame>

## Description

The Get Token Node triggers Microsoft Single Sign-On (SSO) authentication for an [Azure Bot Services Endpoint](/ai/agents/deploy/endpoint-reference/azure-bot-services)'s OAuth connection. Depending on whether the user is already signed in, the Get Token Node either completes the sign-in silently or prompts the user to sign in. Use this Node before your Flow needs to call an API that requires the user's permission, such as the Microsoft Graph API.

## Parameters

| Parameter            | Type          | Description                                                                                                                                                                                                                                              |
| -------------------- | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Overwrite Connection | Toggle        | Sets a different OAuth connection than the one configured in the **SSO Connection Name** parameter in the Endpoint settings. Use this toggle if you use multiple SSO connections with your Azure Bot Services. This parameter is deactivated by default. |
| Connection Name      | CognigyScript | This parameter appears when you activate **Overwrite Connection**. Sets the name of the OAuth connection to get the token for.                                                                                                                           |

## More Information

* [Azure Bot Services Endpoint: Multiple SSO Connections](/ai/agents/deploy/endpoint-reference/azure-bot-services#set-up-multiple-sso-connections-optional-accordion-title)
* [Invalidate Token](/ai/agents/develop/node-reference/service/invalidate-token)
* [Token Status](/ai/agents/develop/node-reference/service/token-status)
* [Cognigy Help Center: SSO and Graph API Nodes](https://support.cognigy.com/hc/en-us/articles/6422300559388-Azure-Bot-Services-Advanced-Deployment-with-SSO-Microsoft-Teams-SSO-and-Graph-API-Nodes)


# Get Transcript
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/service/get-transcript



<a href="/release-notes/2025.20"><Badge>Updated in 2025.20</Badge></a>

<Frame>
  <img alt="Get Transcript Node" />
</Frame>

## Description

You can use the Get Transcript Node to retrieve the conversation transcript for further processing. The Get Transcript Node creates a `transcript` object in the [Input object](#storage) that you can use to:

* Access or modify the transcript before passing it to an AI Agent or LLM Prompt Node, for example, to filter out irrelevant messages or prevent LLM hallucinations.
* Send a summary of the interaction between the user and the AI Agent to your CRM system.

<Note>
  The Get Transcript Node can only retrieve limited context and isn't intended for retrieval of full conversations. To retrieve full conversations, use the [OData endpoint](/ai/agents/analyze/odata) instead.
</Note>

## Restrictions

* The Get Transcript Node stores only plain text in the conversation transcript. Structured data, such as [Text with Buttons, Text with Quick Replies, and adaptive cards](/ai/agents/develop/node-reference/basic/say#default-cognigy-channel), images, and videos aren't included in the transcript. However, information from structured data is added to the transcript by default via [the Include Rich Media Context parameter](#advanced). You can access the full transcript in the [Conversations explorer](/insights/explorers/conversations) or with the [OData endpoint](/ai/agents/analyze/odata).
* The transcript is limited to the last 50 turns. This limit helps protect performance and memory across more conversations.

## Parameters

| Parameter | Type   | Description                                                                                                                                                                                                                                                                                                               |
| --------- | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Limit     | Slider | Sets the number of transcript steps to include, up to a maximum of 50. Each step is an input from any participant in the conversation. For example, if you specify `1`, only the last input is included in the transcript. This input can be, for instance, an AI Agent's message if it was the last in the conversation. |

### Advanced

| Parameter                  | Type   | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
| -------------------------- | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| Include Rich Media Context | Toggle | Controls whether the text from the [Textual Description](/ai/agents/develop/node-reference/basic/say#output-type) field in Say, Question, and Optional Questions Nodes is added to the `transcript` object. This text provides AI Agents with additional information, improving their responses.<br /><br />If you have filled in a **Textual Description** field in other Nodes, this text is added to the transcript as context for rich media, such as Text with Quick Replies. If the **Textual Description** field is empty, the button titles and alt text from the rich media are added to the transcript. By default, the **Include Rich Media Context** parameter is active. When this parameter is inactive, nothing related to the rich media is added to the transcript.<br /><br />**Examples**:<ul><li>If **Textual Description** is filled in:<p>Textual Description: `Select your preferred delivery option: Standard Delivery or Express Delivery`.</p><p>Quick Replies buttons: `Standard Delivery`, `Express Delivery`.</p><p>Context added to the transcript: `Select your preferred delivery option: Standard Delivery or Express Delivery`.</p></li><li>If **Textual Description** is empty:<p>Textual Description: empty.</p><p>Quick Replies buttons: `Standard Delivery`, `Express Delivery`.</p><p>Context added to the transcript: `Standard Delivery`, `Express Delivery`.</p></li><li>If **Include Rich Media Context** is inactive:<p>No information related to the rich media is added to the transcript.</p></li></ul> |

### Storage

| Parameter      | Type          | Description                                                                                                                                                       |
| -------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Store Location | Selector      | Sets where to save the result:<ul><li>**Input** — stores the result in the Input object.</li><li>**Context** — stores the result in the Context object.</li></ul> |
| Input Key      | CognigyScript | The parameter appears when **Input** is selected. The result is stored in the `transcript` Input object by default. You can specify another key.                  |
| Context Key    | CognigyScript | The parameter appears when **Context** is selected. The result is stored in the `transcript` Context object by default. You can specify another key.              |

### Provide Conversation Transcripts to AI Agents

The [AI Agent Node](https://claude.ai/ai/agents/develop/node-reference/ai/ai-agent) and the [LLM Prompt Node](/ai/agents/develop/node-reference/service/llm-prompt) (when **Use Single Prompt Mode** in **Advanced** is deactivated) automatically access the conversation transcript without additional setup.

To modify the transcript before it is processed by one of these Nodes, place the Get Transcript Node above them in the Flow editor:

* [AI Agent Node](/ai/agents/develop/node-reference/ai/ai-agent) — retrieve the transcript from the [Input or Context object](#storage) using CognigyScript in the **Job Description** or **Instructions and Context** field. For example, `{{input.transcript}}`.
* [LLM Prompt Node](/ai/agents/develop/node-reference/service/llm-prompt) — retrieve the transcript from the [Input or Context object](#storage) using CognigyScript in the **System Prompt** field. For example, `{{input.transcript}}`.

## More Information

* [Add Transcript Step](/ai/agents/develop/node-reference/service/add-transcript-step)
* [Agentic AI](/ai/agents/overview)


# Handover to Human Agent
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/service/handover-to-agent



<a href="/release-notes/2026.16"><Badge>Updated in 2026.16.0</Badge></a>

<Frame>
  <img alt="Handover to Human Agent Node" />
</Frame>

## Description

The Handover to Human Agent Node enables AI Agents to pass on conversations to human agents in contact centers.
This Node is useful when your AI Agent is unable to handle the complexity of the inquiry
or when a human touch is necessary to address sensitive issues or provide personalized assistance.
Besides adding this Node to the Flow,
create a [handover provider](/ai/escalate/handover-reference/overview) and an [Endpoint](/ai/agents/deploy/endpoint-reference/overview)
and set up the connection to your contact center.

Alternatively, you can use a Question Node to trigger the handover process either when a [specific Intent is detected](/ai/agents/develop/node-reference/basic/question#escalation-intents) or when the [AI Agent misunderstands the user's input](/ai/agents/develop/node-reference/basic/question#escalation-wrong-answers).

### Handover Process

The handover process contains the following steps:

1. **Identifying Edge Cases**. When a user request is too specific for the AI Agent, it can suggest forwarding the user to a human agent without changing channels.
2. **Integration with Agent Handover**. Once the **Handover to Human Agent** Node is triggered, it sends a handover request to the contact center configured in the selected [handover provider](/ai/escalate/handover-reference/overview), allowing a human agent to take over the conversation.
3. **Flow Pausing and Resumption**. When a handover is initiated, the Flow pauses until the end user cancels the request or an agent finishes the conversation. Upon completion, the Flow resumes execution below the Handover to Human Agent Node.
4. **Handover Status**. The handover status can be:

   * **completed** — the agent finished the Handover.
   * **cancelled** — the user canceled the request.
   * **error** — an error occurred during the Handover.

   The object that is exposed has the following structure:

   ```json theme={null}
   {
     "handover": {
       "status": "completed"
     }
   }
   ```

   In case of an `error` status, the handover object is extended to contain additional information:

   ```json theme={null}
   {
     "handover": {
       "status": "completed",
       "error": {
         "reason": "unsupported",
         "message": ""
       }
     }
   }
   ```

In case a channel doesn't support handover, the error reason is marked as `unsupported`. However, if there is an issue when contacting the handover provider, the error reason is marked as `error`.

Use the **Handover Status** [Token](/ai/platform-features/tokens) to access handover status in Flow.

<Frame>
  <img alt="Flow diagram showing handover status handling with a Lookup Node" />
</Frame>

### Handover Status

To configure the AI Agent's behavior based on the handover status,
use a [Lookup](/ai/agents/develop/node-reference/logic/lookup) Node to handle different outcomes.
Select the **Handover Status** option in the Lookup Node settings.

### Handover Providers

For a list of supported Handover services,
refer to the [Handover Providers Overview](/ai/escalate/handover-reference/overview) documentation.

## Parameters

| Parameter                        | Type          | Description                                                                                                                                                                                                         |
| -------------------------------- | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Handover Provider                | Selector      | Select the handover provider from the list and [configure the handover accordingly](/ai/escalate/handover-reference/overview). The list contains the handover providers created in **Deploy > Handover Providers**. |
| Handover Accepted Message        | CognigyScript | The message to be outputted when the handover is requested.                                                                                                                                                         |
| Repeat Handover Accepted Message | Toggle        | Enable to repeatedly output the Handover Accepted Message if the user sends a message while waiting in the queue for an agent.                                                                                      |

<AccordionGroup>
  <Accordion title="Cancel Handover Options">
    | Parameter          | Type          | Description                                                                       |
    | ------------------ | ------------- | --------------------------------------------------------------------------------- |
    | Cancel Intent      | CognigyScript | The Intent the user has to trigger to cancel the Handover request.                |
    | Cancel Button Text | CognigyScript | The text for a quick reply button that cancels the Handover request when clicked. |
  </Accordion>

  <Accordion title="On Resolve Options">
    | Parameter                           | Type     | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
    | ----------------------------------- | -------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Flow Continuation                   | Selector | Controls where the Flow continues after the human agent leaves the conversation.<br />When the user needs assistance completing the Flow with the AI Agent, the conversation is handed over to a human agent. The human agent can trigger the Flow by sending a `/bot` command as a reply in the conversation interface.<br />For example, the user enters their date of birth in an incorrect format, and the AI Agent can't process the response. The conversation is handed over to a human agent, who identifies the incorrect format and sends the corrected date using `/bot MM/DD/YYYY`. The AI Agent takes over and continues the Flow.<br />You can select one of the following options to determine where the Flow continues:<ul><li>**Below this Node** — the Flow continues from the first Node below the Human Agent Node, regardless of where the human agent stopped assisting the user. This option is selected by default.</li> <li>**At current AgentInject Entrypoint** — the Flow continues from the Node where the human agent stopped assisting the user.</li></ul> |
    | Send resolve event to Virtual Agent | Toggle   | Controls whether the Flow receives the `conversation_status_changed` event with the status set to `resolved`. This parameter is activated by default.<ul><li>When the parameter is activated, the event is sent when the human agent leaves the conversation. As a result, the Flow continues immediately, without waiting for the user to send another message. If the [Lookup Case](/ai/agents/develop/node-reference/logic/lookup) Node is configured with **Handover Status** set to **Completed**, the Node can detect the completed handover and continue the Flow.</li><li>When the parameter is deactivated, the event isn't sent when the human agent leaves the conversation. The Flow continues after the user sends their next message.</li></ul>                                                                                                                                                                                                                                                                                                                             |
  </Accordion>

  <Accordion title="Live Agent Settings">
    | Parameter                       | Description                                                                                                                                                                                                                                                                                                                                                                  |
    | ------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Live Agent Inbox Id             | Enter the Inbox ID that you obtained in the [previous step](/ai/escalate/handover-reference/live-agent).                                                                                                                                                                                                                                                                     |
    | Skills                          | Enter a list of skills that should be used to filter the available agents. For example, `technical support, billing`.                                                                                                                                                                                                                                                        |
    | Languages                       | Enter a list of languages that should be used to filter the available agents. For example, `english, spanish`.                                                                                                                                                                                                                                                               |
    | Priority                        | Select a priority for the conversation from the list. For example, `Urgent`.                                                                                                                                                                                                                                                                                                 |
    | Get Queue Position              | Enable the system to retrieve the end user's current queue position. This parameter is inactive by default.                                                                                                                                                                                                                                                                  |
    | Update Interval                 | Specify the interval, in seconds, for how often to receive updates on the queue status. This parameter appears if **Get Queue Position** is selected.                                                                                                                                                                                                                        |
    | Get Estimated Wait Time         | Enable the retrieval of the estimated wait time for the end user. This parameter is inactive by default.                                                                                                                                                                                                                                                                     |
    | Update Interval                 | Set the interval, in seconds, for how often to receive updates on estimated wait time. This parameter appears if **Get Estimated Wait Time** is selected.                                                                                                                                                                                                                    |
    | Alternative Update              | The alternative update options when certain conditions are met. This parameter is inactive by default.                                                                                                                                                                                                                                                                       |
    | Maximum Queue Position          | Set the threshold for the maximum queue position. If this threshold is reached, the alternative text is sent. This parameter appears if **Alternative Update** is selected.                                                                                                                                                                                                  |
    | Maximum Estimated Wait Time     | Define the maximum estimated wait time in milliseconds. If this duration is exceeded, the alternative text is sent. This parameter appears if **Alternative Update** is selected.                                                                                                                                                                                            |
    | Alternative Text                | The message to be sent when either the maximum queue position or maximum wait time is reached. This parameter appears if **Alternative Update** is selected.                                                                                                                                                                                                                 |
    | Enable User Connects Message    | Notify human agents when an end user reconnects to the chat. The parameter is enabled by default. When the parameter is enabled, the message `User joined the conversation` appears in the chat as soon as the end user returns to the chat tab by clicking the **←** (back arrow) at the top bar in the browser, after having opened a new URL on the same tab as the chat. |
    | Enable User Disconnects Message | Notify human agents when an end user disconnects from the chat. The parameter is enabled by default. When the parameter is enabled, the message `User left the conversation` is sent as soon as the end user closes the tab with the chat or switches to a new URL address within the current tab.                                                                           |

    For more information, read the [Live Agent Handover Setup](/live-agent/getting-started/live-agent-setup/live-agent-setup-handover-flow) and [Real-Time Queue Notifications](/ai/escalate/handover-features/real-time-queue-notifications) documentation.
  </Accordion>

  <Accordion title="Salesforce Settings">
    | Parameter           | Description                                                                                                                                                                                                                                                                                                                                                   |
    | ------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Live Agent Base URL | The Base URL should be retrieved from the Salesforce installation. On the **Home** page, navigate to **Feature Settings > Service > Chat > Chat Settings**. On the **Chat Settings** page, go to the **API Information > Chat API Endpoint**. Copy the Endpoint value without the `/chat/rest` part. Example: `https://x.xxx-xx-xxx.salesforceliveagent.com`. |
    | API Version         | The API version value is `52`.                                                                                                                                                                                                                                                                                                                                |
    | Organization ID     | The ID of your organization in Salesforce.                                                                                                                                                                                                                                                                                                                    |
    | Deployment ID       | The ID of your deployment in Salesforce.                                                                                                                                                                                                                                                                                                                      |
    | Button ID           | The ID of your button for the related deployment in Salesforce. For more information, see [Salesforce Handover](/ai/escalate/handover-reference/salesforce).                                                                                                                                                                                                  |

    For more information, read the [Salesforce Handover](/ai/escalate/handover-reference/salesforce) documentation.
  </Accordion>

  <Accordion title="Salesforce MIAW Settings">
    | Setting                                | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
    | -------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Send Transcript as first message       | Allow Salesforce to receive the conversation transcript as the first message. If the transcript exceeds 3000 characters, it is split into multiple messages. This setting is turned off by default.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
    | Conversation Routing Attributes        | Pass custom data from the chat to Salesforce using key-value pairs. The keys are variables you define in the Omni-Channel flow in Salesforce, and the values are the data you want to send for each conversation. To set this up, follow the [Map Pre-Chat Values in Omni-Channel Flow](https://help.salesforce.com/s/articleView?id=service.miaw_map_messaging_2.htm) guide. Enter the data in JSON format. Example: <br />`{"customerName": "John Doe","issueType": "Billing"}`. Here, `customerName` and `issueType` are Salesforce variables, and `John Doe` and `Billing` are the corresponding values sent from the chat. This parameter allows Salesforce flows to automatically use information collected in the chat to route conversations, pre-fill forms, or trigger specific actions. |
    | Identity Token for Authenticated Users | Specify an access token generated via the [Salesforce API](https://developer.salesforce.com/docs/service/messaging-api/references/miaw-api-reference?meta=generateAccessTokenForAuthenticatedUser) to enable secure interactions for authenticated users with Salesforce MIAW. Before using this token, [set up user verification](https://help.salesforce.com/s/articleView?id=sf.user_verification_setup.htm) on the Salesforce side. This setting is off by default; unauthenticated users can still interact.                                                                                                                                                                                                                                                                                  |
    | Enable User Connects Message           | Notify human agents when an end user reconnects to the chat. The parameter is enabled by default. When the parameter is enabled, the message `User joined the conversation` appears in the chat as soon as the end user returns to the chat tab by clicking the **←** (back arrow) at the top bar in the browser, after having opened a new URL on the same tab as the chat.                                                                                                                                                                                                                                                                                                                                                                                                                       |
    | Enable User Disconnects Message        | Notify human agents when an end user disconnects from the chat. The parameter is enabled by default. When the parameter is enabled, the message `User left the conversation` is sent as soon as the end user closes the tab with the chat or switches to a new URL address within the current tab.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
    | Agent Change Notification              | Send a system message to the end user indicating when the conversation is reassigned to another human agent. This parameter is deactivated by default. When the conversation is reassigned, the system message is sent under the human avatar name set in the [Webchat v3 Endpoint settings](/webchat/v3/configuration#webchat-layout).                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | Get Queue Position                     | Enable the system to retrieve the end user's current queue position. This parameter is inactive by default.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
    | Get Estimated Wait Time                | Enable the retrieval of the estimated wait time for the end user. This parameter is inactive by default.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | Alternative Update                     | Enable sending an alternative message to the end user when the maximum queue position is reached. This parameter is inactive by default.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | Maximum Queue Position                 | Set the threshold for the maximum queue position. If this threshold is reached, the alternative text is sent instead of the actual queue position. This parameter appears if **Alternative Update** is selected.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
    | Alternative Text                       | Specify the message to be sent when the maximum queue position is reached. This parameter appears if **Alternative Update** is selected.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |

    For more information, read the [Salesforce Handover](/ai/escalate/handover-reference/salesforce-miaw) documentation.
  </Accordion>

  <Accordion title="NiCE CXone Settings">
    | Setting                          | Description                                                                                                                                                                            |
    | -------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Send Transcript as first message | Allow NiCE CXone to receive the conversation transcript as the first message when the handover begins. This setting is turned on by default.                                           |
    | Display Agent Details            | Display the human agent's name and avatar from NiCE CXone in the chat for the end user. This parameter is disabled by default.                                                         |
    | Custom Attributes                | Add custom attributes in JSON format to include additional information the handover. For more information, refer to [NiCE CXone](/ai/escalate/handover-reference/nice-cxone/overview). |

    Example for the **Custom Attributes** parameter:

    ```json theme={null}
    {
      "section": "section A",
      "region": "Europe"
    }
    ```
  </Accordion>

  <Accordion title="Genesys Settings">
    | Parameter                       | Description                                                                                                                                                                                                                                                                                                                                                                  |
    | ------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Language                        | Specify a language for the conversation. For example, `english`, `spanish`, `german`.                                                                                                                                                                                                                                                                                        |
    | Skills                          | Define skills for the conversation. For example, `escalation`.                                                                                                                                                                                                                                                                                                               |
    | Priority                        | Set the priority for the conversation. For example, `1`. <br /> If a priority is set, it triggers a flow in Genesys to prioritize or de-prioritize the conversation within the queue. <br /> Note that this functionality requires the appropriate flow to be set up in Genesys.                                                                                             |
    | Enable User Connects Message    | Notify human agents when an end user reconnects to the chat. The parameter is enabled by default. When the parameter is enabled, the message `User joined the conversation` appears in the chat as soon as the end user returns to the chat tab by clicking the **←** (back arrow) at the top bar in the browser, after having opened a new URL on the same tab as the chat. |
    | Enable User Disconnects Message | Notify human agents when an end user disconnects from the chat. The parameter is enabled by default. When the parameter is enabled, the message `User left the conversation` is sent as soon as the end user closes the tab with the chat or switches to a new URL address within the current tab.                                                                           |
    | Display Agent Details           | Display the human agent's name and avatar from Genesys in the chat for the end user. This parameter is disabled by default.                                                                                                                                                                                                                                                  |
    | Custom Attributes               | Add custom attributes in JSON format to include additional information the handover. For more information, refer to [Genesys Cloud Open Messaging](/ai/escalate/handover-reference/genesys-cloud-open-messaging).                                                                                                                                                            |
    | Get Estimated Wait Time         | Retrieve the estimated wait time for the end user. End users then receive wait time updates while they wait to be connected to a human agent. This parameter is inactive by default. It requires an active [Amazon EventBridge Source integration](/ai/escalate/handover-reference/genesys-cloud-open-messaging-with-amazon-eventbridge).                                    |
    | Update Interval                 | Specify the interval, in seconds, for how often to receive updates on the estimated wait time. This parameter appears if **Get Estimated Wait Time** is selected.                                                                                                                                                                                                            |
    | Alternative Update              | Send an alternative message to the end user when the maximum estimated wait time is exceeded. This parameter is inactive by default. This parameter appears if **Get Estimated Wait Time** is selected.                                                                                                                                                                      |
    | Maximum Estimated Wait Time     | Define the maximum estimated wait time in milliseconds. If this duration is exceeded, the alternative text is sent instead of the actual wait time. This parameter appears if **Alternative Update** is selected.                                                                                                                                                            |
    | Alternative Text                | Specify the message to be sent when the maximum estimated wait time is exceeded. This parameter appears if **Alternative Update** is selected.                                                                                                                                                                                                                               |

    Genesys Cloud Open Messaging doesn't provide a queue position. Estimated wait time is the only queue parameter available.

    Example for the **Custom Attributes** parameter:

    ```json theme={null}
    {
      "section": "section A",
      "region": "Europe"
    }
    ```

    For more information,
    read
    the [Genesys Cloud Guest Chat](/ai/escalate/handover-reference/genesys-cloud-guest-chat#prerequisites), [Genesys Cloud Open Messaging](/ai/escalate/handover-reference/genesys-cloud-open-messaging), and [Real-Time Queue Notifications](/ai/escalate/handover-features/real-time-queue-notifications#configure-settings-in-genesys-cloud-open-messaging) documentation.
  </Accordion>

  <Accordion title="8x8 Settings">
    | Parameter           | Description                                                                                                                                                                                                                                                                      |
    | ------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | 8x8 Channel Id      | Specify the channel ID that you created during the 8x8 configuration. <br /> You can find this ID in the 8x8 admin console. Navigate to **Channels > Chat API**. In the channel list, find the channel that you created. Copy the **ID** from the channel row.                   |
    | 8x8 Queue Id        | Specify the header value required for authenticating API calls to the 8x8 system. <br /> You can find this ID in the 8x8 admin console. Navigate to **Channels > Chat API**. In the channel list, find the channel that you created. Copy the **Queue ID** from the channel row. |
    | 8x8 JSON properties | Extend 8x8 properties with custom JSON.                                                                                                                                                                                                                                          |

    For more information, read the [8x8 Handover Setup](/ai/escalate/handover-reference/8x8) documentation.
  </Accordion>

  <Accordion title="RingCentral Engage Settings">
    | Parameter                       | Type   | Description                                                                                                                                                                                                                                                                                                                                                                  |
    | ------------------------------- | ------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Additional Category Ids         |        | IDs to be passed to the RingCentral Engage handover provider.                                                                                                                                                                                                                                                                                                                |
    | Enable User Connects Message    | Toggle | Notify human agents when an end user reconnects to the chat. The parameter is enabled by default. When the parameter is enabled, the message `User joined the conversation` appears in the chat as soon as the end user returns to the chat tab by clicking the **←** (back arrow) at the top bar in the browser, after having opened a new URL on the same tab as the chat. |
    | Enable User Disconnects Message | Toggle | Notify human agents when an end user disconnects from the chat. The parameter is enabled by default. When the parameter is enabled, the message `User left the conversation` is sent as soon as the end user closes the tab with the chat or switches to a new URL address within the current tab.                                                                           |

    For more information, read the [RingCentral Engage Integration](/ai/escalate/handover-reference/ring-central-engage) documentation.
  </Accordion>

  <Accordion title="Event Settings">
    | Parameter         | Type   | Description                                             |
    | ----------------- | ------ | ------------------------------------------------------- |
    | Send Queue Event  | Toggle | Sends an event to the Flow once the Handover is queued. |
    | Send Active Event | Toggle | Sends an event to the Flow once the Handover is active. |
  </Accordion>
</AccordionGroup>

## More Information

* [Handover Providers](/ai/escalate/handover-reference/overview)
* [Check Agent Availability Node](/ai/agents/develop/node-reference/service/check-agent-availability)
* [User Inactivity Detection and Handling](/ai/escalate/handover-features/user-inactivity-detection)


# HTTP Request
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/service/http-request



<a href="/release-notes/4.84"><Badge>Updated in 4.84</Badge></a>

<Frame>
  <img alt="HTTP Request Node configuration panel" />
</Frame>

## Description

The HTTP Request Node lets you make an HTTP request within a Flow. When triggered, the HTTP Request Node performs the request to the specified URL during Flow execution.

Using the HTTP Request Node, you can manage Cognigy resources by creating, updating, and deleting them via the Cognigy API.

You can also use external APIs, such as Google Maps for location data, OpenWeatherMap for weather updates, and NewsAPI for the latest news, or any other external API to enhance chat or voice conversations.
If you use an external API, make sure the server has a public certificate to [avoid connection and secure issues](#handling-certificate-errors).

## Limitations

The default timeout is 15 seconds in a shared SaaS installation and 8 seconds in dedicated SaaS or on-premises installations. You can change this limit only for the following installation types:

* Dedicated SaaS. You can request an increase up to a maximum of 15 seconds by contacting Cognigy technical support.
* On-premises. You can configure the timeout using the `HTTP_NODE_TIMEOUT_IN_SECONDS` environment variable. Setting this value higher than 15 seconds may cause delays and affect AI Agents' performance. Use longer timeouts at your own risk.

## Parameters

## HTTP Request Methods

<AccordionGroup>
  <Accordion title="GET">
    The method retrieves data from the server.

    The results of the GET request are stored in the context of the Flow. You can retrieve the requested data of your GET request by accessing the context with the key you defined in the HTTP Node settings.
  </Accordion>

  <Accordion title="POST, PATCH, and PUT">
    The methods are used for the following purposes:

    * `POST` – submits data to the server to create a new resource.
    * `PATCH` – updates partial data of an existing resource.
    * `PUT` – replaces an existing resource with new data.

    The standard Content-Type header is `application/x-www-form-urlencoded`.
    To use a different Content-Type, set the header value explicitly or use JSON.

    For URL-encoded data, you don't need a specific header.
    URL encoding replaces unsafe characters in URLs with a percent sign (`%`) followed by two hexadecimal digits representing the ASCII code of the character.
    This approach ensures the data is transmitted correctly over the web.

    For example, you can send the data `+49555262626` and `https://handler.twilio.com/twiml/EHf9b7af093c31b5baa1414be891` as a URL-encoded payload:

    ```txt theme={null}
    To=%2B49555262626&Url=https%3A%2F%2Fhandler.twilio.com%2Ftwiml%2FEHf9b7af093c31b5baa1414be891
    ```

    where:

    Each key-value pair is separated by `&`, and each key is separated from its value by `=`:

    * `To=%2B49555262626`:
      * `To` is the key.
      * `%2B49555262626` is the URL-encoded value.
        * `%2B` represents the `+` sign.
        * `49555262626` is the actual phone number.
    * `Url=https%3A%2F%2Fhandler.twilio.com%2Ftwiml%2FEHf9b7af093c31b5baa1414be891`:
      * `Url` is the key.
      * `https%3A%2F%2Fhandler.twilio.com%2Ftwiml%2FEHf9b7af093c31b5baa1414be891` is the URL-encoded value.
        * `%3A` represents `:`.
        * `%2F` represents `/`.
        * The rest of the value is the URL of a Twilio handler endpoint.

    **Payload**

    Define the payload of your `POST`, `PATCH`, or `PUT` request.

    Select one of the following payload formats:

    * **JSON** — consists of key-value pairs. When using a JSON payload to access Cognigy objects such as Input, Context, or Profile, refer to [CognigyScript](/ai/platform-features/cognigyscript) for an example payload.
    * **Text** — consists of unstructured text or basic key-value pairs. For example, `username=JohnDoe&password=12345`.
    * **Form-Data** — consists of key-value pairs. It's used for submitting form data, including file uploads. Each key-value pair is sent as a separate part of the request. The key-value pairs must have the text type. Data isn't supported.

    Inspect the response on the **INFO** tab in the Interaction Panel.
  </Accordion>

  <Accordion title="DELETE">
    The method deletes a resource from the server.

    The `DELETE` request configuration prompt exposes the General Configuration fields.
  </Accordion>
</AccordionGroup>

### URL

The URL to the targeted resource is `https://api-endpoint.com/resource`.

Note that Cognigy expects a URL to the targeted resource in its unencoded form. Decode any encoded URL to make sure the HTTP request can be executed successfully. For more information, see [URL encoding (on en:WP)](https://en.wikipedia.org/wiki/URL%20encoding).

<AccordionGroup>
  <Accordion title="Storage Options">
    Define the context key where you want to store the response from the executed HTTP request.
    This field is required and needs to have a valid value.

    After the HTTP request is successfully executed, you can access the response payload by executing the following [CognigyScript](/ai/platform-features/cognigyscript):

    ```txt theme={null}
    {{ context.<yourContextStore> }}
    ```
  </Accordion>

  <Accordion title="Headers">
    The headers are information sent along with the request, such as content type or authentication tokens.

    Add the headers to your HTTP request. All headers are listed in one JSON object, which should have the following format:

    ```json theme={null}
    {
      "header-name": "header value",
      "another-header-name": "another header value"
    }
    ```

    For example:

    ```json theme={null}
    {
      "Content-Type": "application/json",
      "Accept": "application/json",
    }
    ```

    where:

    * `Content-Type` — specifies the media type of the resource being sent to the server.
    * `Accept` — indicates the media types that the client wants to receive from the server.

    Alternatively, you can click **Show JSON Editor** to switch between the list view and the JSON editor.

    You can also enable capturing the response headers along with the request body in the specified storage location.
  </Accordion>

  <Accordion title="Authentication">
    The supported authentication methods are:

    * **No Auth** — no authentication is required to access the resource.
    * **Basic Auth** — authentication using a username and password encoded in base64.
    * **OAuth2** — a framework for authorization using tokens to securely manage access.
    * **API Key (Authorization: ApiKey)** — authentication using a key sent in the `Authorization` header as API Key.
    * **API Key (X-API-Key)** — authentication using a key sent in the `X-API-Key` header.

    Authentication makes use of [Connections](/ai/platform-features/connections), which means that the actual authentication information can be encrypted.

    When a new authentication connection is created by clicking **+** next to the **Parameters** field,
    the **New Connection** window appears to request the details specific to the authentication type.

    If you select an authorization type other than **No Auth**, additional fields related to the selected authorization type are provided.
  </Accordion>

  <Accordion title="Execution & Caching">
    | Parameter                       | Type   | Description                                                                                                                                                                                                          |
    | ------------------------------- | ------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Execute Requests Asynchronously | Toggle | When you enable the **Execute Requests asynchronously** parameter, the HTTP Request Node executes the request without waiting for a response. This means that the execution of the Flow continues without any delay. |
    | Timeout                         | Number | The timeout for canceling the request (in milliseconds). The default value is `8000`. This setting overrides the [global parameter value](#limitations) for your Node.                                               |
    | Retry attempts                  | Number | The number of retry attempts in case of an error.                                                                                                                                                                    |
    | Cache Results                   | Toggle | When the setting is enabled, the HTTP Node caches the responses.                                                                                                                                                     |
    | Cache expiry time in seconds    | Number | The time in seconds after which the cached response expires.                                                                                                                                                         |
  </Accordion>

  <Accordion title="Security">
    **Allow Insecure SSL**

    By default, Cognigy rejects connecting against insecure SSL endpoints, such as URLs with no or self-signed certificates. Activating this parameter allows the Node to connect against these URLs as well.
  </Accordion>

  <Accordion title="Error Handling & Logging">
    **Error Logging**

    If HTTP requests return a status code greater than `299`, the response is considered an error.
    You can select one of the following options:

    * **Not log the error** — errors aren't recorded in the logs.
    * **Log the error with the server response and the URL** — errors are logged along with the server's response and the URL of the request.
    * **Log the error, the response, the URL, and the request payload** — errors, the server's response, the URL, and the request payload are logged. Selecting this option could expose sensitive data in the logs, so use it with caution.

    **Abort Flow Execution on Error**

    When the setting is enabled, the Flow stops if the response has a status code greater than `299`.

    **Debug Mode Logging**

    Select what information
    to log to [Debug Mode](/ai/agents/test/interaction-panel/chat#debug-mode) when using the Interaction Panel:

    * **Do not log Request & Response** — don't log any request or response data to the Interaction Panel. This option is activated by default.
    * **Request** — log only the request data sent to the server.
    * **Response** — log only the response data received from the server.
    * **Request & Response** — log both the request data sent to the server and the response data received from the server.
  </Accordion>
</AccordionGroup>

## Troubleshooting

### Handling Certificate Errors

When using the HTTP Request Node with a domain that doesn't have a public certificate, errors can occur due to a missing Intermediate Certificate. An example of such an error is `UNABLE_TO_VERIFY_LEAF_SIGNATURE`. If there is no trusted Certificate Authority or a complete certificate chain, the request fails. Note that you can't add a custom Certificate Authority or certificate chain to the Cognigy platform. Your options are:

* Fix the certificate chain by obtaining a certificate from a trusted Certificate Authority or adding the missing Intermediate Certificate to your domain's API services.
* Deactivate certificate verification by enabling the Allow Insecure SSL setting in the Node. However, this option isn't recommended.


# Invalidate Token
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/service/invalidate-token

The Invalidate Token Node signs the user out of an Azure Bot Services OAuth connection and deletes the stored token.

<a href="/release-notes/2026.17"><Badge>Updated in 2026.17</Badge></a>

<Frame>
  <img alt="Invalidate Token Node configuration panel" />
</Frame>

## Description

The Invalidate Token Node signs the user out of an [Azure Bot Services Endpoint](/ai/agents/deploy/endpoint-reference/azure-bot-services)'s OAuth connection and deletes the stored token. Use this Node to let the user log out, or to force a new sign-in before issuing a new token for the same connection.

## Parameters

| Parameter            | Type          | Description                                                                                                                                                                                                                                              |
| -------------------- | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Overwrite Connection | Toggle        | Sets a different OAuth connection than the one configured in the **SSO Connection Name** parameter in the Endpoint settings. Use this toggle if you use multiple SSO connections with your Azure Bot Services. This parameter is deactivated by default. |
| Connection Name      | CognigyScript | This parameter appears when you activate **Overwrite Connection**. Sets the name of the OAuth connection to revoke the token for.                                                                                                                        |

## More Information

* [Azure Bot Services Endpoint: Multiple SSO Connections](/ai/agents/deploy/endpoint-reference/azure-bot-services#set-up-multiple-sso-connections-optional-accordion-title)
* [Get Token](/ai/agents/develop/node-reference/service/get-token)
* [Token Status](/ai/agents/develop/node-reference/service/token-status)
* [Cognigy Help Center: SSO and Graph API Nodes](https://support.cognigy.com/hc/en-us/articles/6422300559388-Azure-Bot-Services-Advanced-Deployment-with-SSO-Microsoft-Teams-SSO-and-Graph-API-Nodes)


# LLM Prompt
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/service/llm-prompt

The Cognigy LLM Prompt Node allows using Generative AI for creating relevant content. This Node supports Tools, including MCP tools, image recognition, and conversation transcripts.

<a href="/release-notes/2026.18"><Badge>Updated in 2026.18</Badge></a>

<Frame>
  <img alt="LLM Prompt Node" />
</Frame>

## Description

<Warning>
  As of Cognigy 2026.18.0, the following parameters are deprecated:

  * **Sampling Method**
  * **Temperature**
  * **Top Percentage**
  * **Maximal Tokens**
  * **Frequency Penalty**
  * **Presence Penalty**
  * **Use Stops**
  * **Stops**
  * **Seed**

  The removal is planned for April 2027. For models that don't support these parameters, activate the **Ignore deprecated parameters** toggle to exclude them from the request to the LLM provider. Otherwise, the request to the LLM provider might fail. If you plan to use these parameters, migrate them to the [Custom Options section](#custom-options) and activate the **Ignore deprecated parameters** toggle. Otherwise, Custom Options might not override them for every provider and model combination.

  For example, OpenAI models such as `gpt-5`, `gpt-5.1`, and `gpt-5.2` don't support these parameters. Similar models from other LLM providers might also not support these parameters. Check your LLM provider's documentation for model-specific support.
</Warning>

The LLM Prompt Node lets you configure the system prompt for your LLM calls, generating both text and structured content. With this Node, you can also:

* Include tools in your LLM calls.
* Enable image handling.
* Set advanced LLM request options.

Additionally, you can refer to your AI Agent configuration using the [Load AI Agent Node](/ai/agents/develop/node-reference/ai/load-ai-agent).

Before using this Node, set the LLM provider in the [Settings](/ai/agents/develop/gen-ai-and-llms/generative-ai).

## Limitations

* The **Description** and **Enum** fields in the **Parameters** section of an [AI Agent tool](#parameters) don't support CognigyScript.

## Parameters

### Parent Node

<AccordionGroup>
  <Accordion title="Large Language Model">
    The selected **Default** model is the model that you specified in **Settings > Generative AI Settings** of your Project.

    You can select a different model from the list or override the selected model using the Custom Model Options parameter.
  </Accordion>

  <Accordion title="System Prompt">
    The system prompt is the message sent to the LLM to guide its responses. The parameter supports CognigyScript, allowing dynamic content and logic within the prompt. This prompt can work either as the input for completion tasks or as the system message in chat-based interactions, setting context and behavior for the AI Agent.

    Additionally, you can inject the recent conversation into the **System Prompt** field by using these tags:

    * `@cognigyRecentConversation` — the tag is replaced with a string that can contain up to 10 recent AI Agent and 10 user outputs, for example:
      ```text theme={null}
      Bot: agentOutput1
      User: userOutput1
      Bot: agentOutput2
      User: userOutput2
      ```
    * `@cognigyRecentUserInputs` — the tag is replaced with a string that can contain up to 10 recent user outputs, for example:

      ```text theme={null}
      User: userOutput1
      User: userOutput2
      ```

    If you want to access only the last user input, specify the `Text` token in the **System Prompt** field.

    When adding a tag, ensure that you leave a line break before and after the tag, for example:

    ```text theme={null}
    A user had a conversation with a chatbot. The conversation history so far is:
    @cognigyRecentConversation

    Describe the user sentiment in one very short line.
    ```

    Both tags can include an optional turn limit parameter, which is appended to the tag.

    Examples:

    ```typescript theme={null}
    @cognigyRecentConversation:3 // returns the last 3 turns of the conversation.
    @cognigyRecentUserInputs:2 // returns the last 2 user inputs.
    ```
  </Accordion>

  <Accordion title="Advanced">
    | Parameter                    | Type   | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
    | ---------------------------- | ------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Ignore deprecated parameters | Toggle | Excludes the deprecated **Sampling Method**, **Temperature**, **Top Percentage**, **Presence Penalty**, **Frequency Penalty**, **Use Stops**, **Stops**, and **Seed** parameters from the request to the LLM provider. Activate this toggle if your model doesn't support these parameters. If you activate this toggle for a model that supports them, the model uses default values.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
    | Maximal Tokens               | Slider | The maximum number of tokens to generate in the completion.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
    | Use Single Prompt Mode       | Toggle | Sends a prompt from the **System Prompt** parameter to the model without any conversation context. The **Use Single Prompt Mode** parameter is deactivated by default and doesn't support multi-turn conversations, making it suitable for simple, one-off completions. When the parameter is activated, the **System Prompt** parameter must not be empty. Empty prompts return no data and may produce empty responses or errors depending on the model you use.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | Transcript Turns             | Slider | The number of previous conversation turns included in the LLM request. A conversation turn includes a user input and the AI Agent’s response. The default value is `50`.<br />This parameter affects only past turns. The current turn is always included, containing the user input, LLM processing, tool calls, tool results, and multi-step tool workflows.<br />For example, setting **Transcript Turns** to `0` excludes all previous turns, while the current turn is still fully included.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | Response Format              | Select | Choose the format for the model's output result. You can select one of the following options:<ul><li>**None** — no response format is specified, or do not request with an LLM provider that does not accept any response format or does not support it or could use provider's default in-built response format. This option is selected by default.</li><li>**Text** — the model returns messages in text format.</li><li>**JSON Object** — the model returns messages in JSON format. To use this option, instruct the model to generate JSON via a system or user message in the **Instruction (System Message/Prompt)** field. For example, `Take the user message and return it as JSON in the following format {"message": "THE_MESSAGE"}`. Note that this parameter may not be supported by all LLMs. For more information, refer to the LLM provider's API documentation.</li></ul>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
    | Timeout                      | Number | The maximum number of milliseconds to wait for a response from the LLM provider.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | Sampling Method              | Select | Methods:<ul><li>**Temperature** — determines the level of randomness in the generated text. A higher temperature allows for more diverse and creative outputs, while a lower temperature leads to more predictable and consistent outputs with the training data.</li><li>**Top Percentage** — specifies the percentage of the most probable outputs for generation, resulting in more consistent output.</li></ul>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
    | Temperature                  | Slider | Define the sampling temperature, which ranges between 0 and 1. Higher values, such as 0.8, make the output more random, while lower values, such as 0.2, make it more focused and deterministic.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | Top Percentage               | Slider | Control the Top-p (nucleus) sampling, ranging from 0 to 1. Higher values allow more diverse word choices, while lower values make the output more focused. For example, 0.9 means the model selects from the smallest set of words with a combined probability of 90%.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
    | Presence Penalty             | Slider | Number between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood of talking about new topics.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | Frequency Penalty            | Slider | Number between -2.0 and 2.0. The penalty assigns a lower probability to tokens frequently appearing in the generated text, encouraging the model to generate more diverse and unique content.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
    | Use Stops                    | Toggle | Whether to use a list of stop words to let Generative AI know where the sentence stops.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
    | Stops                        | Text   | Up to 4 sequences where the API will stop generating further tokens. The returned text will not contain the stop sequence.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
    | Seed                         | Number | Use this parameter for consistent output when referring to the same LLM Prompt Node multiple times. Specify any integer number, for example, `123`. The number in the Seed field and the prompt in the **Instruction (System Message/Prompt)** field should remain unchanged during subsequent references to the Node.<br />Note that in OpenAI, this parameter is in [Beta](https://platform.openai.com/docs/api-reference/chat/create#chat-create-seed) and is supported only by [certain models](https://cookbook.openai.com/examples/reproducible_outputs_with_the_seed_parameter).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
    | Include Rich Media Context   | Toggle | Controls whether the text from the [Textual Description](/ai/agents/develop/node-reference/basic/say#output-type) field in Say, Question, and Optional Questions Nodes is added to the AI Agent's prompt. This text provides AI Agents with additional information, improving their responses.<br /><br />If you have filled in a **Textual Description** field in other Nodes, this text is added to the prompt as context for rich media, such as Text with Quick Replies. If the **Textual Description** field is empty, the button titles and alt text from the rich media are added to the prompt. By default, the **Include Rich Media Context** parameter is active. When this parameter is inactive, nothing related to the rich media is added to the prompt.<br /><br />**Examples**:<ul><li>If **Textual Description** is filled in:<p>Textual Description: `Select your preferred delivery option: Standard Delivery or Express Delivery`.</p><p>Quick Replies buttons: `Standard Delivery`, `Express Delivery`.</p><p>Context added to the prompt: `Select your preferred delivery option: Standard Delivery or Express Delivery`.</p></li><li>If **Textual Description** is empty:<p>Textual Description: empty.</p><p>Quick Replies buttons: `Standard Delivery`, `Express Delivery`.</p><p>Context added to the prompt: `Standard Delivery`, `Express Delivery`.</p></li><li>If **Include Rich Media Context** is inactive:<p>No information related to the rich media is added to the prompt.</p></li></ul> |
  </Accordion>

  <Accordion title="Storage & Streaming Options">
    | Parameter                     | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
    | ----------------------------- | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | How to handle the result      | Select        | Determine how to handle the prompt result:<ul><li>**Store in Input** — stores the result in the Input object. To print the prompt result, use the LLM Prompt Result Token in the Say Node or enable the **Output result immediately** option.</li><li>**Store in Context** — stores the result in the Context object. To print the prompt result, use the LLM Prompt Result Token in the Say Node or enable the **Output result immediately** option.</li><li>**Stream to Output** — streams the result directly into the output. This means that the model provides prompts directly into the conversation chat, as soon as a Stream Buffer Flush Token is matched, and you don't need to use the LLM Prompt Result token and Say Node. By default, this result won't be stored in either the Input or the Context. You can change this behavior by activating the **Store Copy in Input** option. Note that streaming may not be supported by all [Cognigy LLM Prompt Node](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) providers, such as Google<sup>[1](#footnote1)</sup>. If streaming is not supported, enabling the **Store Copy in Input** option will save the result to the Input object.</li></ul> |
    | Input Key to store Result     | CognigyScript | The parameter appears when **Store in Input** is selected. The result is stored in the `promptResult` Input object by default. You can specify another key.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
    | Context Key to store Result   | CognigyScript | The parameter appears when **Store in Context** is selected. The result is stored in the `promptResult` Context object by default. You can specify another key.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
    | Stream Buffer Flush Tokens    | Text Array    | The parameter appears when **Stream to Output** is selected. It defines tokens that trigger the stream buffer to flush to the output. The tokens can be punctuation marks or symbols, such as `\n`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | Stream Buffer Flush Overrides | Text Array    | The parameter appears when **Stream to Output** is selected. It allows using regular expressions (without leading or trailing slashes) to control stream buffer flushing. A trailing `$` is automatically added to match patterns at the end of the buffer. For example, `\d+\.` checks for a number followed by a dot at the end of the string.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
    | Output result immediately     | Toggle        | The parameter appears when you select either **Store in Input** or **Store in Context**. This parameter allows you to output results immediately without using the Say Node and LLM Prompt token.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |
    | Store Detailed Results        | Toggle        | The parameter appears when you select either **Store in Input** or **Store in Context**, or when you enable **Store Copy in Input**. This parameter allows you to save detailed results of the LLM's generated output. By default, the result is stored in the `promptResult` object. You can specify another value in the **Context Key for the Result** field to save it in the Context object, or in the **Input Key for the Result** to save it in the Input object. The object contains keys such as `result`, `finishReason`, and `usage`. It may also include `detailedResult` if completion models are used, as well as `firstChunk` and `lastChunk` in some streaming results, depending on the LLM provider.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         |
    | Store Copy in Input           | Toggle        | The parameter appears when **Stream to Output** is selected. In addition to streaming the result to the output, store a copy in the Input object by specifying a value in the **Input Key to store Result** field.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |

    **Streaming Results**

    In Stream mode, LLM generates tokens and returns them one by one to Cognigy to ensure low-latency responses.
    Cognigy monitors delimiter tokens (Stream Buffer Flush Tokens),
    which serve as markers indicating when to output the token buffer.
    These tokens could be `.`, `!`, `?`, or any other symbols that act as delimiters for complete logical statements.
    When Cognigy detects one of these tokens, it promptly flushes the token buffer into the voice or text chat.

    The preconfigured overrides are listed in the table.

    | Regex                                       | Description                                | Example        |
    | ------------------------------------------- | ------------------------------------------ | -------------- |
    | `\d+\.`                                     | A number followed by a dot.                | `26.08`        |
    | `\b(?:Dr\|Ms\|Mr\|Mrs\|Prof\|Sr\|Jr\|ca)\.` | Common abbreviations followed by a dot.    | `Mr.`          |
    | `\b[A-Z]\.`                                 | A single capital letter followed by a dot. | `M. Smith`     |
    | `\.\.\.`                                    | Three dots used for omission.              | `...`          |
    | `\b.\..\.`                                  | Two-letter abbreviations.                  | `i.e.`, `e.g.` |
  </Accordion>

  <Accordion title="Tool Settings">
    The process of setting up a tool is the same as for the AI Agent Node. See the example in the [AI Agent Tool Settings](/ai/agents/develop/node-reference/ai/ai-agent#child-nodes) section.

    | Parameter       | Type     | Description                                                                                                                                                                                                                                                                                                                         |
    | --------------- | -------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Tool Choice     | Selector | If supported by your LLM Model, this will determine how tools should be selected by the AI Agent:<ul><li>**Auto** — tools (or none) are automatically selected by the AI Agent when needed.</li><li>**Required** — your AI Agent will always use one of its Tools.</li><li>**None** — your AI Agent won't use a tool.</li></ul>     |
    | Use Strict mode | Toggle   | When the parameter is enabled, strict mode (if supported by the LLM provider) ensures that the arguments passed to a tool call precisely match the expected parameters. Enabling this feature can help prevent errors. However, it may cause a slight delay in the response, especially during the first call after making changes. |
  </Accordion>

  <Accordion title="Image Handling">
    | Parameter            | Type     | Description                                                                                                                                                                                                                                                                                                                              |
    | -------------------- | -------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Process Images       | Toggle   | Enable the AI Agent to read and understand image attachments. Make sure that your LLM provider supports image processing; refer to your provider's documentation. In addition, make sure that attachments are supported by and activated in your Endpoint, for example, Webchat.                                                         |
    | Images in Transcript | Selector | Configure how images older than the last turn are handled to reduce token usage:<ul><li>**Minify** — reduces the size of these images to 512x512px.</li><li>**Drop** — excludes the images.</li><li>**Keep** — sends the max size (this option consumes more tokens).</li></ul>Limitations and token consumption depend on the LLM used. |
  </Accordion>

  <Accordion title="Error Handling">
    | Parameter                      | Type   | Description                                                                                                                                                                                                                                                                                                                                                                                                              |
    | ------------------------------ | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Log to System Logs             | Toggle | Log errors to the system logs. They can be viewed on the [Logs](/ai/agents/test/logs) page of your Project. The parameter is inactive by default.                                                                                                                                                                                                                                                                        |
    | Select Error Handling Approach | Select | You can select one of the Error Handling options:<ul><li>**Stop Flow Execution** — terminate the current Flow execution.</li><li>**Continue Flow Execution** — allow the Flow to continue executing, bypassing the error and proceeding to the next steps.</li><li>**Go to Node** — redirect the workflow to a specific Node in the Flow, which can be useful for error recovery or customized error handling.</li></ul> |
    | Error Message (optional)       | Text   | The parameter appears when **Continue Flow Execution** is selected. Add a message to output if the LLM Prompt Node fails.                                                                                                                                                                                                                                                                                                |
    | Select Flow                    | Select | The parameter appears when **Go to Node** is selected. Select a Flow from the available options.                                                                                                                                                                                                                                                                                                                         |
    | Select Node                    | Select | The parameter appears when **Go to Node** is selected. Select a Node from the available options.                                                                                                                                                                                                                                                                                                                         |
  </Accordion>

  <Accordion title="Custom Options">
    These options let you use parameters that aren't included in the LLM Prompt Node or overwrite set configurations.

    | Parameter              | Type | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          | Example                                                                      |
    | ---------------------- | ---- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ---------------------------------------------------------------------------- |
    | Custom Options         | JSON | Set custom parameters to the LLM configuration. You can set individual parameters as well as entire functions. These parameters customize the behavior of the model, such as adjusting `temperature`, `top_k`, or `presence_penalty`. This field overrides the parameters set in the LLM Prompt Node. To view the full list of available parameters for your model, refer to the LLM provider's API documentation, for example, [OpenAI](https://platform.openai.com/docs/api-reference/chat/create) or [Azure OpenAI](https://learn.microsoft.com/en-us/azure/ai-services/openai/). | `{ "model": "gpt-4o", "temperature": 0.7, "max_tokens": 4000 }`              |
    | Custom Request Options | JSON | Set custom parameters to customize the request to the LLM provider. These parameters override the configurations set in this Node. For example, setting a `{ "stream": true }` parameter overrides the **Output result immediately** option. For more information, refer to the LLM provider's API documentation.                                                                                                                                                                                                                                                                    | `{ "stream": true }`, `{ "headers": { "Authorization": "Bearer <token>" } }` |

    **Forcing Model Versions**

    You can force the LLM Prompt Node to use a specific model version by including it in the Custom Options.
    This means that the LLM Prompt Node will use the specified version of the language model instead of the default or any other available versions. This allows for more control over the behavior of the LLM Prompt Node, ensuring it uses the desired model version for generating prompts or responses.

    You can use models from any LLM provider supported by Cognigy, including those not yet directly integrated.
    However, you can only replace a model with another from the same provider.

    Let's consider an example with Anthropic:
    You can force the LLM Prompt Node to use the model version `claude-opus-4-1-20250805`, despite the LLM resource defaulting to the `claude-opus-4-5-20251101` model:

    1. [Create an Anthropic LLM resource](/ai/agents/develop/gen-ai-and-llms/providers/anthropic) for Claude, for example, `claude-opus-4-5-20251101`.
    2. Create a Flow and add an LLM Prompt Node to it.
    3. In the LLM Prompt Node, select the model `claude-opus-4-5-20251101` from the **Large Language Model** list.
    4. In the **Custom Model Options** field, add `{ "model": "claude-opus-4-1-20250805" }` to force the use of the `claude-opus-4-1-20250805` model.
    5. Click **Save Node**.

    The LLM Prompt Node now utilizes the `claude-opus-4-1-20250805` model.

    Below, you'll find documentation for supported models:

    * [Aleph Alpha Models](https://docs.aleph-alpha.com/)
    * [Anthropic Models](https://docs.anthropic.com/claude/docs/models-overview)
    * [Azure OpenAI](https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/models)
    * [OpenAI Models](https://platform.openai.com/docs/models)
  </Accordion>

  <Accordion title="Debugging Settings">
    When using the Interaction Panel, you can trigger two types of debug logs. These logs are only available when using the Interaction Panel and are not intended for production debugging. You can also combine both log types.

    | Parameter                      | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                            |
    | ------------------------------ | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Show Token Count               | Toggle        | Sends a debug message containing the input, output, and total token count. The message appears in the Interaction Panel when [debug mode](/ai/agents/test/interaction-panel/chat#debug-mode) is enabled. Cognigy uses the GPT-3 tokenizer algorithm, so actual token usage may vary depending on the model used. The parameter is inactive by default.                                                 |
    | Log System Prompt & Completion | Toggle        | Sends a debug message containing the system prompt sent to the LLM provider and the subsequent completion. The message appears in the Interaction Panel when [debug mode](/ai/agents/test/interaction-panel/chat#debug-mode) is enabled. The parameter is inactive by default.                                                                                                                         |
    | Log Tool Definitions           | Toggle        | Sends a debug message containing information about the configured AI Agent tools. The message appears in the Interaction Panel when [debug mode](/ai/agents/test/interaction-panel/chat#debug-mode) is enabled. The parameter is inactive by default.                                                                                                                                                  |
    | Log LLM Latency                | Toggle        | Sends a debug message containing key latency metrics for the request to the model, including the time taken for the first output and the total time to complete the request. The message appears in the Interaction Panel when [debug mode](/ai/agents/test/interaction-panel/chat#debug-mode) is enabled. The parameter is inactive by default.                                                       |
    | Send request logs to Webhook   | Toggle        | Sends the request sent to the LLM provider and the subsequent completion to a webhook service, including metadata, the request body, and custom logging data. With this parameter, you can use a webhook service to view detailed logs of the request to the LLM. The parameter is inactive by default.                                                                                                |
    | Webhook URL                    | CognigyScript | Sets the URL of the webhook service to send the request logs to.                                                                                                                                                                                                                                                                                                                                       |
    | Custom Logging Data            | CognigyScript | Sets custom data to send with the request to the webhook service.                                                                                                                                                                                                                                                                                                                                      |
    | Condition for Webhook Logging  | CognigyScript | Sets the condition for the webhook logging.                                                                                                                                                                                                                                                                                                                                                            |
    | Webhook Headers                | Input fields  | Sets the headers to send with the request to the webhook service. Use the **Key** and **Value** fields to enter a header. The **Value** field supports CognigyScript. After entering the header key, new empty **Key** and **Value** fields are automatically added, in case you need to add more headers. Alternatively, you can click **Show JSON Editor** and add input examples in the code field. |
  </Accordion>
</AccordionGroup>

### Child Nodes

#### Tool

Tools are child Nodes of AI Agent Nodes. They define the actions the AI Agent can take. If an AI Agent wants to execute the tool, the branch below the child Node is executed.

At the end of a tool branch, it is advisable to use a [Resolve Tool Action](/ai/agents/develop/node-reference/ai/resolve-tool-action) Node to return to the AI Agent.

Clicking the Tool Node lets you define a tool, set its parameters, and allows debugging by enabling detailed messages about the tool's execution.

<AccordionGroup>
  <Accordion title="Tool">
    | Parameter   | Type          | Description                                                                                                                                          |
    | ----------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Tool ID     | CognigyScript | Provide a meaningful name as a Tool ID. This ID can contain only letters, numbers, underscores (`_`), or dashes (`-`). For example, `update_user-1`. |
    | Description | CognigyScript | Provide a detailed description of what the tool does, when it should be used, and its parameters.                                                    |
  </Accordion>

  <Accordion title="Parameters">
    Configure the parameters that the AI Agent will collect before the tool is called. You can switch between the Graphical and JSON editors. When editing the JSON, follow the [JSON Schema specification](https://json-schema.org).

    | Parameter       | Type     | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         |
    | --------------- | -------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Use Parameters  | Toggle   | Activate this toggle to add parameters in addition to the tool name and description. The AI Agent will collect all data it needs and call a Tool with these parameters filled as arguments. These values can be accessed directly in the `input.aiAgent.toolArgs` object.                                                                                                                                                                                                                                                                                                                                                                           |
    | Name            | Text     | Specify the name of the parameter. The name should be clear and concise, and describe the purpose of the parameter.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
    | Type            | Selector | Select a type of the parameter:<ul><li>**String** — a sequence of characters. For example, `"hello"`, `"123"`.</li><li>**Number** — a numerical value, which can be either an integer (for example,`5`) or a floating point number (for example, `3.14`).</li><li>**Boolean** — a logical value representing `true` or `false`.</li><li>**Array** — a collection of elements, which can contain multiple values of any type. For example, `["apple", "banana", "cherry"]`.</li><li>**Object** — a collection of key-value pairs, where each key is a string and the value can be of any type. For example, `{"name": "John", "age": 30}`.</li></ul> |
    | Description     | Text     | Explain what the parameter means by providing a brief description of the parameter's usage.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         |
    | Enum (optional) | Enum     | Define a set of values that the parameter can accept. The enum restricts the input to one of the specified values, ensuring only valid options are chosen. The enum is only available for string-type parameters in the Graphical editor. For other types, use the JSON editor. May not be supported by all LLM providers.                                                                                                                                                                                                                                                                                                                          |
    | Add Parameter   | Button   | Add a new parameter.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
  </Accordion>

  <Accordion title="Debug Settings">
    | Parameter                 | Type   | Description                                                                                                       |
    | ------------------------- | ------ | ----------------------------------------------------------------------------------------------------------------- |
    | Debug Message when called | Toggle | Enable the output of a debug message when the tool is called to provide detailed information about the tool call. |
  </Accordion>

  <Accordion title="Advanced">
    | Parameter | Type          | Description                                                                                                                                                                                                                                                                                                                                                                        |
    | --------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Condition | CognigyScript | The tool will be enabled only if the condition is evaluated as true. If false, the tool isn't part of the AI Agent's Tools within this execution. For example, when using the `unlock_account` tool, you can specify a condition like `context.accountStatus === "locked"`. This checks the value in the context, and if it is missing or different, the tool will not be enabled. |
  </Accordion>
</AccordionGroup>

#### MCP Tool

MCP Tool Nodes are child Nodes of AI Agent Nodes. The MCP Tool Nodes connect to a remote [MCP server](https://modelcontextprotocol.io/introduction) to load tools that the AI Agent can execute. If an AI Agent wants to execute one of the loaded tools, the branch below the MCP Tool Node is triggered.

Clicking the MCP Tool Node lets you define the connection, filter loaded tools, and allows debugging by enabling detailed messages about the tool's execution.

<AccordionGroup>
  <Accordion title="MCP Tool">
    | Parameter          | Type          | Description                                                                                                                                                                            |
    | ------------------ | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Name               | CognigyScript | Provide a name for the MCP connection. This name helps you identify the source of the loaded tool.                                                                                     |
    | MCP Server SSE URL | CognigyScript | Provide the URL to an SSE (Server-Sent Events) endpoint from a remote [MCP server](https://modelcontextprotocol.io/introduction). Ensure that you connect only to trusted MCP servers. |
    | Timeout            | Slider        | Set the timeout time for the MCP connection in seconds.                                                                                                                                |
  </Accordion>

  <Accordion title="Authentication">
    | Parameter           | Type     | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
    | ------------------- | -------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Authentication Type | Selector | The authentication type to use for the MCP server connection. Select one of the following options: <ul><li>**None** — no authentication. This option is used by default.</li><li>**OAuth2** — authenticate using OAuth 2.0, which provides improved security and control compared to API keys. OAuth 2.0 supports scoped permissions, expiring tokens, and reduced exposure through short-lived tokens.</li></ul> If you want to use a different authentication method, such as an API key, you can add it as a custom header in **Custom Headers** of the **Advanced** section.                                                                                                                                                                                                                                                             |
    | OAuth2 Parameters   | Selector | Opens the fields required to configure OAuth 2.0 authentication. These fields allow the AI Agent to securely request and use access tokens when communicating with the MCP server. Configure the following fields: <ul><li>**Connection Name** — enter a descriptive name for the connection, for example, `MCP Tool OAuth`.</li><li>**OAuth URL** — enter the token endpoint URL exposed by the MCP server, for example: `https://mcp.example.com/oauth2/token`.</li><li>**Client ID** — enter the application (client) ID issued by the MCP server.</li><li>**Client Secret** — enter the application secret issued by the MCP server. Keep it secure.</li><li>**Scope** — add a list of scopes that define the permissions required by your application, for example: `read write admin`. Separate multiple scopes with spaces.</li></ul> |
  </Accordion>

  <Accordion title="Debug Settings">
    | Parameter             | Type   | Description                                                                                                                                                                                                                                                               |
    | --------------------- | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Debug Loaded Tools    | Toggle | Enable this parameter to display a debug message listing all tools loaded from the MCP server. The debug message also includes tools filtered out in the Advanced section. This parameter shows whether tools were loaded from the cache or directly from the MCP server. |
    | Debug with Parameters | Toggle | Enable this parameter to include the Tool Parameters in the debug message.                                                                                                                                                                                                |
    | Debug calling Tool    | Toggle | Enable the output of a debug message when the tool is called to provide detailed information about the tool call.                                                                                                                                                         |
  </Accordion>

  <Accordion title="Advanced">
    | Parameter      | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | -------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Cache Tools    | Toggle        | Disables caching of loaded tools while developing. Ensure that caching is enabled in production for performance reasons. The caching time is 10 minutes. In debug mode, you can verify whether the cache was used. To do this, make sure **Debug Loaded Tools** is enabled. For more information, see the [Debug Settings](#debug-settings) section.                                                                                                                    |
    | Condition      | CognigyScript | Sets the condition under which the tool will be activated. If the condition is evaluated as false, the tool isn't part of the AI Agent's Tools during execution. For example, when using the `unlock_account` tool, you can specify a condition like `context.accountStatus === "locked"`. This checks the value in the context, and if it is missing or different, the tool will not be enabled.                                                                       |
    | Tool Filter    | Select        | Controls if tools should be excluded from execution. You can select one of the following options:<ul><li>**None** — no tool filtering is applied, and all tools are available for execution. This option is selected by default.</li><li>**Whitelist** — only tools on the list are allowed for execution, while all other tools are excluded.</li><li>**Blacklist** — tools on the list are excluded from execution, while all other tools remain available.</li></ul> |
    | Blacklist      | CognigyScript | This parameter appears if you select **Blacklist** in **Tool Filter**. Specify the tools that should be blocked from execution. Specify only one tool per field.                                                                                                                                                                                                                                                                                                        |
    | Whitelist      | CognigyScript | This parameter appears if you select **Whitelist** in **Tool Filter**. Specify the tools you want to allow for execution. Specify only one tool per field.                                                                                                                                                                                                                                                                                                              |
    | Custom Headers | -             | Sets custom authentication headers to send with the request to the MCP server. Use the **Key** and **Value** fields to enter a header. The **Value** field supports CognigyScript. After entering the header key, new empty **Key** and **Value** fields are automatically added, in case you need to add more headers. Alternatively, you can click **Show JSON Editor** and enter the headers in the code field.                                                      |
  </Accordion>
</AccordionGroup>

#### Call MCP Tool

In the Flow editor, when you add an MCP Tool Node, a Call MCP Tool Node is automatically created below it. These two Nodes work together to define and execute the chosen tool. The Call MCP Tool Node sets the actual execution point of the chosen tool. This way, you can verify or modify the tool call arguments in the `input.aiAgent.toolArgs` object, or add a Say Node before the tool call. When the Call MCP Tool Node is executed, the tool call is sent to the remote MCP server, where the Tool is executed remotely with any arguments set by the AI Agent.

To return the tool result to the AI Agent, the **Resolve Immediately** setting can be enabled to send the full result returned from the remote MCP server to the AI Agent.

As an alternative, use a [Resolve Tool Action](/ai/agents/develop/node-reference/ai/resolve-tool-action) Node to return a specific result to the AI Agent.

<AccordionGroup>
  <Accordion title="Call MCP Tool">
    | Parameter           | Type   | Description                                                                                           |
    | ------------------- | ------ | ----------------------------------------------------------------------------------------------------- |
    | Resolve Immediately | Toggle | Enable this parameter to immediately resolve the tool action with the full result as the tool answer. |
  </Accordion>

  <Accordion title="Storage Options">
    | Parameter                   | Type          | Description                                                                                                                                                                                                                               |
    | --------------------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | How to handle the result    | Select        | Determine how to handle the MCP tool call result:<ul><li>**Store in Input** — stores the result in the Input object.</li><li>**Store in Context** — stores the result in the Context object.</li></ul>                                    |
    | Input Key to store Result   | CognigyScript | The parameter appears when **Store in Input** is selected. The result is stored in the `input.aiAgent.toolResult` object by default. You can specify another value, but the **MCP Tool Result** Token won't work if the value is changed. |
    | Context Key to store Result | CognigyScript | The parameter appears when **Store in Context** is selected. The result is stored in the `context.aiAgent.toolResult` object by default.                                                                                                  |
  </Accordion>

  <Accordion title="Debug Settings">
    | Parameter         | Type   | Description                                                                             |
    | ----------------- | ------ | --------------------------------------------------------------------------------------- |
    | Debug Tool Result | Toggle | Enable the output of a debug message with the tool call result after a successful call. |
  </Accordion>
</AccordionGroup>

#### Knowledge Tool

The Knowledge Tool Node is a child Node of the AI Agent Node. It allows the AI Agent to directly access [Knowledge Stores](/ai/agents/develop/knowledge-ai/knowledge-store) to provide context-aware responses.

In the Knowledge Tool Node, you can select the [Knowledge Store](/ai/agents/develop/knowledge-ai/knowledge-store) to search, [Source Tags](/ai/agents/develop/knowledge-ai/knowledge-source/knowledge-source-tags) to refine the search, and allows for detailed debug messages about the tool's execution.

<AccordionGroup>
  <Accordion title="Knowledge Tool">
    | Parameter       | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
    | --------------- | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Knowledge Store | Selector      | Select a Knowledge Store.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
    | Tool ID         | CognigyScript | Enter a name as a tool ID. This ID can contain only letters, numbers, underscores (`_`), or dashes (`-`). If you use more than one Knowledge tool, the tool ID provides context to trigger the correct Knowledge tool. In this case, enter a clear tool ID for each Knowledge tool, for example, `search_appliances`                                                                                                                                                                                                                                                                                                                        |
    | Description     | CognigyScript | Enter instructions to guide the AI Agent to call the Knowledge tool. The description field provides context to trigger the Knowledge tool. If you want to use more than one Knowledge tool, enter clear instructions for the cases when each Knowledge tool should be used. For example, `Find the answer to prompts or questions about appliances by searching the attached data sources. Use this tool when a customer asks about appliance items such as washing machines, dryers, and other household appliances. Focus exclusively on a knowledge search and does not execute tasks like small talk, calculations, or script running.` |
  </Accordion>

  <Accordion title="Debug Settings">
    | Parameter                 | Type   | Description                                                                                                       |
    | ------------------------- | ------ | ----------------------------------------------------------------------------------------------------------------- |
    | Debug Message when called | Toggle | Enable the output of a debug message when the tool is called to provide detailed information about the tool call. |
  </Accordion>

  <Accordion title="Advanced">
    | Parameter                   | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
    | --------------------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Top K                       | Slider        | Set how many Knowledge Chunks to return. Providing more results gives the AI Agent additional context, but it also increases noise and token usage.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
    | Store Location              | Selector      | Select whether and where to store the knowledge search results. Select one of the following options:<ul><li>**Don't store** — the content isn't stored. This option is set by default.</li><li>**Input** — the content is stored in the Input object.</li><li>**Context** — the content is stored in the Context object.</li></ul>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
    | Input Key to store result   | CognigyScript | Appears when **Store Location** is set to **Input**. The property in the Input object where the result is stored. For example, `input.knowledgeSearch`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | Context Key to store result | CognigyScript | Appears when **Store Location** is set to **Context**. The property in the Context object where the result is stored. For example, `context.knowledgeSearch`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
    | Source Tags                 | CognigyScript | Enter [Knowledge Source Tags](/ai/agents/develop/knowledge-ai/knowledge-source/knowledge-source-tags) to refine the scope of your knowledge search, including only the most relevant Knowledge Chunks in the Knowledge Store. Before entering tags, ensure they are included in the Knowledge Sources. Add tags by entering each separately and pressing <kbd>Enter</kbd>. Max 5 tags. When multiple Source Tags are specified, the Search Extract Output Node defaults to the `AND` operator, meaning it only considers Sources that have all specified tags. To change this behavior, adjust the **Match Type for Source Tags** parameter.                                                                                                                                                                      |
    | Match type for Source Tags  | Selector      | The operator to filter Knowledge Sources by Source Tags. Select one of the following options:<ul><li>**AND** — the default value, requires all Source Tags to match across multiple Knowledge Sources. Consider the following example: there are Knowledge Sources with Source Tags `S-a`, `S-b`, and `S-c`. When you use the `AND` operator to filter by `S-a` and `S-b`, only Sources with both Tags `S-a` and `S-b` are included in the search results.</li><li>**OR** — requires at least one Source Tag to match across multiple Knowledge Sources. Consider the following example: there are Knowledge Sources with Source Tags `S-a`, `S-b`, and `S-c`. When you use the `OR` operator to filter by `S-a` or `S-b`, any Source with either Tag `S-a` or `S-b` is included in the search results.</li></ul> |
    | Condition                   | CognigyScript | The Knowledge tool is activated only if the condition is evaluated as true. If false, the tool isn't included in the current execution. For example, when using the Knowledge tool, you can enter a condition such as `context.productCategory === "appliances"`. This checks the value in the context, and if it is missing or different, the tool isn't activated.                                                                                                                                                                                                                                                                                                                                                                                                                                              |
  </Accordion>
</AccordionGroup>

#### Send Email Tool

The Send Email tool lets your AI Agent send emails directly to users. The Send Email tool uses the same configuration and restrictions as the [Email Notification Node](/ai/agents/develop/node-reference/data/email-notification) but adds more flexibility. While the Email Notification Node sends emails at a fixed step in a Flow, the Send Email tool allows the AI Agent to send emails dynamically, based on user input, conversation context, or instructions. This tool makes automation more flexible and minimizes extra Flow steps.

<AccordionGroup>
  <Accordion title="Tool">
    | Parameter                    | Type          | Description                                                                                                                                              |
    | ---------------------------- | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Tool ID                      | CognigyScript | Provide a meaningful name as a tool ID. This ID can contain only letters, numbers, underscores (`_`), or dashes (`-`). The default name is `send_email`. |
    | Description                  | CognigyScript | Provide a detailed description of what the tool should do. The default description is `Create and send a new email message.`                             |
    | Recipient TO Email Addresses | CognigyScript | A comma-separated list of email addresses to which the email will be sent.                                                                               |
  </Accordion>

  <Accordion title="Debug Settings">
    | Parameter                 | Type   | Description                                                                                                       |
    | ------------------------- | ------ | ----------------------------------------------------------------------------------------------------------------- |
    | Debug Message when called | Toggle | Enable the output of a debug message when the tool is called to provide detailed information about the tool call. |
  </Accordion>

  <Accordion title="Advanced">
    | Parameter | Type          | Description                                                                                                                                                                                                                                                                                                                                                                        |
    | --------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Condition | CognigyScript | The tool will be enabled only if the condition is evaluated as true. If false, the tool isn't part of the AI Agent's Tools within this execution. For example, when using the `unlock_account` tool, you can specify a condition like `context.accountStatus === "locked"`. This checks the value in the context, and if it is missing or different, the tool will not be enabled. |
  </Accordion>
</AccordionGroup>

#### Handover to AI Agent Tool

The Handover to AI Agent tool lets you transfer a conversation to another AI Agent in the same or a different Flow. This approach ensures the conversation keeps its context, different AI Agents can handle specific tasks, and multi-step conversations run smoothly across Flows.

<AccordionGroup>
  <Accordion title="Tool">
    | Parameter   | Type          | Description                                                                                                                                                                                                                                                     |
    | ----------- | ------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Tool ID     | CognigyScript | Provide a meaningful name as a tool ID. This ID can contain only letters, numbers, underscores (`_`), or dashes (`-`). The default name is `handover_to_ai_agent`.                                                                                              |
    | Description | CognigyScript | Describe what the receiving AI Agent should do. For example: `Handle product recommendations`, `Perform technical troubleshooting`, or `Assist with billing questions`. Be specific so the handover rules clearly indicate when this AI Agent should take over. |
    | Select Flow | Selector      | Select the target Flow to hand over to. By default, all Flows in the Project are displayed in the list. The selected Flow must contain an AI Agent Node.                                                                                                        |
    | Select Node | Selector      | Select the AI Agent Node to hand over to.                                                                                                                                                                                                                       |
  </Accordion>

  <Accordion title="Debug Settings">
    | Parameter                 | Type   | Description                                                                                                       |
    | ------------------------- | ------ | ----------------------------------------------------------------------------------------------------------------- |
    | Debug Message when called | Toggle | Enable the output of a debug message when the tool is called to provide detailed information about the tool call. |
  </Accordion>

  <Accordion title="Advanced">
    | Parameter | Type          | Description                                                                                                                                                                                                                                                                                                                                                                        |
    | --------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Condition | CognigyScript | The tool will be enabled only if the condition is evaluated as true. If false, the tool isn't part of the AI Agent's Tools within this execution. For example, when using the `unlock_account` tool, you can specify a condition like `context.accountStatus === "locked"`. This checks the value in the context, and if it is missing or different, the tool will not be enabled. |
  </Accordion>
</AccordionGroup>

#### Handover to Human Agent Tool

The Handover to Human Agent tool lets you transfer a conversation to a human agent through handover providers. This tool allows your AI Agents to refer complex tasks to a human agent if they can't address the user's request.

<AccordionGroup>
  <Accordion title="Tool">
    | Parameter                        | Type          | Description                                                                                                                                                                                                                                                                                                                                            |
    | -------------------------------- | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Tool ID                          | CognigyScript | Provide a meaningful name as a tool ID. This ID can contain only letters, numbers, underscores (`_`), or dashes (`-`). The default name is `handover_to_human_agent`.                                                                                                                                                                                  |
    | Description                      | CognigyScript | Describe the situation when the AI Agent should refer to a human agent. For example: `Use this tool if the user explicitly asks for a human.`, or `Use this tool if you fail to answer the user's question three times.`. Be specific so the handover rules clearly indicate when the conversation should be handed over to the receiving human agent. |
    | Handover Provider                | Selector      | Select the handover provider from the list. The list contains the handover providers created in **Deploy > Handover Providers**.                                                                                                                                                                                                                       |
    | Handover Accepted Message        | CognigyScript | The message to be outputted when the handover is requested.                                                                                                                                                                                                                                                                                            |
    | Repeat Handover Accepted Message | Toggle        | Enable to repeatedly output the Handover Accepted Message if the user sends a message while waiting in the queue for an agent.                                                                                                                                                                                                                         |
  </Accordion>

  <Accordion title="Live Agent Settings">
    | Parameter                       | Description                                                                                                                                                                                                                                                                                                                                                                  |
    | ------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Live Agent Inbox Id             | Enter the Inbox ID that you obtained in the [previous step](/ai/escalate/handover-reference/live-agent).                                                                                                                                                                                                                                                                     |
    | Skills                          | Enter a list of skills that should be used to filter the available agents. For example, `technical support, billing`.                                                                                                                                                                                                                                                        |
    | Languages                       | Enter a list of languages that should be used to filter the available agents. For example, `english, spanish`.                                                                                                                                                                                                                                                               |
    | Priority                        | Select a priority for the conversation from the list. For example, `Urgent`.                                                                                                                                                                                                                                                                                                 |
    | Get Queue Position              | Enable the system to retrieve the end user's current queue position. This parameter is inactive by default.                                                                                                                                                                                                                                                                  |
    | Update Interval                 | Specify the interval, in seconds, for how often to receive updates on the queue status. This parameter appears if **Get Queue Position** is selected.                                                                                                                                                                                                                        |
    | Get Estimated Wait Time         | Enable the retrieval of the estimated wait time for the end user. This parameter is inactive by default.                                                                                                                                                                                                                                                                     |
    | Update Interval                 | Set the interval, in seconds, for how often to receive updates on estimated wait time. This parameter appears if **Get Estimated Wait Time** is selected.                                                                                                                                                                                                                    |
    | Alternative Update              | The alternative update options when certain conditions are met. This parameter is inactive by default.                                                                                                                                                                                                                                                                       |
    | Maximum Queue Position          | Set the threshold for the maximum queue position. If this threshold is reached, the alternative text is sent. This parameter appears if **Alternative Update** is selected.                                                                                                                                                                                                  |
    | Maximum Estimated Wait Time     | Define the maximum estimated wait time in milliseconds. If this duration is exceeded, the alternative text is sent. This parameter appears if **Alternative Update** is selected.                                                                                                                                                                                            |
    | Alternative Text                | The message to be sent when either the maximum queue position or maximum wait time is reached. This parameter appears if **Alternative Update** is selected.                                                                                                                                                                                                                 |
    | Enable User Connects Message    | Notify human agents when an end user reconnects to the chat. The parameter is enabled by default. When the parameter is enabled, the message `User joined the conversation` appears in the chat as soon as the end user returns to the chat tab by clicking the **←** (back arrow) at the top bar in the browser, after having opened a new URL on the same tab as the chat. |
    | Enable User Disconnects Message | Notify human agents when an end user disconnects from the chat. The parameter is enabled by default. When the parameter is enabled, the message `User left the conversation` is sent as soon as the end user closes the tab with the chat or switches to a new URL address within the current tab.                                                                           |

    For more information, read the [Live Agent Handover Setup](/live-agent/getting-started/live-agent-setup/live-agent-setup-handover-flow) and [Real-Time Queue Notifications](/ai/escalate/handover-features/real-time-queue-notifications) documentation.
  </Accordion>

  <Accordion title="Chatwoot Settings">
    | Parameter         | Description                                                                                                                                                                                                                                                                                          |
    | ----------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Chatwoot Inbox Id | Enter the Chatwoot Inbox ID. You can find the Inbox ID in your Chatwoot URL. For example, `https://app.chatwoot.com/app/account/6607/inbox/5446`, where the Inbox ID is `5446`. For more information, read the [Chatwoot handover provider](/ai/escalate/handover-reference/chatwoot) documentation. |
  </Accordion>

  <Accordion title="Salesforce Chat Settings">
    | Parameter                        | Description                                                                                                                                                                                                                                        |
    | -------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Send Transcript as first message | Activate this parameter to allow Salesforce to receive the conversation transcript as the first message. This setting is turned off by default.                                                                                                    |
    | Button ID                        | Enter the ID of the Button you created in the Salesforce installation. This setting will override the default Button ID that you specified in the Endpoint configuration. You can use different Button IDs for individual Handover to Agent Nodes. |
    | Pre-chat Entities                | Enter the pre-chat entities. For more information, read the [Salesforce Handover](/ai/escalate/handover-reference/salesforce) documentation.                                                                                                       |
    | Pre-chat Details                 | Enter the pre-chat details. For more information, read the [Salesforce Handover](/ai/escalate/handover-reference/salesforce) documentation.                                                                                                        |
  </Accordion>

  <Accordion title="NiCE CXone Settings">
    | Parameter                        | Description                                                                                                                                                                            |
    | -------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Send Transcript as first message | Activate this parameter to allow NiCE CXone to receive the conversation transcript as the first message when the handover begins. This setting is turned on by default.                |
    | Display Agent Details            | Display the human agent's name and avatar from NiCE CXone in the chat for the end user. This parameter is disabled by default.                                                         |
    | Custom Attributes                | Add custom attributes in JSON format to include additional information the handover. For more information, refer to [NiCE CXone](/ai/escalate/handover-reference/nice-cxone/overview). |

    Example for the **Custom Attributes** parameter:

    ```json theme={null}
    {
      "section": "section A",
      "region": "Europe"
    }
    ```
  </Accordion>

  <Accordion title="Genesys Settings">
    | Parameter                       | Description                                                                                                                                                                                                                                                                                                                                                                  |
    | ------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Language                        | Specify a language for the conversation. For example, `english`, `spanish`, `german`.                                                                                                                                                                                                                                                                                        |
    | Skills                          | Define skills for the conversation. For example, `escalation`.                                                                                                                                                                                                                                                                                                               |
    | Priority                        | Set the priority for the conversation. For example, `1`. <br /> If a priority is set, it triggers a flow in Genesys to prioritize or de-prioritize the conversation within the queue. <br /> Note that this functionality requires the appropriate flow to be set up in Genesys.                                                                                             |
    | Enable User Connects Message    | Notify human agents when an end user reconnects to the chat. The parameter is enabled by default. When the parameter is enabled, the message `User joined the conversation` appears in the chat as soon as the end user returns to the chat tab by clicking the **←** (back arrow) at the top bar in the browser, after having opened a new URL on the same tab as the chat. |
    | Enable User Disconnects Message | Notify human agents when an end user disconnects from the chat. The parameter is enabled by default. When the parameter is enabled, the message `User left the conversation` is sent as soon as the end user closes the tab with the chat or switches to a new URL address within the current tab.                                                                           |
    | Custom Attributes               | Add custom attributes in the `Key: Value` format, which allows you to include additional information.                                                                                                                                                                                                                                                                        |
    | Get Estimated Wait Time         | Retrieve the estimated wait time for the end user. End users then receive wait time updates while they wait to be connected to a human agent. This parameter is inactive by default. It requires an active [Amazon EventBridge Source integration](/ai/escalate/handover-reference/genesys-cloud-open-messaging-with-amazon-eventbridge).                                    |
    | Update Interval                 | Specify the interval, in seconds, for how often to receive updates on the estimated wait time. This parameter appears if **Get Estimated Wait Time** is selected.                                                                                                                                                                                                            |
    | Alternative Update              | Send an alternative message to the end user when the maximum estimated wait time is exceeded. This parameter is inactive by default. This parameter appears if **Get Estimated Wait Time** is selected.                                                                                                                                                                      |
    | Maximum Estimated Wait Time     | Define the maximum estimated wait time in milliseconds. If this duration is exceeded, the alternative text is sent instead of the actual wait time. This parameter appears if **Alternative Update** is selected.                                                                                                                                                            |
    | Alternative Text                | Specify the message to be sent when the maximum estimated wait time is exceeded. This parameter appears if **Alternative Update** is selected.                                                                                                                                                                                                                               |

    Genesys Cloud Open Messaging doesn't provide a queue position. Estimated wait time is the only queue parameter available.

    Example for the **Custom Attributes** parameter:

    ```json theme={null}
    {
      "section": "section A",
      "region": "Europe"
    }
    ```

    For more information,
    read
    the [Genesys Cloud Guest Chat](/ai/escalate/handover-reference/genesys-cloud-guest-chat#prerequisites), [Genesys Cloud Open Messaging](/ai/escalate/handover-reference/genesys-cloud-open-messaging), and [Real-Time Queue Notifications](/ai/escalate/handover-features/real-time-queue-notifications#configure-settings-in-genesys-cloud-open-messaging) documentation.
  </Accordion>

  <Accordion title="8x8 Settings">
    | Parameter           | Description                                                                                                                                                                                                                                                                      |
    | ------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | 8x8 Channel Id      | Specify the channel ID that you created during the 8x8 configuration. <br /> You can find this ID in the 8x8 admin console. Navigate to **Channels > Chat API**. In the channel list, find the channel that you created. Copy the **ID** from the channel row.                   |
    | 8x8 Queue Id        | Specify the header value required for authenticating API calls to the 8x8 system. <br /> You can find this ID in the 8x8 admin console. Navigate to **Channels > Chat API**. In the channel list, find the channel that you created. Copy the **Queue ID** from the channel row. |
    | 8x8 JSON properties | Extend 8x8 properties with custom JSON.                                                                                                                                                                                                                                          |

    For more information, read the [8x8 Handover Setup](/ai/escalate/handover-reference/8x8) documentation.
  </Accordion>

  <Accordion title="RingCentral Engage Settings">
    | Parameter                       | Type          | Description                                                                                                                                                                                                                                                                                                                                                                  |
    | ------------------------------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Additional Category Ids         | CognigyScript | IDs to be passed to the RingCentral Engage handover provider.                                                                                                                                                                                                                                                                                                                |
    | Enable User Connects Message    | Toggle        | Notify human agents when an end user reconnects to the chat. The parameter is enabled by default. When the parameter is enabled, the message `User joined the conversation` appears in the chat as soon as the end user returns to the chat tab by clicking the **←** (back arrow) at the top bar in the browser, after having opened a new URL on the same tab as the chat. |
    | Enable User Disconnects Message | Toggle        | Notify human agents when an end user disconnects from the chat. The parameter is enabled by default. When the parameter is enabled, the message `User left the conversation` is sent as soon as the end user closes the tab with the chat or switches to a new URL address within the current tab.                                                                           |

    For more information, read the [RingCentral Engage Integration](/ai/escalate/handover-reference/ring-central-engage) documentation.
  </Accordion>

  <Accordion title="Cancel Handover Options">
    | Parameter          | Type          | Description                                                                                          |
    | ------------------ | ------------- | ---------------------------------------------------------------------------------------------------- |
    | Cancel Intent      | CognigyScript | The Intent the user has to trigger to cancel the Handover request.                                   |
    | Cancel Button Text | CognigyScript | The text for a quick reply button that cancels the handover request when the user clicks the button. |
  </Accordion>

  <Accordion title="On Resolve Options">
    | Parameter                           | Type     | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | ----------------------------------- | -------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Flow Continuation                   | Selector | Controls where the Flow continues after the human agent leaves the conversation.<br />When the user needs assistance completing the Flow with the AI Agent, the conversation is handed over to a human agent. The human agent can trigger the Flow by sending a `/bot` command as a reply in the conversation interface.<br />For example, the user enters their date of birth in an incorrect format, and the AI Agent can't process the response. The conversation is handed over to a human agent, who identifies the incorrect format and sends the corrected date using `/bot MM/DD/YYYY`. The AI Agent takes over and continues the Flow.<br /><ul><li>**Below this Node** — the Flow continues from the first Node below the Human Agent Node, regardless of where the human agent stopped assisting the user. This option is selected by default.</li> <li>**At current AgentInject Entrypoint** — the Flow continues from the Node where the human agent stopped assisting the user.</li></ul> |
    | Send resolve event to Virtual Agent | Toggle   | Controls whether the Flow receives the `conversation_status_changed` event with the status set to `resolved`. This parameter is activated by default.<ul><li>When the parameter is activated, the event is sent when the human agent leaves the conversation. As a result, the Flow continues immediately, without waiting for the user to send another message. If the [Lookup Case](/ai/agents/develop/node-reference/logic/lookup) Node is configured with **Handover Status** set to **Completed**, the Node can detect the completed handover and continue the Flow.</li><li>When the parameter is deactivated, the event isn't sent when the human agent leaves the conversation. The Flow continues after the user sends their next message.</li></ul>                                                                                                                                                                                                                                           |
  </Accordion>

  <Accordion title="Event Settings">
    | Parameter         | Type   | Description                                             |
    | ----------------- | ------ | ------------------------------------------------------- |
    | Send Queue Event  | Toggle | Sends an event to the Flow once the handover is queued. |
    | Send Active Event | Toggle | Sends an event to the Flow once the handover is active. |
  </Accordion>

  <Accordion title="Debug Settings">
    | Parameter                 | Type   | Description                                                                                                       |
    | ------------------------- | ------ | ----------------------------------------------------------------------------------------------------------------- |
    | Debug Message when called | Toggle | Enable the output of a debug message when the tool is called to provide detailed information about the tool call. |
  </Accordion>

  <Accordion title="Advanced">
    | Parameter | Type          | Description                                                                                                                                                                                                                                                                                                                                                                        |
    | --------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Condition | CognigyScript | The tool will be enabled only if the condition is evaluated as true. If false, the tool isn't part of the AI Agent's Tools within this execution. For example, when using the `unlock_account` tool, you can specify a condition like `context.accountStatus === "locked"`. This checks the value in the context, and if it is missing or different, the tool will not be enabled. |
  </Accordion>
</AccordionGroup>

#### Execute Workflow Tool

The Execute Workflow tool lets the AI Agent execute another Flow. After the target Flow executes, the conversation returns to the AI Agent.

The target Flow inherits the [Context object](/ai/agents/develop/ai-agent-memory/context) from the AI Agent. If the Context object is changed during the target Flow execution, these changes are also available to the AI Agent.

| Parameter   | Description                                                                                                                                                                                                                                                                                     |
| ----------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Tool ID     | Enter a meaningful name as a tool ID. This ID can contain only letters, numbers, underscores (`_`), or dashes (`-`). The default name is `execute_workflow`.                                                                                                                                    |
| Description | Describe when to trigger the AI Agent to execute the target Flow. For example, if you have a specific Flow for dealing with package returns, enter: `Use this tool if the user explicitly asks to return a package.`. Be specific so the AI Agent only executes the target Flow when necessary. |
| Target Flow | Select the target Flow that the AI Agent executes when this tool is called. Click <img alt="open target flow" /> to open the target Flow.                                                                                                                                                       |
| Select Node | Select the target Node that the AI Agent executes when the target Flow is called. You must set the Node you want to select as an Entrypoint with the [Set Entrypoint](/ai/agents/develop/nodes/overview#set-nodes-as-entrypoints) setting before you can select it.                             |

<AccordionGroup>
  <Accordion title="Debug Settings">
    | Parameter                 | Type   | Description                                                                                                         |
    | ------------------------- | ------ | ------------------------------------------------------------------------------------------------------------------- |
    | Debug Message when called | Toggle | Activate the output of a debug message when the tool is called to provide detailed information about the tool call. |
  </Accordion>

  <Accordion title="Advanced">
    | Parameter      | Type   | Description                                                                                                                                                                                                                                                                                                                                                                                                                              |
    | -------------- | ------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Parse Intents  | Toggle | Sets whether Intents should be parsed in the target Flow. This parameter is activated by default.<br /><br />If this parameter is enabled, Cognigy NLU reparses the Intents in the target Flow based on `text` in the [Input object](/ai/agents/develop/ai-agent-memory/input#accessing-input-properties) and generates Intent scores using the Intents in the target Flow instead of the ones in the Flow the AI Agent is currently in. |
    | Parse Slots    | Toggle | Sets whether Slots should be parsed in the target Flow. This parameter is activated by default.<br /><br />If this parameter is enabled, Cognigy NLU reparses the Slots in the target Flow based on `text` in the [Input object](/ai/agents/develop/ai-agent-memory/input#accessing-input-properties) and generates Slot scores using the Slots in the target Flow instead of the ones in the Flow the AI Agent is currently in.         |
    | Absorb Context | Toggle | Sets whether the [default Context object](/ai/agents/develop/ai-agent-memory/context#default-context) of the target Flow is used in the context of the Flow the AI Agent is currently in.<br /><br />If the default Context of the target Flow has a value, it is copied into the Context object of the AI Agent as soon as the target Flow is executed.                                                                                 |
  </Accordion>
</AccordionGroup>

## Examples

### Tool

In this example, the `unlock_account` tool unlocks a user account by providing the email and specifying the reason for the unlocking.

Parameter configuration in JSON:

```json theme={null}
{
  "type": "object",
  "properties": {
    "email": {
      "type": "string",
      "description": "User's login email for their account."
    }
  },
  "required": ["email"],
  "additionalProperties": false
}
```

where:

* `type` — the type for a tool parameter schema, which must always be `object`.
* `properties` — defines the parameters for the tool configuration:
  * `email` — a required tool parameter for unlocking the account.
    * `type` — defines the data type for the tool parameter.
    * `description` — a brief explanation of what the property represents.
* `required` — lists `email` as a required parameter, ensuring that this value is always provided when the tool is called.
* `additionalProperties` — ensures that the input contains only the `email` tool parameter, and no others are allowed.

### MCP Tool and Call MCP Tool

#### Use Zapier's Remote MCP server

You can create a custom MCP server with personalized tools by using one of the [provided SDKs](https://github.com/modelcontextprotocol). For a quicker setup, you can use a third-party provider. For example, [Zapier](https://zapier.com/) allows you to [configure your MCP server](https://actions.zapier.com/settings/mcp/), which can be connected to multiple application APIs.

To use Zapier as a remote MCP server, follow these steps:

1. Log in to your Zapier account, go to the [MCP settings](https://actions.zapier.com/settings/mcp/) page, and configure your MCP server.
2. Copy the SSE URL and paste it into the **MCP Server SSE URL** field of your MCP Tool Node.
3. In the Zapier MCP settings, create an action to connect to various APIs. For example, you can create a Zapier action to automatically generate a Google Doc.

Once the setup is complete, the configured MCP Actions will be loaded when the AI Agent is executed.

You will see the following debug message in the Interaction Panel, indicating the result of the tool call after a successful execution:

```txt theme={null}
AI Agent: MCP Tool
Fetched tools from MCP Tool "zapier"

- google_docs_create_document_from_: Create a new document from text. Also supports limited HTML.
  Parameters:
  - title (string): Document Name
  - file (string): Document Content
```

### Knowledge Tool

In this example, you have two Knowledge tools to search two different Knowledge Stores. Each Knowledge Store refers to a different product category, in this example, appliances and furniture. To guide the AI Agent to use the correct Knowledge tool, enter a clear tool ID and description:

* Knowledge Tool 1:
  * Tool ID: `search_appliances`
  * Description: `Find the answer to prompts or questions about appliances by searching the attached data sources. Use this tool when a customer asks about appliance items such as washing machines, dryers, and other household appliances. Focus exclusively on a knowledge search and does not execute tasks like small talk, calculations, or script running.`
* Knowledge Tool 2:
  * Tool ID: `search_furniture`
  * Description: `Find the answer to prompts or questions about furniture by searching the attached data sources. Use this tool when a customer asks about furniture items such as sofas, tables, and other items. Focus exclusively on a knowledge search and does not execute tasks like small talk, calculations, or script running.`

## More Information

* [LLM](/ai/agents/develop/gen-ai-and-llms/llms)
* [Generative AI](/ai/agents/develop/gen-ai-and-llms/generative-ai)

***

<sup>1</sup>: Note that not all LLM models support streaming.


# MongoDB
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/service/mongo-db



<a href="/release-notes/4.88"><Badge>Updated in 4.88</Badge></a>

<Frame>
  <img alt="MongoDB Node configuration panel with find operation" />
</Frame>

## Description

The MongoDB Nodes let you execute a range of operations on your [MongoDB](https://www.mongodb.com/docs/).

## Restrictions

* MongoDB Nodes are compatible with MongoDB 3.6 or later versions.

## Nodes

### Find

| Parameter          | Type          | Description                                                                                                            |
| ------------------ | ------------- | ---------------------------------------------------------------------------------------------------------------------- |
| MongoDB Connection | List          | Select MongoDB database connection. To create a new connection, click <img alt="add MongoDB connection" />.            |
| Collection         | CognigyScript | The name of the MongoDB collection.                                                                                    |
| Query              | JSON          | The MongoDB query. [MongoDB Documentation: Find](https://docs.mongodb.com/manual/reference/method/db.collection.find/) |

<AccordionGroup>
  <Accordion title="Result Storage">
    | Parameter                 | Type          | Description                                                                                                                         |
    | ------------------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------- |
    | Where to store the result | List          | <ul><li>**Input** — stores the result in the Input object.</li><li>**Context** — stores the result in the Context object.</li></ul> |
    | Input Key to store Result | CognigyScript | Enter the key where the result should be stored.                                                                                    |
  </Accordion>

  <Accordion title="Advanced">
    | Parameter     | Type          | Description                                                                                                                                                            |
    | ------------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Projection    | JSON          | Read the MongoDB documentation for more information on Projection. [MongoDB Documentation: Projection](https://docs.mongodb.com/manual/reference/operator/projection/) |
    | Options       | JSON          | Read the MongoDB documentation for configurable options. [MongoDB Documentation](https://docs.mongodb.com/manual)                                                      |
    | Index         | CognigyScript | Refer to the MongoDB documentation on Indexes. [MongoDB Documentation: Indexes](https://docs.mongodb.com/manual/indexes/)                                              |
    | Stop on Error | Toggle        | Lets you configure whether to stop flow execution when an error occurs or not.                                                                                         |
  </Accordion>

  <Accordion title="Caching">
    | Parameter                               | Type   | Description                                                     |
    | --------------------------------------- | ------ | --------------------------------------------------------------- |
    | Cache Response from MongoDB Operation   | Toggle | Whether to cache the MongoDB responses to optimize performance. |
    | The expiration time of the cache (in s) | Number | The expiration time in seconds of the cache.                    |
  </Accordion>
</AccordionGroup>

### Find One

| Parameter          | Type          | Description                                                                                                                  |
| ------------------ | ------------- | ---------------------------------------------------------------------------------------------------------------------------- |
| MongoDB Connection | List          | Select MongoDB database connection. To create a new connection, click <img alt="add MongoDB connection" />.                  |
| Collection         | CognigyScript | The name of the MongoDB collection.                                                                                          |
| Query              | JSON          | The MongoDB query. [MongoDB Documentation: FindOne](https://docs.mongodb.com/manual/reference/method/db.collection.findOne/) |

<AccordionGroup>
  <Accordion title="Result Storage">
    | Parameter                 | Type          | Description                                                                                                                         |
    | ------------------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------- |
    | Where to store the result | List          | <ul><li>**Input** — stores the result in the Input object.</li><li>**Context** — stores the result in the Context object.</li></ul> |
    | Input Key to store Result | CognigyScript | Enter the key where the result should be stored.                                                                                    |
  </Accordion>

  <Accordion title="Advanced">
    | Parameter     | Type   | Description                                                                                                                                                            |
    | ------------- | ------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Projection    | JSON   | Read the MongoDB documentation for more information on Projection. [MongoDB Documentation: Projection](https://docs.mongodb.com/manual/reference/operator/projection/) |
    | Options       | JSON   | Read the MongoDB documentation for configurable options. [MongoDB Documentation](https://docs.mongodb.com/manual)                                                      |
    | Stop on Error | Toggle | Lets you configure whether to stop flow execution when an error occurs or not.                                                                                         |
  </Accordion>

  <Accordion title="Caching">
    | Parameter                               | Type   | Description                                                     |
    | --------------------------------------- | ------ | --------------------------------------------------------------- |
    | Cache Response from MongoDB Operation   | Toggle | Whether to cache the MongoDB responses to optimize performance. |
    | The expiration time of the cache (in s) | Number | The expiration time in seconds of the cache.                    |
  </Accordion>
</AccordionGroup>

### Update One

| Parameter          | Type          | Description                                                                                                                                             |
| ------------------ | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------- |
| MongoDB Connection | List          | Select MongoDB database connection. To create a new connection, click <img alt="add MongoDB connection" />.                                             |
| Collection         | CognigyScript | The name of the MongoDB collection.                                                                                                                     |
| Update             | JSON          | The MongoDB update payload in JSON format.[MongoDB Documentation: UpdateOne](https://docs.mongodb.com/manual/reference/method/db.collection.updateOne/) |

<AccordionGroup>
  <Accordion title="Result Storage">
    | Parameter                 | Type          | Description                                                                                                                         |
    | ------------------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------- |
    | Where to store the result | List          | <ul><li>**Input** — stores the result in the Input object.</li><li>**Context** — stores the result in the Context object.</li></ul> |
    | Input Key to store Result | CognigyScript | Enter the key where the result should be stored.                                                                                    |
  </Accordion>

  <Accordion title="Advanced">
    | Parameter     | Type   | Description                                                                                                                                                                        |
    | ------------- | ------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Use Operators | Toggle | If enabled, then you can use operators like `$push` in the update query. If not, then we will automatically \$set everything that you write in the update query into the database. |
    | Options       | JSON   | Read the MongoDB documentation for configurable options. [MongoDB Documentation](https://docs.mongodb.com/manual)                                                                  |
    | Stop on Error | Toggle | Lets you configure whether to stop flow execution when an error occurs or not.                                                                                                     |
  </Accordion>
</AccordionGroup>

### Update Many

[MongoDB Documentation: UpdateMany](https://docs.mongodb.com/manual/reference/method/db.collection.updateMany/)

### Insert

[MongoDB Documentation: Insert](https://docs.mongodb.com/manual/reference/method/db.collection.insert/)

### Remove

[MongoDB Documentation: Remove](https://docs.mongodb.com/manual/reference/method/db.collection.remove/)

### MongoDB Aggregate

[MongoDB Documentation: Aggregate](https://docs.mongodb.com/manual/reference/method/db.collection.aggregate/)


# MS SQL
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/service/ms-sql



<Frame>
  <img alt="MS SQL Node configuration panel" />
</Frame>

## Description

The SQL Node lets you run queries, transactions, or stored procedures on MS SQL Server.

## Run Query

Runs a single SQL query statement.

| Parameter         | Type          | Description                                                                                               |
| ----------------- | ------------- | --------------------------------------------------------------------------------------------------------- |
| MS SQL Connection | List          | Select MS SQL database connection. To create a new connection, click <img alt="add MS SQL connection" />. |
| SQL Query         | CognigyScript | Enter a valid SQL query in the Query field.                                                               |

<AccordionGroup>
  <Accordion title="Result Storage">
    | Parameter                 | Type          | Description                                                                                                                         |
    | ------------------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------- |
    | Where to store the result | List          | <ul><li>**Input** — stores the result in the Input object.</li><li>**Context** — stores the result in the Context object.</li></ul> |
    | Input Key to store Result | CognigyScript | Enter the key where the result should be stored.                                                                                    |
  </Accordion>

  <Accordion title="Advanced">
    | Parameter     | Type   | Description                                                                    |
    | ------------- | ------ | ------------------------------------------------------------------------------ |
    | Stop on Error | Toggle | Lets you configure whether to stop Flow execution when an error occurs or not. |
  </Accordion>
</AccordionGroup>

## Run Transaction

Runs a transaction.

| Parameter         | Type          | Description                                                                                                                                                                                                                                                                                        |
| ----------------- | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| MS SQL Connection | List          | Select MS SQL database connection. To create a new connection, click <img alt="add MS SQL connection" />.                                                                                                                                                                                          |
| Query             | CognigyScript | Enter a valid SQL transaction query in the Query field. You do not need to add the "BEGIN TRANSACTION" and "COMMIT" fields, since this will be done implicitly. If you want to work with Try / catch blocks and do more advanced transactions with rollbacks, use [Run Query](#run-query) instead. |

<AccordionGroup>
  <Accordion title="Result Storage">
    | Parameter                 | Type          | Description                                                                                                                         |
    | ------------------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------- |
    | Where to store the result | List          | <ul><li>**Input** — stores the result in the Input object.</li><li>**Context** — stores the result in the Context object.</li></ul> |
    | Input Key to store Result | CognigyScript | Enter the key where the result should be stored.                                                                                    |
  </Accordion>

  <Accordion title="Advanced">
    | Parameter     | Type   | Description                                                                    |
    | ------------- | ------ | ------------------------------------------------------------------------------ |
    | Stop on Error | Toggle | Lets you configure whether to stop Flow execution when an error occurs or not. |
  </Accordion>
</AccordionGroup>

## Run Stored Procedure

Runs a stored procedure. The stored procedure must exist in the database specified under Connection.

| Parameter             | Type          | Description                                                                                                                                                                                                                                                                                          |
| --------------------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| MS SQL Connection     | List          | Select MS SQL database connection. To create a new connection, click <img alt="add MS SQL connection" />.                                                                                                                                                                                            |
| Stored Procedure Name | CognigyScript | Enter the name of the stored procedure.                                                                                                                                                                                                                                                              |
| Inputs                | JSON          | Enter an array of input parameters of your stored procedure as defined in the database. You can forward parameters from the Context to the database, for example. Parameters must be specified in an array of objects that define the input parameter in the following format: `{name,value,type}`   |
| Outputs               | JSON          | Enter an array of output parameters of your stored procedure as defined in the database. You can forward parameters from the database to the Context, for example. Parameters must be specified in an array of objects that define the output parameter in the following format: `{name,value,type}` |

<AccordionGroup>
  <Accordion title="Result Storage">
    | Parameter                 | Type          | Description                                                                                                                         |
    | ------------------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------- |
    | Where to store the result | List          | <ul><li>**Input** — stores the result in the Input object.</li><li>**Context** — stores the result in the Context object.</li></ul> |
    | Input Key to store Result | CognigyScript | Enter the key where the result should be stored.                                                                                    |
  </Accordion>

  <Accordion title="Advanced">
    | Parameter     | Type   | Description                                                                    |
    | ------------- | ------ | ------------------------------------------------------------------------------ |
    | Stop on Error | Toggle | Lets you configure whether to stop Flow execution when an error occurs or not. |
  </Accordion>
</AccordionGroup>


# Service Overview
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/service/overview



Service Nodes are used to:

* Serve as connectors to APIs and databases. They assist in tasks such as accessing your user database to enhance the customization of your Flow.
* Help to operate Flows that include handover logic.

<CardGroup>
  <Card title="Add Transcript Step" href="/ai/agents/develop/node-reference/service/add-transcript-step">
    The Add Transcript Step Node adds a step for one of the conversation actors to simulate an input or output.
  </Card>

  <Card title="Check Agent Availability" href="/ai/agents/develop/node-reference/service/check-agent-availability">
    The Check Agent Availability Node checks and reports the availability of agents with specific skills, or suggest an alternative response.
  </Card>

  <Card title="Close Handover" href="/ai/agents/develop/node-reference/service/close-handover">
    The Close Handover Node closes the handover conversation on the handover provider side and returns the user to the AI Agent.
  </Card>

  <Card title="Get Token" href="/ai/agents/develop/node-reference/service/get-token">
    The Get Token Node triggers Microsoft SSO authentication for an Azure Bot Services OAuth connection.
  </Card>

  <Card title="Get Transcript" href="/ai/agents/develop/node-reference/service/get-transcript">
    The Get Transcript Node lets you receive a real-time conversation transcript and store the result in the [Input](/ai/agents/develop/ai-agent-memory/input) or [Context](/ai/agents/develop/ai-agent-memory/context) object.
  </Card>

  <Card title="Handover to Human Agent" href="/ai/agents/develop/node-reference/service/handover-to-agent">
    With the Handover to Human Agent Node, AI Agents can hand transfer conversations to human agents in contact centers.
  </Card>

  <Card title="HTTP Request" href="/ai/agents/develop/node-reference/service/http-request">
    The HTTP Request Node executes HTTP requests to other interfaces (APIs).
  </Card>

  <Card title="Invalidate Token" href="/ai/agents/develop/node-reference/service/invalidate-token">
    The Invalidate Token Node signs the user out of an Azure Bot Services OAuth connection and deletes the stored token.
  </Card>

  <Card title="LLM Prompt" href="/ai/agents/develop/node-reference/service/llm-prompt">
    This Node combines the capabilities of the AI Agent Node with support for tools, image recognition, and conversation transcripts.
  </Card>

  <Card title="MongoDB Nodes" href="/ai/agents/develop/node-reference/service/mongo-db">
    A built-in MongoDB connector that you can use directly in the Flow. Includes different operations for MongoDB.
  </Card>

  <Card title="Microsoft SQL Nodes" href="/ai/agents/develop/node-reference/service/ms-sql">
    A built-in Microsoft SQL connector that you can use directly in the Flow. Includes different operations for Microsoft SQL.
  </Card>

  <Card title="Set Handover Inactivity" href="/ai/agents/develop/node-reference/service/set-handover-inactivity">
    The Set Handover Inactivity Node detects and handles user inactivity, freeing up capacity for the human agent team.
  </Card>

  <Card title="Token Status" href="/ai/agents/develop/node-reference/service/token-status">
    The Token Status Node checks whether a valid Microsoft SSO token exists for an Azure Bot Services OAuth connection, and branches the Flow into Granted or Not Granted child Nodes for further execution.
  </Card>

  <Card title="Trigger Function" href="/ai/agents/develop/node-reference/service/trigger-function">
    Use the Trigger Function Node to execute a [Function](/ai/for-developers/functions).
  </Card>
</CardGroup>


# Set Handover Inactivity
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/service/set-handover-inactivity



<Frame>
  <img alt="Set Handover Inactivity Node configuration panel" />
</Frame>

## Description

This Node detects and handles user inactivity, freeing up capacity for the human agent team. To configure this Node, specify the time in minutes during which the user can be inactive.

## Parameters

| Parameter | Type   | Description                                                |
| --------- | ------ | ---------------------------------------------------------- |
| Timer     | Number | The time in minutes during which the user can be inactive. |

## How to Use

To use the Node, first configure the [Lookup](/ai/agents/develop/node-reference/logic/lookup) Node to enable different inactivity behaviors.

<Frame>
  <img alt="Lookup Node configured for handover inactivity behavior" />
</Frame>

To manage the user inactivity time, use the `input.handover.inactivityCount >= <times inactivity was reached>` condition in the **Condition** field of the [If](/ai/agents/develop/node-reference/logic/if) Node. For example, `input.handover.inactivityCount >= 2`. You can also adjust the count value as needed.

<Frame>
  <img alt="If Node with handover inactivity count condition" />
</Frame>

To improve the user experience, after the Set Handover Inactivity Node, add a [Say](/ai/agents/develop/node-reference/basic/say) Node message that prompts the user with a message such as `Are you still there? The conversation will be automatically closed soon`.

<Frame>
  <img alt="Say Node with inactivity prompt message for the user" />
</Frame>

Using the [Say](/ai/agents/develop/node-reference/basic/say#handover-settings) Node, you can determine if the agent, user, or both receive messages from the AI Agent after the time has expired.

To close a handover based on the inactivity event, use the [Close Handover](/ai/agents/develop/node-reference/service/close-handover) Node. Alternatively, you can initiate another inactivity behavior if necessary.

## More Information

* [User Inactivity Detection and Handling](/ai/escalate/handover-features/user-inactivity-detection)
* [Agent Handover](/ai/agents/develop/node-reference/service/handover-to-agent)
* [Set Handover Inactivity](/ai/agents/develop/node-reference/service/set-handover-inactivity)
* [Close Handover](/ai/agents/develop/node-reference/service/close-handover)


# Token Status
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/service/token-status

The Token Status Node checks whether a valid Microsoft SSO token exists for an Azure Bot Services OAuth connection, and branches the Flow into Granted or Not Granted child Nodes for further execution.

<a href="/release-notes/2026.17"><Badge>Updated in 2026.17</Badge></a>

<Frame>
  <img alt="Token Status Node configuration panel" />
</Frame>

## Description

The Token Status Node checks whether a valid Microsoft SSO token exists for an [Azure Bot Services Endpoint](/ai/agents/deploy/endpoint-reference/azure-bot-services)'s OAuth connection, and branches the Flow into **Granted** or **Not Granted** child Nodes for further execution.

Place this Node after a [Get Token](/ai/agents/develop/node-reference/service/get-token) Node to check whether the sign-in completed, before continuing with an action that requires the token.

## Parameters

| Parameter            | Type          | Description                                                                                                                                                                                                                                              |
| -------------------- | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Overwrite Connection | Toggle        | Sets a different OAuth connection than the one configured in the **SSO Connection Name** parameter in the Endpoint settings. Use this toggle if you use multiple SSO connections with your Azure Bot Services. This parameter is deactivated by default. |
| Connection Name      | CognigyScript | This parameter appears when you activate **Overwrite Connection**. Sets the name of the OAuth connection to check the token for.                                                                                                                         |

## More Information

* [Azure Bot Services Endpoint: Multiple SSO Connections](/ai/agents/deploy/endpoint-reference/azure-bot-services#set-up-multiple-sso-connections-optional-accordion-title)
* [Get Token](/ai/agents/develop/node-reference/service/get-token)
* [Invalidate Token](/ai/agents/develop/node-reference/service/invalidate-token)
* [Cognigy Help Center: SSO and Graph API Nodes](https://support.cognigy.com/hc/en-us/articles/6422300559388-Azure-Bot-Services-Advanced-Deployment-with-SSO-Microsoft-Teams-SSO-and-Graph-API-Nodes)


# Trigger Function
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/service/trigger-function



<Frame>
  <img alt="Trigger Function Node configuration panel" />
</Frame>

## Description

This Node executes a [Function](/ai/for-developers/functions).

If the Function was successfully scheduled, it will execute the **On Scheduled** branch, otherwise it will execute the **On Scheduling Failed** branch.

## Parameters

| Parameter           | Type | Description                            |
| ------------------- | ---- | -------------------------------------- |
| Function            | List | Select the Function to be executed.    |
| Function Parameters | JSON | Enter the parameters for the Function. |


# Call Recording
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/audiocodes/call-recording



<Frame>
  <img alt="AudioCodes Call Recording Node configuration panel" />
</Frame>

## Description

<Warning>
  * As of Cognigy 4.96.0, AudioCodes Nodes and Generic Voice Nodes are deprecated. Flows with these Nodes continue to work, and you can still edit and clone them as well as export them as a Package. However, you can't create new ones in Cognigy 4.96 or later.
  * The full removal of AudioCodes Nodes is planned for Q2 2026. After the removal, AI Agents using these Nodes will stop working properly. We strongly recommend migrating to [Voice Gateway Nodes](/ai/agents/develop/node-reference/voice/voice-gateway/overview) before the removal date.
</Warning>

The Call Recording Node controls the voice recording capabilities of the conversation.

Recording is done by the SBC (SRC) using the SIPRec protocol and functioning as the Session Recording Client (SRC).
The session recording server (SRS) can be, for example, AudioCodes SmartTAP recording solution or any other SRS.

| Parameter             | Type          | Description                                                                                               |
| --------------------- | ------------- | --------------------------------------------------------------------------------------------------------- |
| Activity              | Select        | Whether to start or stop a call recording.                                                                |
| Call Recording Server | CognigyScript | Defines the SRS as an IP Group name (as configured on the SBC) to record the call.                        |
| Call Recording ID     | CognigyScript | Defines the recording ID session, which is forwarded by the SBC to the SRS in the XML body.               |
| Destination Username  | CognigyScript | Defines the username that is used in the SIP Request-URI and To headers of the INVITE request for SIPRec. |


# Hang Up
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/audiocodes/hangup



<Frame>
  <img alt="AudioCodes Hang Up Node configuration panel" />
</Frame>

## Description

<Warning>
  * As of Cognigy 4.96.0, AudioCodes Nodes and Generic Voice Nodes are deprecated. Flows with these Nodes continue to work, and you can still edit and clone them as well as export them as a Package. However, you can't create new ones in Cognigy 4.96 or later.
  * The full removal of AudioCodes Nodes is planned for Q2 2026. After the removal, AI Agents using these Nodes will stop working properly. We strongly recommend migrating to [Voice Gateway Nodes](/ai/agents/develop/node-reference/voice/voice-gateway/overview) before the removal date.
</Warning>

The Hangup Node will instruct AudioCodes to end the call.

| Parameter | Type          | Description                                              |
| --------- | ------------- | -------------------------------------------------------- |
| Reasons   | CognigyScript | The reason for hanging up. Will show in AudioCodes logs. |


# AudioCodes
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/audiocodes/overview



<Warning>
  * As of Cognigy 4.96.0, AudioCodes Nodes and Generic Voice Nodes are deprecated. Flows with these Nodes continue to work, and you can still edit and clone them as well as export them as a Package. However, you can't create new ones in Cognigy 4.96 or later.
  * The full removal of AudioCodes Nodes is planned for Q2 2026. After the removal, AI Agents using these Nodes will stop working properly. We strongly recommend migrating to [Voice Gateway Nodes](/ai/agents/develop/node-reference/voice/voice-gateway/overview) before the removal date.
</Warning>

AudioCodes Nodes help you build a voice AI Agent using [AudioCodes configuration](/ai/agents/deploy/endpoint-reference/audiocodes).

### [Call Recording](/ai/agents/develop/node-reference/voice/audiocodes/call-recording)

This Node controls the voice recording capabilities of the conversation.

### [Hang Up](/ai/agents/develop/node-reference/voice/audiocodes/hangup)

This Node instructs AudioCodes to end the call.

### [Play URL](/ai/agents/develop/node-reference/voice/audiocodes/play-url)

This Node lets you retrieve a WAV file from a URL location and play it back to the user.

### [Send Message](/ai/agents/develop/node-reference/voice/audiocodes/send-message)

This Node sends a text (optionally with SSML) to AudioCodes, which then turns this text into speech (TTS) and outputs it to the caller.

### [Send Metadata](/ai/agents/develop/node-reference/voice/audiocodes/send-meta-data)

This Node sends metadata via SIP INFO messages to the connected SIP trunk.

### [Set Session Parameters](/ai/agents/develop/node-reference/voice/audiocodes/set-session-params)

This Node configures the AudioCodes Session. All parameters set within this Node are valid for the remainder of the session unless overwritten, either directly or by activity parameters.

### [Transfer](/ai/agents/develop/node-reference/voice/audiocodes/transfer-vg)

This Node enables the forwarding of an active call to a Tel or SIP target.


# Play URL
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/audiocodes/play-url



<Frame>
  <img alt="AudioCodes Play URL Node configuration panel" />
</Frame>

## Description

<Warning>
  * As of Cognigy 4.96.0, AudioCodes Nodes and Generic Voice Nodes are deprecated. Flows with these Nodes continue to work, and you can still edit and clone them as well as export them as a Package. However, you can't create new ones in Cognigy 4.96 or later.
  * The full removal of AudioCodes Nodes is planned for Q2 2026. After the removal, AI Agents using these Nodes will stop working properly. We strongly recommend migrating to [Voice Gateway Nodes](/ai/agents/develop/node-reference/voice/voice-gateway/overview) before the removal date.
</Warning>

The Play URL Node lets you retrieve a WAV file from a URL location and play it back to the user.

| Parameter               | Type          | Description                                                                                         |
| ----------------------- | ------------- | --------------------------------------------------------------------------------------------------- |
| Media URL               | CognigyScript | The URL to the WAV file. Must be accessible to your AudioCodes instance.                            |
| Media Format            | Select        | Options:<br />- wav/lpcm15<br />- raw/lpcm16                                                        |
| Cache Audio             | Toggle        | Whether to cache the audio file or not. Caching will make future playbacks quicker. Default: false. |
| Set Activity Parameters | Toggle        | Option to add parameters which will be valid for this activity only                                 |

<Accordion title="This Node supports Activity Parameters">
  To learn more about activity parameters, visit our Help Center at: [https://support.cognigy.com/hc/en-us/articles/360017413959](https://support.cognigy.com/hc/en-us/articles/360017413959)
</Accordion>


# Send Message
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/audiocodes/send-message



<Frame>
  <img alt="AudioCodes Send Message Node configuration panel" />
</Frame>

## Description

<Warning>
  * As of Cognigy 4.96.0, AudioCodes Nodes and Generic Voice Nodes are deprecated. Flows with these Nodes continue to work, and you can still edit and clone them as well as export them as a Package. However, you can't create new ones in Cognigy 4.96 or later.
  * The full removal of AudioCodes Nodes is planned for Q2 2026. After the removal, AI Agents using these Nodes will stop working properly. We strongly recommend migrating to [Voice Gateway Nodes](/ai/agents/develop/node-reference/voice/voice-gateway/overview) before the removal date.
</Warning>

The Send Message Node sends a text (optionally with SSML) to AudioCodes, which then turns this text into speech (TTS) and outputs it to the caller.

| Parameter               | Type                    | Description                                                                                                |
| ----------------------- | ----------------------- | ---------------------------------------------------------------------------------------------------------- |
| Text                    | CognigyScript with SSML | The text to send back to AudioCodes. Can contain SSML (either entered manually or via the provided editor) |
| Set Activity Parameters | Toggle                  | Option to add parameters which will be valid for this activity only.                                       |

<Accordion title="This Node supports Activity Parameters">
  To learn more about activity parameters, visit our Help Center at: [https://support.cognigy.com/hc/en-us/articles/360017413959](https://support.cognigy.com/hc/en-us/articles/360017413959)
</Accordion>


# Send Metadata
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/audiocodes/send-meta-data



<Frame>
  <img alt="AudioCodes Send Metadata Node configuration panel" />
</Frame>

## Description

<Warning>
  * As of Cognigy 4.96.0, AudioCodes Nodes and Generic Voice Nodes are deprecated. Flows with these Nodes continue to work, and you can still edit and clone them as well as export them as a Package. However, you can't create new ones in Cognigy 4.96 or later.
  * The full removal of AudioCodes Nodes is planned for Q2 2026. After the removal, AI Agents using these Nodes will stop working properly. We strongly recommend migrating to [Voice Gateway Nodes](/ai/agents/develop/node-reference/voice/voice-gateway/overview) before the removal date.
</Warning>

The Send Metadata Node will send metadata via SIP INFO messages to the connected SIP trunk.

| Parameter | Type | Description                                                  |
| --------- | ---- | ------------------------------------------------------------ |
| Metadata  | JSON | The metadata to send via SIP INFO. See below for an example. |

```json theme={null}
{
  "myParamName": "myParamValue"
}
```


# Set Session Params
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/audiocodes/set-session-params



<Frame>
  <img alt="AudioCodes Set Session Parameters Node configuration panel" />
</Frame>

## Description

<Warning>
  * As of Cognigy 4.96.0, AudioCodes Nodes and Generic Voice Nodes are deprecated. Flows with these Nodes continue to work, and you can still edit and clone them as well as export them as a Package. However, you can't create new ones in Cognigy 4.96 or later.
  * The full removal of AudioCodes Nodes is planned for Q2 2026. After the removal, AI Agents using these Nodes will stop working properly. We strongly recommend migrating to [Voice Gateway Nodes](/ai/agents/develop/node-reference/voice/voice-gateway/overview) before the removal date.
</Warning>

The Set Session Params Node configures the AudioCodes Session. All parameters set here are valid for the remainder of the session unless overwritten, either directly or by activity parameters.

To learn more about session and activity parameters, go to the [Cognigy Help Center](https://support.cognigy.com/hc/en-us/articles/360017413959).


# Transfer
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/audiocodes/transfer-vg



<Frame>
  <img alt="AudioCodes Transfer Node configuration panel" />
</Frame>

## Description

<Warning>
  * As of Cognigy 4.96.0, AudioCodes Nodes and Generic Voice Nodes are deprecated. Flows with these Nodes continue to work, and you can still edit and clone them as well as export them as a Package. However, you can't create new ones in Cognigy 4.96 or later.
  * The full removal of AudioCodes Nodes is planned for Q2 2026. After the removal, AI Agents using these Nodes will stop working properly. We strongly recommend migrating to [Voice Gateway Nodes](/ai/agents/develop/node-reference/voice/voice-gateway/overview) before the removal date.
</Warning>

The Transfer Node lets you pass an active call to a tel or sip target.

| Parameter            | Type          | Description                                                                                                                     |
| -------------------- | ------------- | ------------------------------------------------------------------------------------------------------------------------------- |
| Reasons              | CognigyScript | The reason for the handover. Will show in AudioCode logs.                                                                       |
| Target               | CognigyScript | The target URI to pass the call to. Typically either a "tel" or "sip" URI.                                                      |
| Referral URL         | CognigyScript | Defines the party (URL) who initiated the call referral                                                                         |
| Transfer SIP Headers | JSON          | Array of objects listing SIP headers that should be sent to the transferee. Each object comprises a name and a value attribute. |


# Barge In - Config
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/generic/barge-in-config



<Frame>
  <img alt="Generic Barge In Config Node configuration panel" />
</Frame>

## Description

<Warning>
  * As of Cognigy 4.96.0, AudioCodes Nodes and Generic Voice Nodes are deprecated. Flows with these Nodes continue to work, and you can still edit and clone them as well as export them as a Package. However, you can't create new ones in Cognigy 4.96 or later.
  * The full removal of AudioCodes Nodes is planned for Q2 2026. After the removal, AI Agents using these Nodes will stop working properly. We strongly recommend migrating to [Voice Gateway Nodes](/ai/agents/develop/node-reference/voice/voice-gateway/overview) before the removal date.
</Warning>

The Barge In - Config Node is a session config that enables the interruption of the voice agent.

When executed, the settings will apply for the remainder of the session.

| Parameter                | Type          | Description                                                                                                                                     |
| ------------------------ | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------- |
| Barge In On Speech       | Toggle        | Enables users to interrupt the voice agent with speech.                                                                                         |
| Barge In Minimum Words   | Slider        | Defines the minimum number of words that the user must say for the Voice Gateway to consider it a barge in.                                     |
| Barge In On DTMF         | Toggle        | Enables users to interrupt the voice agent with DTMF digits. Capturing of DTMF signals is enabled by default.                                   |
| DTMF Inter Digit Timeout | Number        | Defines the timeout in between collected DTMF digits.                                                                                           |
| DTMF Max Digits          | Number        | Defines the maximum number of digits which can be entered by the user, the digits are being submitted automatically once this limit is reached. |
| DTMF Submit Digit        | CognigyScript | Defines the DTMF submit digit which will be used for submitting the previously entered digits. This overrides the minimum digits validation.    |


# Continuous ASR - Config
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/generic/continuous-asr-config



<a href="/release-notes/4.90"><Badge>Updated in 4.90</Badge></a>

<Frame>
  <img alt="Generic Continuous ASR Config Node configuration panel" />
</Frame>

## Description

<Warning>
  * As of Cognigy 4.96.0, AudioCodes Nodes and Generic Voice Nodes are deprecated. Flows with these Nodes continue to work, and you can still edit and clone them as well as export them as a Package. However, you can't create new ones in Cognigy 4.96 or later.
  * The full removal of AudioCodes Nodes is planned for Q2 2026. After the removal, AI Agents using these Nodes will stop working properly. We strongly recommend migrating to [Voice Gateway Nodes](/ai/agents/develop/node-reference/voice/voice-gateway/overview) before the removal date.
</Warning>

Continuous ASR allows the system to combine multiple STT recognitions from the user and send them as a single textual message to the voice agent.

Once the Node is executed, the settings will remain active for the rest of the session.

| Parameter                   | Type          | Description                                                                                                                                                   |
| --------------------------- | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Enable Continuous ASR       | Toggle        | Enables or disables Continuous ASR.                                                                                                                           |
| Continuous ASR Submit Digit | CognigyScript | Defines a special DTMF key that sends the accumulated recognitions to the Flow.                                                                               |
| Continuous ASR Timeout      | Number        | Defines the number of milliseconds of silence before the accumulated recognitions are sent to the Flow. The default value is 3000. The minimum value is 2000. |


# DTMF Collect - Config
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/generic/dtmf-collect-config



<Frame>
  <img alt="Generic DTMF Collect Config Node configuration panel" />
</Frame>

## Description

<Warning>
  * As of Cognigy 4.96.0, AudioCodes Nodes and Generic Voice Nodes are deprecated. Flows with these Nodes continue to work, and you can still edit and clone them as well as export them as a Package. However, you can't create new ones in Cognigy 4.96 or later.
  * The full removal of AudioCodes Nodes is planned for Q2 2026. After the removal, AI Agents using these Nodes will stop working properly. We strongly recommend migrating to [Voice Gateway Nodes](/ai/agents/develop/node-reference/voice/voice-gateway/overview) before the removal date.
</Warning>

Enables capturing of DTMF signals by the voice agent during the entire call session.

When executed, the settings will apply for the remainder of the session.

| Parameter                | Type          | Description                                                                                                                                    |
| ------------------------ | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------- |
| Capture DTMF signals     | Toggle        | Enables capturing of DTMF signals by the voice agent                                                                                           |
| DTMF Inter Digit Timeout | Number        | Defines the timeout in between collected DTMF digits                                                                                           |
| DTMF Max Digits          | Number        | Defines the maximum number of digits which can be entered by the user, the digits are being submitted automatically once this limit is reached |
| DTMF Submit Digit        | CognigyScript | Defines the DTMF submit digit which will be used for submitting the previously entered digits. This overrides the minimum digits validation    |


# Hang Up
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/generic/hang-up



<Frame>
  <img alt="Generic Hang Up Node configuration panel" />
</Frame>

## Description

<Warning>
  * As of Cognigy 4.96.0, AudioCodes Nodes and Generic Voice Nodes are deprecated. Flows with these Nodes continue to work, and you can still edit and clone them as well as export them as a Package. However, you can't create new ones in Cognigy 4.96 or later.
  * The full removal of AudioCodes Nodes is planned for Q2 2026. After the removal, AI Agents using these Nodes will stop working properly. We strongly recommend migrating to [Voice Gateway Nodes](/ai/agents/develop/node-reference/voice/voice-gateway/overview) before the removal date.
</Warning>

The Hang Up Node will instruct the system to end the call.

| Parameter          | Type | Description                                            |
| ------------------ | ---- | ------------------------------------------------------ |
| Reason for hang up | Text | The reason for hanging up, will be shown in VG/AC logs |


# Mute Speech Input
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/generic/mute-speech-input



<a href="/release-notes/4.54"><Badge>Updated in 4.54</Badge></a>

<Frame>
  <img alt="Generic Mute Speech Input Node configuration panel" />
</Frame>

<Warning>
  * As of Cognigy 4.96.0, AudioCodes Nodes and Generic Voice Nodes are deprecated. Flows with these Nodes continue to work, and you can still edit and clone them as well as export them as a Package. However, you can't create new ones in Cognigy 4.96 or later.
  * The full removal of AudioCodes Nodes is planned for Q2 2026. After the removal, AI Agents using these Nodes will stop working properly. We strongly recommend migrating to [Voice Gateway Nodes](/ai/agents/develop/node-reference/voice/voice-gateway/overview) before the removal date.
</Warning>

## Description

<div />

This Node controls when speech user inputs can be collected in your Flow.
It prevents speech interruptions in specific cases, such as [xApp Flows](/ai/agents/develop/node-reference/xApp/overview), allowing for a smoother conversation experience.
By enabling or disabling speech gathering as needed,
you can maintain better control and ensure a seamless workflow in your application.

## Settings

<Warning>
  Do not turn off a speech user input throughout the call to receive additional messages from the user, especially when the AI Agent needs to send multiple messages with time gaps in between.
</Warning>

| Parameter         | Type   | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
| ----------------- | ------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Mute Speech Input | Toggle | Disables speech input recognition for the current call. By default, the setting is deactivated, allowing the Nodes below the Mute Speech Input Node to accept speech inputs. <br /><br /> When this setting is activated, the Nodes below the Mute Speech Input Node will not accept speech inputs or digits as the answer. It affects all speech inputs made through the following Nodes:<br /> - Nodes with Set Activity Parameters. <br /> - Say and Play Nodes with enabled barge-in can no longer be interrupted.<br /> - Question and Optional Question Nodes. |

## Example

In the following example, the Mute Node allows you to control when speech input is accepted or disabled within the Flow:

1. Say Node: `Hello, you will receive an xApp`.
2. xApp: Init Session Node.
3. Mute Speech Input (enabled).
4. Question Node: `Please rate your experience with the xApp on a scale of 1 to 5`.<br />When the Mute Node STT is enabled, the question node will not accept speech inputs or digits as the answer. The user must provide the rating through other means, such as selecting a number using buttons or typing the response. This process ensures that the speech gathering is temporarily turned off during the rating question to avoid unintended interference from background speech.
5. Mute Node STT (disabled). <br /> After the rating question, the Mute Node STT is disabled, allowing the subsequent Nodes to accept speech inputs normally.

## More Information

* [Voice Gateway Mute Speech Input Node](/ai/agents/develop/node-reference/voice/voice-gateway/mute-speech-input)
* [xApps](/xApps/overview)
* [xApp Nodes](/ai/agents/develop/node-reference/xApp/overview)


# Generic Voice Nodes
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/generic/overview



<Warning>
  * As of Cognigy 4.96.0, AudioCodes Nodes and Generic Voice Nodes are deprecated. Flows with these Nodes continue to work, and you can still edit and clone them as well as export them as a Package. However, you can't create new ones in Cognigy 4.96 or later.
  * The full removal of AudioCodes Nodes is planned for Q2 2026. After the removal, AI Agents using these Nodes will stop working properly. We strongly recommend migrating to [Voice Gateway Nodes](/ai/agents/develop/node-reference/voice/voice-gateway/overview) before the removal date.
</Warning>

The Generic Voice Nodes are a set of Nodes for creating voice Flows that are not specific to one Endpoint.
These Voice Nodes are compatible with both voice integrations: AudioCodes and Voice Gateway.

### [Barge In - Config](/ai/agents/develop/node-reference/voice/generic/barge-in-config)

This Node is a session config that enables the interruption of the voice AI Agent.

When executed, the settings will apply for the remainder of the session.

### [Continuous ASR - Config](/ai/agents/develop/node-reference/voice/generic/continuous-asr-config)

Continuous ASR enables the system to concatenate multiple STT recognitions of the user
and then send them as a single textual message to the voice AI Agent.

### [DTMF Collect - Config](/ai/agents/develop/node-reference/voice/generic/dtmf-collect-config)

Enables capturing of DTMF signals by the voice AI Agent during the entire call session.

When executed, the settings will apply for the remainder of the session.

### [Hang Up](/ai/agents/develop/node-reference/voice/generic/hang-up)

This Node instructs the system to end the call.

### [Mute Speech Input](/ai/agents/develop/node-reference/voice/generic/mute-speech-input)

This Node controls when speech user inputs can be collected in your Flow.

### [Play URL](/ai/agents/develop/node-reference/voice/generic/play-url)

The Play URL Node lets you stream a sound file into the call with the ability to loop it. MP3 and WAV encoded files are supported.

### [Send Metadata](/ai/agents/develop/node-reference/voice/generic/send-metadata)

The Send Metadata Node sends metadata via SIP INFO messages to the connected SIP trunk.

### [Session Speech Parameters - Config](/ai/agents/develop/node-reference/voice/generic/session-speech-parameters-config)

This Node enables the change of speech parameters during the Flow.

### [Transfer](/ai/agents/develop/node-reference/voice/generic/transfer)

This Node lets you pass an active call to a tel or sip target.

### [User Input Timeout - Config](/ai/agents/develop/node-reference/voice/generic/user-input-timeout-config)

This Node defines the action that the AI Agent should take if it does not receive any input from the user within a certain timeframe.


# Play URL
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/generic/play-url



<Frame>
  <img alt="Generic Play URL Node configuration panel" />
</Frame>

## Description

<Warning>
  * As of Cognigy 4.96.0, AudioCodes Nodes and Generic Voice Nodes are deprecated. Flows with these Nodes continue to work, and you can still edit and clone them as well as export them as a Package. However, you can't create new ones in Cognigy 4.96 or later.
  * The full removal of AudioCodes Nodes is planned for Q2 2026. After the removal, AI Agents using these Nodes will stop working properly. We strongly recommend migrating to [Voice Gateway Nodes](/ai/agents/develop/node-reference/voice/voice-gateway/overview) before the removal date.
</Warning>

The Play URL Node lets you stream a sound file into the call.

This Node is useful for playing advertisements, pre-recorded jingles, or other sound files for callers. It supports MP3 and WAV encoded files.

## Parameters

| Parameter               | Type          | Description                                                                                                                                                                                                                                                                                                                                                                             |
| ----------------------- | ------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Media URL               | CognigyScript | Accepts URL links to MP3 and WAV tracks. The URL should point to the storage location where you want to host it and doesn't need to include the `.mp3` or `.wav` extension. For example, `https://abc.xyz/music.mp3` or `https://audio.jukehost.co.uk/N5pnlULbup8KabGRE7dsGwHTeIZAwWdr`. Make sure the file is publicly accessible.                                                     |
| Set Activity Parameters | Toggle        | Enables parameters which will be valid for this activity only. The available parameters are: <br /> - [Recognizer (STT)](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details) <br /> - [Barge In](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details) <br /> - [DTMF](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details) |

## More Information

* [AudioCodes Play URL Node](/ai/agents/develop/node-reference/voice/audiocodes/play-url)
* [Voice Gateway Play Node](/ai/agents/develop/node-reference/voice/voice-gateway/play)
* [Parameter Details](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details)


# Send Metadata
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/generic/send-metadata



<a href="/release-notes/4.45"><Badge>Updated in 4.45</Badge></a>

<Frame>
  <img alt="Generic Send Metadata Node configuration panel" />
</Frame>

## Description

<Warning>
  * As of Cognigy 4.96.0, AudioCodes Nodes and Generic Voice Nodes are deprecated. Flows with these Nodes continue to work, and you can still edit and clone them as well as export them as a Package. However, you can't create new ones in Cognigy 4.96 or later.
  * The full removal of AudioCodes Nodes is planned for Q2 2026. After the removal, AI Agents using these Nodes will stop working properly. We strongly recommend migrating to [Voice Gateway Nodes](/ai/agents/develop/node-reference/voice/voice-gateway/overview) before the removal date.
</Warning>

The Send Metadata Node sends metadata via SIP INFO messages to the connected SIP trunk.

| Parameter | Type | Description                        |
| --------- | ---- | ---------------------------------- |
| Metadata  | JSON | The metadata to send via SIP INFO. |

Example:

```json theme={null}
{
  "myParamName": "myParamValue"
}
```

<Note>
  This Node supports Activity Parameters. You can find more information about activity parameters on the [Parameter Details](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details) page.
</Note>


# Session Speech Parameters Config
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/generic/session-speech-parameters-config



<a href="/release-notes/4.65"><Badge>Updated in 4.65</Badge></a>

<Frame>
  <img alt="Generic Session Speech Parameters Config Node configuration panel" />
</Frame>

## Description

<Warning>
  * As of Cognigy 4.96.0, AudioCodes Nodes and Generic Voice Nodes are deprecated. Flows with these Nodes continue to work, and you can still edit and clone them as well as export them as a Package. However, you can't create new ones in Cognigy 4.96 or later.
  * The full removal of AudioCodes Nodes is planned for Q2 2026. After the removal, AI Agents using these Nodes will stop working properly. We strongly recommend migrating to [Voice Gateway Nodes](/ai/agents/develop/node-reference/voice/voice-gateway/overview) before the removal date.
</Warning>

This Node enables the change of speech parameters during the Flow.

When executed, the settings will apply for the remainder of the session.

## Synthesizer - Text-To-Speech - Settings

The TTS settings can be chosen from a pre-filled dropdown for Microsoft Azure, AWS, Google, Nuance, or a custom vendor.

| Parameter                  | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |
| -------------------------- | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| TTS Vendor                 | Dropdown      | Defines the desired TTS Vendor. You can select a custom vendor.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
| Custom (Vendor)            | CognigyScript | Allows for specifying a TTS Vendor, which is not in the dropdown list. This option is only available on Voice Gateway. For preinstalled providers, use all lowercase letters, for example, `microsoft`, `google`, `aws`. For custom providers, use the name that you specified on the [Speech Service](/voice-gateway/webapp/speech-services) page in Voice Gateway. <br /><br /> The **Custom** field appears if you selected **Custom** from the **TTS Vendor** list.                                                                                                                                                                                  |
| TTS Language               | Dropdown      | Defines the language of the Voice AI Agent output.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
| Custom (Language)          | CognigyScript | Allows for choosing a TTS language, which is not in the dropdown list. Defines the language of the AI Agent output. The format to use depends on the TTS Vendor, for example, `de-DE`, `fr-FR`, `en-US`. <br /><br /> The **Custom** field appears if you selected **Custom** from the **TTS Language** list.                                                                                                                                                                                                                                                                                                                                            |
| TTS Voice                  | Dropdown      | Defines the voice that should be used for the voice AI Agent output.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
| Custom (Voice)             | CognigyScript | Allows for choosing a TTS voice, which is not in the dropdown list. This setting can be the case for region-specific voices. The format to use depends on the TTS Vendor, for example, `de-DE-ConradNeural`.                                                                                                                                                                                                                                                                                                                                                                                                                                             |
| Enable Advanced TTS Config | Toggle        | Enables the addition of a URL for an Azure Custom Voice Endpoint. <br /><br /> This setting is only available and appears for TTS Vendor Microsoft.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
| Disable TTS Audio Caching  | Toggle        | Disables TTS audio caching.<br /><br /> By default, the setting is deactivated. In this case, previously requested TTS audio results are stored in the AI Agent cache. When a new TTS request is made and the audio text has been previously requested, the AI Agent retrieves the cached result instead of sending another request to the TTS provider.<br /><br /> When the setting is activated, the AI Agent caches TTS results but doesn't use them. In this case, each request is directly sent to your speech provider.  <br /><br /> Note that disabling caching can increase TTS costs. For detailed information, contact your speech provider. |

## Recognizer - Speech-To-Text - Settings

The STT settings can be chosen from a prefilled dropdown for supported vendors, or a custom vendor.

| Parameter                       | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
| ------------------------------- | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| STT Vendor                      | Dropdown      | Defines the desired STT Vendor. You can select a custom vendor.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
| Custom (Vendor)                 | CognigyScript | Allows for specifying an STT Vendor, which is not in the dropdown list. Select the desired STT Vendor. This option is only available on Voice Gateway. For preinstalled providers, use all lowercase letters like `microsoft`, `google`, `aws`. For custom providers, use the name that you specified on the [Speech Service page](/voice-gateway/webapp/speech-services) in Voice Gateway. <br /><br /> The **Custom** field appears if you selected **Custom** from the **STT Vendor** list.                                                                                                                                                                                                   |
| STT Language                    | Dropdown      | Defines the language that should be recognized.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
| Custom (Language)               | CognigyScript | Allows for choosing an STT language, which is not in the dropdown list. This setting can be the case for region-specific voices. The format to use depends on the STT Vendor, for example, `de-DE`, `fr-FR`, `en-US`. <br /><br /> The **Custom** field appears if you selected **Custom** from the **STT Language** list.                                                                                                                                                                                                                                                                                                                                                                       |
| STT Hints                       | Text          | Array of words or phrases to assist speech detection. If you want to use multiple hints, enter each hint into a separate input field. For instance, you can enter `Germany` in the first field, `France` in the second field, and `Netherlands` in the third field. The STT provider will receive the data in array format: \["Germany", "France", "Netherlands"]. <br /> Note: This requires support from the STT engine. The field is not available for the Nuance speech vendor.                                                                                                                                                                                                              |
| Dynamic Hints                   | CognigyScript | Uses context or input for adding array hints. For example, <code>\{\{context.hints}}</code> or <code>\{\{input.hints}}</code>.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
| Google Model                    | Dropdown      | This parameter is active only when Google is selected in the STT Vendor setting.<br /><br />Uses one of Google Cloud Speech-to-Text transcription models. The default value is `latest_short`. For a detailed list of Google models, refer to the [Transcription models](https://cloud.google.com/speech-to-text/docs/transcription-model#transcription_models) section in the Google Documentation. Keep in mind that the `default` value is a Google Model type that can be used if other models don't suit your use case.                                                                                                                                                                     |
| Enable Voice Activity Detection | Toggle        | Delays connection to cloud recognizer until the speech is detected.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |
| Disable STT Punctuation         | Toggle        | This parameter appears only when Google, Deepgram, or Deepgram Flux is selected in the STT Vendor setting.<br /><br /> Prevents the transcription returned to the AI Agent from including punctuation marks.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
| Deepgram Model                  | Dropdown      | This parameter is active only when Deepgram is selected in the STT Vendor setting. <br /><br /> Choose a model for processing submitted audio. Each model is associated with a tier. Ensure that the selected tier is available for the chosen STT language. For detailed information about Deepgram models, refer to the [Deepgram documentation](https://developers.deepgram.com/docs/model).                                                                                                                                                                                                                                                                                                  |
| Endpointing                     | Toggle        | This parameter is active only when Deepgram is selected in the STT Vendor setting. <br /><br /> Deepgram's Endpointing feature watches streaming audio for long pauses that signal the end of speech. When it spots an endpoint, it finalizes predictions and returns the transcript, marking it as complete with the `speech_final` parameter set to `true`. For detailed information about Deepgram Endpointing, refer to the [Deepgram documentation](https://developers.deepgram.com/docs/endpointing).<br /><br />The duration for detecting the end of speech is preconfigured with a default value (10 milliseconds). If you want to change this value, use the Endpointing Time setting. |
| Endpointing Time                | Number        | This parameter is active only when Deepgram is selected in the STT Vendor setting and the Endpointing toggle is enabled. <br /><br /> Customize the duration (in milliseconds) for detecting the end of speech. The default is 10 milliseconds of silence. Transcripts are sent after detecting silence, and the system waits until the speaker resumes or the required silence time is reached. Once either condition is met, a transcript is sent back with `speech_final` set to `true`.                                                                                                                                                                                                      |
| Smart Formatting                | Toggle        | This parameter is active only when Deepgram is selected in the STT Vendor setting. <br /><br /> Deepgram's Smart Format feature applies additional formatting to transcripts to optimize them for human readability. Smart Format capabilities vary between models. When Smart Formatting is turned on, Deepgram will always apply the best-available formatting for your chosen model, tier, and language combination. For detailed examples, refer to the [Deepgram documentation](https://developers.deepgram.com/docs/smart-format).                                                                                                                                                         |
| Enable Advanced TTS Config      | Toggle        | Enables the addition of an ID for an Azure's Custom Speech model deployment.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
| Enable Audio Logging            | Toggle        | Enables recording and logging of audio from the user on Azure.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |


# Transfer
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/generic/transfer



<a href="/release-notes/4.64"><Badge>Updated in 4.65</Badge></a>

<Frame>
  <img alt="Generic Transfer Node configuration panel" />
</Frame>

## Description

<Warning>
  * As of Cognigy 4.96.0, AudioCodes Nodes and Generic Voice Nodes are deprecated. Flows with these Nodes continue to work, and you can still edit and clone them as well as export them as a Package. However, you can't create new ones in Cognigy 4.96 or later.
  * The full removal of AudioCodes Nodes is planned for Q2 2026. After the removal, AI Agents using these Nodes will stop working properly. We strongly recommend migrating to [Voice Gateway Nodes](/ai/agents/develop/node-reference/voice/voice-gateway/overview) before the removal date.
</Warning>

The Transfer Node lets you pass an active call to a tel or sip target.

| Parameter            | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
| -------------------- | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| Reason               | CognigyScript | The reason for the handover. Will show in VG/AC logs                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
| Transfer Destination | CognigyScript | E.164 syntax or a SIP URI are supported                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
| Referred By          | String        | This parameter is optional.<br /><br /> This setting allows you to change the original Referred By value, which can be a SIP URI or a user identifier such as a phone number. To define the Referred By value, you can use the following patterns:<br />- **SIP URI** - `sip:[referred-by]@custom.domain.com`. In this case, the entire SIP URI will be sent as the Referred-By header. Example: `"Referred-by": "sip:CognigyOutbound@custom.domain.com"`.<br />- **User Identifier** - `sip:[referred-by]@[SIP FROM Domain from carrier config]`. Example, `"Referred-By": "sip:CognigyOutbound@sip.cognigy.ai"`. |
| Custom SIP Headers   | JSON          | Array of objects listing SIP headers that should be sent to the transferee. Each object comprises a name and a value attribute.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |


# User Input Timeout - Config
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/generic/user-input-timeout-config



<a href="/release-notes/4.45"><Badge>Updated in 4.45</Badge></a>

<Frame>
  <img alt="Generic User Input Timeout Config Node configuration panel" />
</Frame>

## Description

<Warning>
  * As of Cognigy 4.96.0, AudioCodes Nodes and Generic Voice Nodes are deprecated. Flows with these Nodes continue to work, and you can still edit and clone them as well as export them as a Package. However, you can't create new ones in Cognigy 4.96 or later.
  * The full removal of AudioCodes Nodes is planned for Q2 2026. After the removal, AI Agents using these Nodes will stop working properly. We strongly recommend migrating to [Voice Gateway Nodes](/ai/agents/develop/node-reference/voice/voice-gateway/overview) before the removal date.
</Warning>

The User Input Timeout - Config Node defines the action that the AI Agent should take if it does not receive any input from the user within a certain timeframe.

Once the settings are executed, they will remain in effect for the duration of the session.

| Parameter             | Type     | Description                                                                                                                            |
| --------------------- | -------- | -------------------------------------------------------------------------------------------------------------------------------------- |
| User No Input Mode    | Dropdown | Define the action if a user does not provide input to the AI Agent in time.                                                            |
| User No Input Timeout | Number   | Define the timeout for user input in milliseconds.                                                                                     |
| User No Input Retries | Number   | Define the number of retries for the AI Agent to request input from the user before ending the call. The maximum value is 999 retries. |


# Voice Nodes
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/overview



<Warning>
  * As of Cognigy 4.96.0, AudioCodes Nodes and Generic Voice Nodes are deprecated. Flows with these Nodes continue to work, and you can still edit and clone them as well as export them as a Package. However, you can't create new ones in Cognigy 4.96 or later.
  * The full removal of AudioCodes Nodes is planned for Q2 2026. After the removal, AI Agents using these Nodes will stop working properly. We strongly recommend migrating to [Voice Gateway Nodes](/ai/agents/develop/node-reference/voice/voice-gateway/overview) before the removal date.
</Warning>

Voice Nodes are necessary for developing a voice AI Agent.

<CardGroup>
  <Card title="Generic Voice Nodes" href="/ai/agents/develop/node-reference/voice/generic/overview">
    The Generic Voice Nodes create voice Flows that are not specific to one Endpoint. They are compatible with both AudioCodes and Voice Gateway.
  </Card>

  <Card title="Voice Gateway Nodes" href="/ai/agents/develop/node-reference/voice/voice-gateway/overview">
    Voice Gateway Nodes help you build a voice AI Agent using Voice Gateway configuration.
  </Card>

  <Card title="AudioCodes" href="/ai/agents/develop/node-reference/voice/audiocodes/overview">
    AudioCodes Nodes help you build a voice AI Agent using [AudioCodes configuration](https://www.cognigy.com/products/voice-gateway).
  </Card>
</CardGroup>


# Call Recording
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/voice-gateway/call-recording



<a href="/release-notes/4.45">
  <Badge>
    Updated in 4.45
  </Badge>
</a>

<Frame>
  <img alt="Call Recording Node" />
</Frame>

## Description

The Call Recording Node starts, stops, pauses, or resumes recording of the current call. Using the [SIPREC protocol (RFC 7866)](https://datatracker.ietf.org/doc/html/rfc7866), Voice Gateway streams a copy of the call audio in real time to a session recording server (SRS) that you operate. The SRS can be any server that supports the SIPREC standard.

<Note>
  The Call Recording Node uses SIPREC only. For WebSocket-based recording, use the [Call Recording](/voice-gateway/webapp/accounts#call-recording-configuration) setting in the [Voice Gateway Self-Service Portal](/voice-gateway/webapp/overview) instead.
</Note>

## Parameters

| Parameter             | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
| --------------------- | ------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Activity              | Select        | Whether to `start`, `stop`, `pause`, or `resume` the call recording.                                                                                                                                                                                                                                                                                                                                                                                            |
| Call Recording Server | CognigyScript | Enter the SIP URI of your SRS. Voice Gateway uses this URI as the destination for the `SIPREC INVITE` request that carries the recorded media. You can specify either a single host that resolves to one or more SRS instances via DNS/SRV records, or a comma-separated list of SIP URIs to send recordings to multiple recorders in parallel. For more information about supported SIP URI formats, refer to the [SIP URI Formats](#sip-uri-formats) section. |
| Call Recording ID     | CognigyScript | Defines the recording session identifier. Voice Gateway forwards the value of this parameter to the SRS, allowing you to correlate recordings. This parameter is optional.                                                                                                                                                                                                                                                                                      |

### SIP URI Formats

### Supported formats

<AccordionGroup>
  <Accordion title="Valid SIP URI formats">
    | Use case                         | Example value                               |
    | -------------------------------- | ------------------------------------------- |
    | Single SRS by hostname           | `sip:siprec.example.com`                    |
    | Single SRS by hostname with port | `sip:siprec.example.com:5060`               |
    | Single SRS by IP address         | `sip:3.208.57.175`                          |
    | Single SRS over TLS              | `sips:siprec.example.com`                   |
    | Multiple parallel recorders      | `sip:srs1.example.com,sip:srs2.example.com` |

    When multiple URIs are provided, Voice Gateway opens a parallel SIPREC dialog to each URI. The recording is considered successful if at least one dialog is accepted.
  </Accordion>

  <Accordion title="Invalid SIP URI formats">
    Voice Gateway accepts only SIP URIs. The following formats aren't supported:

    * WebSocket URLs such as `wss://…` or `ws://…`.
    * HTTP URLs such as `https://…` or `http://…`.
    * Plain hostnames or IP addresses without a SIP scheme. For example, `siprec.example.com` or `10.0.0.5`.
  </Accordion>
</AccordionGroup>

## Troubleshooting

| Symptom                                                                            | Likely cause                                                                                                           | Recommended Action                                                                             |
| ---------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------- |
| Recording never starts. SBC logs show `503` response to `startCallRecording` INFO. | `Call Recording Server` isn't a valid SIP URI, for example `wss://…`, `https://…`, or missing `sip:` / `sips:` scheme. | Use a valid `sip:` or `sips:` URI. See [Supported formats](#supported-formats).                |
| Logs show `Cannot start recording when recording is already started`.              | The Flow triggered `start` twice without a `stop` in between.                                                          | Ensure `start` is called only once per session, or send a `stop` before calling `start` again. |
| Logs show `discarding duplicate startCallRecording request for a call`.            | A duplicate `start` request was sent to the SBC for a call that is already being recorded.                             | Ensure the Flow issues only one `start` per recording session.                                 |
| Logs show `startCallRecording request is missing X-Srs-Url header`.                | The `Call Recording Server` field was left empty, or a CognigyScript expression evaluated to empty.                    | Set a static value or verify your expression resolves at runtime.                              |
| SRS receives the request but no audio arrives.                                     | Network is blocking audio between systems.                                                                             | Allow RTP traffic from SBC to SRS.                                                             |

## FAQ

<Accordion title="Does SIPREC count as an active session?">
  No. A recorded call counts as one active session. Additional SIP dialogs and RTP streams are handled by the SRS. Those streams don't count against Voice Gateway quotas, even when multiple SRS instances receive the audio.
</Accordion>


# DTMF
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/voice-gateway/dtmf



<Frame>
  <img alt="Voice Gateway DTMF Node configuration panel" />
</Frame>

## Description

The DTMF Node sends RTP payloads to the caller's SIP provider to play DTMF digit tones to the caller. If the caller's SIP provider supports RTP payloads according to [RFC 2833](https://datatracker.ietf.org/doc/html/rfc2833), the customer will hear the voice tones of the DTMF digits.

This Node is useful when paired with a [Handover to Human Agent](/ai/agents/develop/node-reference/service/handover-to-agent) Node or a [Transfer](/ai/agents/develop/node-reference/voice/voice-gateway/transfer) Node.
By playing the DTMF digit tones,
the DTMF Node informs the caller
that they are being transferred to another agent — either a voice AI Agent or a human agent in a contact center.

The Node can also be used in combination with DTMF input
(the **Capture DTMF Signals** parameter in the [Set Session Config](/ai/agents/develop/node-reference/voice/voice-gateway/set-session-config) Node) to provide feedback on the caller's input.
For example, after a caller enters digits, the system plays back the corresponding DTMF tones for each digit entered to confirm the input.

## Parameters

| Parameter       | Type   | Description                                                                                                                                                          |
| --------------- | ------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Digits Sequence | Text   | Specifies the sequence of digits to be played to the caller. The sequence of digits can be a telephone extension code or the number for the customer service center. |
| Pause time      | Slider | Specifies the pause length in milliseconds (ms), between each DTMF digit tone. The default value is 500 ms.                                                          |

## Example

A company provides an option
for customers to order equipment by leaving a voice message with the details about the equipment
or to get assistance with equipment they have already received.
An AI Agent guides the customer through this process using DTMF digit tones.

To build a Flow with the DTMF Node for the previous example, follow these steps:

<Accordion title="Example DTMF Flow">
  1. In the Flow editor, click **+** and select **Extensions > VG > Set Session Config**.

  2. In the **DTMF** section of the **Set Session Config** Node, activate the **Capture DTMF Signals** parameter, remove `#` from the **DTMF Submit Digit** field, and save.

  3. Below the Set Session Config Node, add a Question Node and configure the following settings:

     * **Question Type**: select **Number**.
     * **Output Type**: Select **Text**.
     * **Text**:`Hello! Welcome to our Support. We can help you with your equipment needs. If you want to place an order for equipment, press 1. If you have already received your equipment and need help with setup or are experiencing issues, press 2.`

  4. Below the Question Node, add a Lookup Node, select the **Text** type, and click **Save Node**.

  5. Configure the logic in the branch for the first Case Node:

     1. In the **Value** field, specify `1`.
     2. Below the Case Node, add a Say Node. In the **Text** field of the Say Node, specify the following sentence: `After the signal, please leave a voice message with the information about the equipment you wish to order, and we will call you back`.
     3. Below the Say Node, add a DTMF Node. Specify `1` in the **Digit Sequence** field to correspond the tone to the dial number that the caller presses.
     4. (*Optional*) Add a Call Recording Node to record the caller speech.

  6. Configure the logic in the branch for the second Case Node:

     1. In the **Value** field, specify `2`.
     2. Below the second Case Node, add a Say Node. In the **Text** field of the Say Node, specify the following sentence: `After the signal, you will be transferred to a contact center`.
     3. Below the second Say Node, add another DTMF Node. Specify `2` in the **Digit Sequence** field to correspond the tone to the dial number that the caller presses.
     4. Add a Handover to Human Agent Node to route the caller to the contact center.

     <Frame>
       <img alt="DTMF example Flow with Question, Lookup, and Handover Nodes" />
     </Frame>

  Test your Flow via the Interaction Panel. Use the [Voice Call](/ai/agents/test/interaction-panel/voice-call) setting.
</Accordion>

## More Information

* [DTMF Parameter](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details)
* [DTMF Verb](/voice-gateway/references/verbs/dtmf)
* [DTMF Collect Config](/ai/agents/develop/node-reference/voice/generic/dtmf-collect-config)


# Hang Up
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/voice-gateway/hangup



<a href="/release-notes/4.99"><Badge>Updated in 4.99</Badge></a>

<Frame>
  <img alt="Voice Gateway Hang Up Node configuration panel" />
</Frame>

## Description

The Hang Up Node instructs Voice Gateway to end the call.

| Parameter               | Type   | Description                                                                                                                                                                                                                                              |
| ----------------------- | ------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Reason for hang up      | Text   | The reason for hanging up, which is shown in logs                                                                                                                                                                                                        |
| Enable Immediate Hangup | Toggle | When this setting is enabled, the ongoing call will be terminated immediately, bypassing any active operations that are currently awaiting execution. A typical use case is to automatically end an outbound call when an answering machine is detected. |


# Mute Speech Input
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/voice-gateway/mute-speech-input



<a href="/release-notes/4.61"><Badge>Updated in 4.61</Badge></a>

<Frame>
  <img alt="Voice Gateway Mute Speech Input Node configuration panel" />
</Frame>

## Description

This Node controls when speech and Dual-Tone Multi-Frequency (DTMF) user inputs can be collected in your Flow.
It prevents speech and DTMF interruptions in specific cases, such as [xApp Flows](/xApps/overview), allowing for a smoother conversation experience.
By enabling or disabling parameters in this Node as needed,
you can maintain better control and ensure a seamless workflow in your application.

## Parameters

<Warning>
  Do not turn off a speech user input throughout the call to receive additional messages from the user,
  especially when the AI Agent needs to send multiple messages with time gaps in between.
</Warning>

| Parameter         | Type   | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |
| ----------------- | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| Mute Speech Input | Toggle | Disables speech input recognition for the current call. By default, the parameter is deactivated, allowing the Nodes below the Mute Speech Input Node to accept speech inputs.<br /><br />When this parameter is activated, the Nodes below the Mute Speech Input Node don't accept speech inputs or digits as the answer. It affects all speech inputs made through the following Nodes:<ul><li>Nodes with Set Activity Parameters</li><li>Say and Play Nodes with enabled barge-in can no longer be interrupted</li><li>Question and Optional Question Nodes</li></ul> |
| Mute DTMF Input   | Toggle | Disables DTMF input recognition for the current call. By default, this parameter is deactivated, allowing the Nodes below the Mute Input Node to accept DTMF inputs.<br /><br />When this parameter is activated, the Nodes below the Mute DTMF Input Node don't accept DTMF inputs. This parameter affects all DTMF inputs made through the following Nodes:<ul><li>Nodes with Set Activity Parameters</li><li>DTMF and Play Nodes with the enabled DTMF parameter can no longer be interrupted</li></ul>                                                               |

## Examples

<AccordionGroup>
  <Accordion title="Mute Speech Input Example">
    You can control when speech input is accepted or not in the Flow:

    1. Say Node: `Hello, you will receive an xApp`.
    2. xApp: Init Session Node.
    3. Mute Speech Input (enabled).
    4. Question Node: `Please rate your experience with the xApp on a scale of 1 to 5`.<br />When the Mute Node STT is enabled, the Question Node will not accept speech inputs or digits as the answer. The user must provide the rating through other means, such as selecting a number using buttons or typing the response. This process ensures that the speech gathering is temporarily turned off during the rating question to avoid unintended interference from background speech.
    5. Mute Node STT (disabled).

    After the rating question, the Mute Node STT is disabled, allowing the subsequent Nodes to accept speech inputs normally.
  </Accordion>

  <Accordion title="Mute DTMF Input Example">
    You can control when DTMF input is accepted or not in the Flow:

    1. Say Node: `Welcome to the interactive menu`.
    2. xApp: Init Session Node.
    3. Mute DTMF Input (enabled).
    4. Question Node: `Please select an option from the menu using your keypad`. <br />When the Mute DTMF Input Node is enabled, the Question Node will not accept DTMF inputs. Users must interact with the menu using other means, such as speech recognition or selecting options through buttons or touchscreen. This prevents unintended DTMF input interference while speech recognition or other input methods are active.
    5. Mute DTMF Input (disabled).

    After the menu interaction, the Mute DTMF Input Node is disabled, allowing subsequent Nodes to accept DTMF inputs normally.
  </Accordion>
</AccordionGroup>

## More Information

* [Generic Mute Speech Input Node](/ai/agents/develop/node-reference/voice/voice-gateway/mute-speech-input)
* [xApps](/ai/agents/develop/node-reference/xApp/overview)
* [xApp Nodes](/ai/agents/develop/node-reference/xApp/overview)


# Voice Gateway Nodes
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/voice-gateway/overview



Voice Gateway Nodes help you build a voice AI Agent using [Voice Gateway configuration](/voice-gateway/overview).

<CardGroup>
  <Card title="Call Recording" href="/ai/agents/develop/node-reference/voice/voice-gateway/call-recording">
    The Call Recording Node starts or stops a conversation recording.
  </Card>

  <Card title="DTMF" href="/ai/agents/develop/node-reference/voice/voice-gateway/dtmf">
    The DTMF Node plays DTMF digits to the caller.
  </Card>

  <Card title="Hang Up" href="/ai/agents/develop/node-reference/voice/voice-gateway/hangup">
    The Hang Up Node instructs Voice Gateway to end the call.
  </Card>

  <Card title="Mute Speech Input" href="/ai/agents/develop/node-reference/voice/voice-gateway/mute-speech-input">
    The Mute Speech Input Node controls when speech and DTMF user inputs are collected in your Flow.
  </Card>

  <Card title="Play" href="/ai/agents/develop/node-reference/voice/voice-gateway/play">
    The Play Node[^1] lets you stream a sound file into the call, also in a loop. The Node supports encoded MP3 and WAV files.
  </Card>

  <Card title="Send Metadata" href="/ai/agents/develop/node-reference/voice/voice-gateway/send-metadata">
    The Send Metadata Node sends metadata via SIP INFO messages to the connected SIP trunk.
  </Card>

  <Card title="Set Session Config" href="/ai/agents/develop/node-reference/voice/voice-gateway/set-session-config">
    The Set Session Config Node configures the Voice Gateway session. This configuration applies for the rest of the session unless overwritten, either directly or by activity parameters.
  </Card>

  <Card title="Transfer" href="/ai/agents/develop/node-reference/voice/voice-gateway/transfer">
    The Transfer Node lets you transfer an ongoing call to a Tel or SIP target.
  </Card>
</CardGroup>

[^1]: This Node can be configured with [Parameter Details](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details).


# Voice Gateway Parameter Details
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details



<a href="/release-notes/2026.11"><Badge>Updated in 2026.11</Badge></a>

Cognigy Voice Gateway has many configuration settings that are controlled directly from within your Flow. These settings can be applied individually to the scopes:

* **Setting Session Parameters**. Session parameters can comfortably be set with the [Set Session Config](/ai/agents/develop/node-reference/voice/voice-gateway/set-session-config) Node. When executed, the settings apply for the remainder of the session.
* **Setting Activity Parameters**. Activity parameters can be set for activities or Nodes. If, for example, Barge In is set on the [Play](/ai/agents/develop/node-reference/voice/voice-gateway/play) Node, Barge In is only activated during the execution of this Node. Therefore, the user can interrupt the AI Agent during the output but not afterward. These configurations are also available in the [Say](/ai/agents/develop/node-reference/basic/say), [Question](/ai/agents/develop/node-reference/basic/question), and [Optional Question](/ai/agents/develop/node-reference/basic/optional-question) Nodes.

## Parameters

<AccordionGroup>
  <Accordion title="Synthesizer (TTS)">
    The Text-To-Speech (TTS) settings can be chosen from a pre-filled dropdown for Microsoft Azure, AWS, Google, Nuance, or a custom vendor.

    | Parameter                  | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
    | -------------------------- | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | TTS Vendor                 | Dropdown      | Defines the TTS vendor. You can select a custom vendor.<br /><br />The **Default** option's behavior in the Set Session Config Node depends on whether you select this option in the first or in a subsequent Node of this type:<ul><li>**First Set Session Config Node** — switches the TTS vendor to the default configuration set in the [Voice Gateway Self-Service Portal](/voice-gateway/webapp/applications#add-additional-tts-and-stt-vendor)</li><li>**Subsequent Set Session Config Node** — keeps the TTS vendor set in the previous Set Session Config Node, allowing you to change other parameters without overwriting the TTS vendor.</li></ul>The Speechmatics TTS vendor is hidden behind a feature flag. Activate Speechmatics on the Cognigy and Voice Gateway sides:<ul><li>If you have a SaaS installation, contact [Cognigy technical support](/help/get-help).</li><li>If you have an on-premises installation, specify the [feature flags](/voice-gateway/webapp/speech-services#speechmatics) in `values.yaml`.</li></ul> |
    | Custom (Vendor)            | CognigyScript | Allows for specifying a TTS Vendor, which isn't in the dropdown list. This option is only available on Voice Gateway. For preinstalled providers, use all lowercase letters, for example, `microsoft`, `google`, `aws`. For custom providers, use the name that you specified on the [Speech Service](/voice-gateway/webapp/speech-services) page in Voice Gateway. <br /><br /> The **Custom** field appears if you selected **Custom** from the **TTS Vendor** list.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | TTS Language               | Dropdown      | Defines the language of the Voice AI Agent output.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 |
    | Custom (Language)          | CognigyScript | Allows for choosing a TTS language, which isn't in the dropdown list. Defines the language of the AI Agent output. The format to use depends on the TTS Vendor, for example, `de-DE`, `fr-FR`, `en-US`. <br /><br /> The **Custom** field appears if you selected **Custom** from the **TTS Language** list.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
    | TTS Voice                  | Dropdown      | Defines the voice that should be used for the voice AI Agent output.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
    | Custom (Voice)             | CognigyScript | Allows for choosing a TTS voice, which isn't in the dropdown list. This setting can be the case for region-specific voices. The format to use depends on the TTS Vendor, for example, `de-DE-ConradNeural`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
    | TTS Label                  | CognigyScript | The alternative name of the vendor is the one you [specify in the Voice Gateway Self-Service Portal](/voice-gateway/webapp/applications#add-additional-tts-and-stt-vendor). If you have created multiple speech services from the same vendor, use the label to specify which service to use.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
    | Enable Advanced TTS Config | Toggle        | Enables the addition of a URL for an Azure Custom Voice Endpoint.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
    | Disable TTS Audio Caching  | Toggle        | Disables TTS audio caching.<br /><br /> By default, the setting is deactivated. In this case, previously requested TTS audio results are stored in the AI Agent cache. When a new TTS request is made, and the audio text has been previously requested, the AI Agent retrieves the cached result instead of sending another request to the TTS provider.<br /><br /> When the setting is activated, the AI Agent no longer caches TTS results. In this case, each request is directly sent to your speech provider.  <br /><br /> Note that disabling caching can increase TTS costs. For detailed information, contact your speech provider.                                                                                                                                                                                                                                                                                                                                                                                                     |
  </Accordion>

  <Accordion title="Recognizer (STT)">
    The Speech-To-Text (STT) settings can be chosen from a pre-filled dropdown for supported vendors, or you can select a custom vendor.

    | Parameter                           | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
    | ----------------------------------- | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | STT Vendor                          | Dropdown      | Defines the STT vendor. You can select a custom vendor.<br /><br />The **Default** option's behavior in the Set Session Config Node depends on whether you select this option in the first or in a subsequent Node of this type:<ul><li>**First Set Session Config Node** — switches the STT vendor to the default configuration set in [Voice Gateway Self-Service Portal](/voice-gateway/webapp/applications#add-additional-tts-and-stt-vendor)</li><li>**Subsequent Set Session Config Node** — keeps the TTS vendor set in the previous Set Session Config Node, allowing you to change other parameters without overwriting the STT vendor.</li></ul> The ElevenLabs STT vendor is hidden behind a feature flag. Activate ElevenLabs on the Cognigy side: <ul><li>If you have a SaaS installation, contact [Cognigy technical support](/help/get-help). </li><li>If you have an on-premises installation, specify the `FEATURE_ENABLE_ELEVENLABS_STT_VENDOR=true` in the `values.yaml`.</li></ul>The Speechmatics STT vendor is hidden behind a feature flag. Activate Speechmatics on the Cognigy and Voice Gateway sides:<ul><li>If you have a SaaS installation, contact [Cognigy technical support](/help/get-help).</li><li>If you have an on-premises installation, specify the [feature flags](/voice-gateway/webapp/speech-services#speechmatics) in `values.yaml`.</li></ul> |
    | STT Language                        | Dropdown      | Select the desired STT Language. For custom languages, use the following format: de-DE, fr-FR, en-US.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
    | Custom (Vendor)                     | CognigyScript | Allows for specifying an STT Vendor, which isn't in the dropdown list. Select the desired STT Vendor. This option is only available on Voice Gateway. For preinstalled providers, use all lowercase letters like `microsoft`, `google`, `aws`. For custom providers, use the name that you specified on the [Speech Service page](/voice-gateway/webapp/speech-services) in Voice Gateway. <br /><br /> The **Custom** field appears if you selected **Custom** from the **STT Vendor** list.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |
    | STT Language                        | Dropdown      | Defines the language that should be recognized.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | Custom (Language)                   | CognigyScript | Allows for choosing an STT language, which isn't in the dropdown list. This setting can be the case for region-specific voices. The format to use depends on the STT Vendor, for example, `de-DE`, `fr-FR`, `en-US`. <br /><br /> The **Custom** field appears if you selected **Custom** from the **STT Language** list.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
    | Deepgram Model                      | Dropdown      | This parameter is active only if Deepgram is selected from the **STT Vendor** list. <br /><br /> Choose a model for processing submitted audio. Each model is associated with a tier. Ensure that the selected tier is available for the chosen STT language. For detailed information about Deepgram models, refer to the [Deepgram documentation](https://developers.deepgram.com/docs/model).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | Endpointing                         | Toggle        | This parameter is active only if Deepgram or Deepgram Flux is selected from the **STT Vendor** list: <ul><li>**Deepgram** – detects the end of speech by monitoring user pauses. When a pause is detected, Deepgram finalizes the transcript with `speech_final = true`. The duration for detecting the end of speech is preconfigured with the default value of 10 milliseconds. If you want to change this value, use the **Endpointing Time** parameter.</li><li>**Deepgram Flux** – detects the end of speech using a confidence threshold. When a user's speech ends or the optional timeout is reached, the transcript is finalized. This approach ensures accurate, responsive multiturn transcription.</li></ul> For detailed information about the endpointing configuration, refer to the documentation for [Deepgram](https://developers.deepgram.com/docs/endpointing) or [Deepgram Flux](https://developers.deepgram.com/docs/flux/configuration).                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | Endpointing Time                    | Number        | This parameter is active only when Deepgram is selected from the **STT Vendor** list and the **Endpointing** toggle is enabled. <br /><br /> Customize the duration (in milliseconds) for detecting the end of speech. The default is 10 milliseconds of silence. After detecting silence, the system waits until either the speaker resumes or the required silence duration is reached. Once one of these conditions is met, the transcript is sent with `speech_final` set to `true`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
    | Smart Formatting                    | Toggle        | This parameter is active only if Deepgram is selected from the **STT Vendor** list. <br /><br /> Deepgram's Smart Format feature applies additional formatting to transcripts to optimize them for human readability. Smart Format capabilities vary between models. When Smart Formatting is turned on, Deepgram always applies the best-available formatting for your chosen model, tier, and language combination. For detailed examples, refer to the [Deepgram documentation](https://developers.deepgram.com/docs/smart-format). <br /><br /> Note that when Smart Formatting is turned on, punctuation is activated, even if you have the Disable STT Punctuation setting enabled.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
    | End of Turn Threshold               | Slider        | This parameter is active only if Deepgram Flux is selected from the **STT Vendor** list. Defines the confidence level for detecting when a user's speech ends. Higher values improve accuracy but may increase latency. The default value is `0.7`, meaning the system waits until it is 70% confident that the user has finished speaking before finalizing the transcript.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
    | End of Turn Timeout in Milliseconds | Number        | This parameter is active only if Deepgram Flux is selected from the **STT Vendor** list. Defines the maximum number of milliseconds to wait before finalizing a turn. The timeout acts as a backup if the user stops speaking unexpectedly. The default value is `5000` (5 seconds), meaning the turn will be closed after 5 seconds of silence.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | STT Hints                           | Text          | Array of words or phrases to assist speech detection. If you want to use multiple hints, enter each hint into a separate input field. For instance, you can enter `Germany` in the first field, `France` in the second field, and `Netherlands` in the third field. The STT provider receives the data in array format: \["Germany", "France", "Netherlands"]. <br /> Note: This requires support from the STT engine. The field isn't available for the Nuance speech vendor.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | Dynamic Hints                       | CognigyScript | Uses context or input for adding array hints. For example, <code>\{\{context.hints}}</code> or <code>\{\{input.hints}}</code>. You can override these settings using Advanced parameters.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
    | STT Label                           | CognigyScript | The alternative name of the vendor is the one you [specify in the Voice Gateway Self-Service Portal](/voice-gateway/webapp/applications#add-additional-tts-and-stt-vendor). If you have created multiple speech services from the same vendor, use the label to specify which service to use.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |
    | Google Model                        | Dropdown      | This parameter is active only when Google is selected from the **STT Vendor** list.<br /><br />Uses one of Google Cloud Speech-to-Text transcription models. The default value is `latest_short`. For a detailed list of Google models, refer to the [Transcription models](https://cloud.google.com/speech-to-text/docs/transcription-model#transcription_models) section in the Google Documentation. Keep in mind that the `default` value is a Google model type that can be used if other models don't suit your use case.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | Enable Voice Activity Detection     | Toggle        | Delays connection to cloud recognizer until the speech is detected.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
    | VAD Sensitivity                     | Slider        | Detection sensitivity, the lowest value has the highest sensitivity.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
    | Minimal Voice Duration              | Slider        | Milliseconds of speech activity required before connecting to the cloud recognizer.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
    | Disable STT Punctuation             | Toggle        | This parameter appears only when Google, Deepgram, or Deepgram Flux is selected from the **STT Vendor** list.<br /><br /> Prevents the transcription returned to the AI Agent from including punctuation marks.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | Enable Advanced TTS Config          | Toggle        | Enables the addition of an ID for an Azure's Custom Speech model deployment.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
    | Profanity Filter                    | Dropdown      | Defines how offensive, inappropriate, or abusive language is handled in transcripts when using Microsoft Azure Speech Services as the STT vendor. Possible values:<ul><li>**Raw** - profanity shows up in transcripts. This option is selected by default.</li><li>**Masked** - replaces profanity with asterisks (`****`) in transcripts.</li><li>**Removed** - removes profanity from transcripts.</li></ul>For more information, see the [Microsoft Azure Speech Services documentation on the profanity filter](https://learn.microsoft.com/en-us/azure/ai-services/speech-service/display-text-format?pivots=programming-language-csharp#profanity-filter).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | Enable Audio Logging                | Toggle        | Enables recording and logging of audio from the user on Azure.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | Recognize Language                  | Toggle        | Enables the addition of alternative languages for recognition. You can select a maximum of 3 languages. To reuse these languages in other Nodes, such as the child Nodes of the [Lookup](/ai/agents/develop/node-reference/logic/lookup) Node,  use the following format: `de-DE`, `fr-FR`, `en-US`. For the parent Node of the Lookup Node, specify `input.data.payload.speech.language_code`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
  </Accordion>

  <Accordion title="Barge In">
    <Warning>
      Barge In uses the [TTS and STT vendor](/voice-gateway/references/tts-and-stt-vendors) to listen throughout the entire conversation. Consequently, Barge In may lead to increased subscription costs with your vendor.
    </Warning>

    Barge In is a feature that allows the caller to interrupt the voice AI Agent by using speech input or DTMF digits during the entire call. By default, this feature is turned off.

    Before release 4.80, this feature couldn't be controlled when the call was transferred to the contact center. Barge In was always active, allowing the caller to interrupt the voice AI Agent at any time.

    Starting with release 4.80, you can enable or disable Barge In when the call is redirected to the contact center. This improvement lets you decide whether the caller should listen to the voice AI Agent's messages fully or have the option to interrupt them.
    This way, the caller can't use Barge In to skip, for example, important legal information such as the GDPR.

    To ensure Barge In works correctly after the call is transferred to the contact center, place the Set Session Config Node above the Handover to Human Agent Node.

    | Parameter              | Type   | Description                                                                                                                                                                                                                                                                                                                                      |
    | ---------------------- | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Barge In On Speech     | Toggle | Enables interrupting the voice AI Agent with speech. The user is able to interrupt the voice AI Agent's responses even after the handover has taken place and a human agent communicates with the user through TTS. This parameter is disabled by default. It retains its setting throughout the whole conversation.                             |
    | Barge In On DTMF       | Toggle | Enables interrupting the voice AI Agent with DTMF digits. The user is able to interrupt the voice AI Agent's responses by pressing any digit, even after the handover has taken place and a human agent communicates with the user through TTS. This parameter is disabled by default. It retains its setting throughout the whole conversation. |
    | Barge In Minimum Words | Slider | Defines the minimum number of words that the user must say for the Voice Gateway to consider it a barge in.                                                                                                                                                                                                                                      |
  </Accordion>

  <Accordion title="User Input Timeout">
    This feature defines what should happen when there is no input from the user.

    Before the release [4.81](/release-notes/4.81), User Input Timeout was always enabled and users had to determine the number of milliseconds before timeout occurred. Starting from release [4.81](/release-notes/4.81), users can enable or disable User Input Timeout using a toggle. This setting keeps the voice AI Agent on the call even if the caller takes a while to respond. When the User Input Timeout is disabled, the voice AI Agent waits for the caller's response.

    | Parameter                    | Type     | Description                                                                                                                                                                                                                  |
    | ---------------------------- | -------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Enable User No Input Timeout | Toggle   | Enables or disables the User No Input Timeout parameter. This parameter is enabled by default.                                                                                                                               |
    | User No Input Mode           | Dropdown | This parameter is active only when Enable User No Input Timeout is enabled. <br /><br /> Defines the action if a user doesn't provide an input to the AI Agent in time.                                                      |
    | User No Input Timeout        | Number   | This parameter is active only when Enable User No Input Timeout is enabled. <br /><br /> Defines the timeout duration for user input, specified in milliseconds (ms).                                                        |
    | User No Input Retries        | Number   | This parameter is active only when Enable User No Input Timeout is enabled. <br /><br /> Defines how often the voice AI Agent should retry to get an input from a user before completing the call. The default value is `5`. |
  </Accordion>

  <Accordion title="Flow Input Timeout">
    This feature is designed for use cases where response delays occur in a Flow, such as when utilizing large language models (LLMs), waiting for responses from external services, or managing complex processing tasks. It helps maintain user engagement by proactively delivering information or prompts during these delays.

    For example, you can inform end users that their request is being processed or that assistance is on the way.

    | Parameter                    | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | ---------------------------- | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Enable Flow No Input Timeout | Toggle        | This parameter is disabled by default. When enabled, this setting plays a track or speaks a prompt while waiting for a response from the Flow.                                                                                                                                                                                                                                                                                                                          |
    | Flow No Input Mode           | Dropdown      | This parameter is active only when **Enable Flow No Input Timeout** is enabled. Select one of the following modes:<ul><li>**Speak** - outputs a spoken message, allowing you to communicate important information or prompts to the user.</li><li>**Play** - plays an audio track, such as music, sound effects, or pre-recorded messages.</li></ul>                                                                                                                    |
    | Flow No Input URL            | URL           | This parameter is active only when the **Play** option in **Enable Flow No Input Timeout** is selected. <br /><br /> Define the URL that is played if the Flow doesn't continue in time. MP3 and WAV files are supported.                                                                                                                                                                                                                                               |
    | Flow No Input Speech         | CognigyScript | This parameter is active only when the **Speak** option in **Enable Flow No Input Timeout** is selected. Define a prompt to be spoken if the Flow doesn't continue in time.                                                                                                                                                                                                                                                                                             |
    | Flow No Input Timeout        | Number        | This parameter is active only when **Enable Flow No Input Timeout** is enabled. <br /><br /> Defines how frequently the voice AI Agent should retry to obtain input from the Flow before completing the call.                                                                                                                                                                                                                                                           |
    | Flow No Input Retries        | Number        | This parameter is active only when **Enable Flow No Input Timeout** is enabled. Define how many times the AI Agent should retry to obtain input from the Flow before executing an action.                                                                                                                                                                                                                                                                               |
    | AI Agent Fails on Error      | Toggle        | This parameter is disabled by default. It defines a failure condition when the maximum number of retry attempts is reached. If [Call Failover](/ai/agents/deploy/endpoint-reference/voice-gateway#call-failover) is enabled in the Voice Gateway Endpoint Settings, the call is transferred either to another voice AI Agent or to a human agent in the contact center. If Call Failover isn't enabled, the call disconnects, leaving the user without further support. |
  </Accordion>

  <Accordion title="DTMF">
    Enables DTMF collection.

    | Parameter                | Type          | Description                                                                                                                                     |
    | ------------------------ | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------- |
    | Capture DTMF signals     | Toggle        | Enables capturing DTMF signals by the AI Agent.                                                                                                 |
    | DTMF Inter Digit Timeout | Number        | Defines the timeout between collected DTMF digits.                                                                                              |
    | DTMF Max Digits          | Number        | Defines the maximum number of digits the user can enter. The digits are submitted automatically once this limit is reached.                     |
    | DTMF Min Digits          | Number        | Defines the minimum number of digits before they are forwarded to the AI Agent. A submit digit can override this.                               |
    | DTMF Submit Digit        | CognigyScript | Defines the DTMF submit digit, which is used for submitting the previously entered digits. This action overrides the minimum digits validation. |
  </Accordion>

  <Accordion title="Continuous ASR">
    Continuous ASR enables the Voice Gateway to concatenate multiple STT recognitions of the user and then send them as a single textual message to the AI Agent.

    | Parameter                   | Type          | Description                                                                                                                                    |
    | --------------------------- | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------- |
    | Enable Continuous ASR       | Toggle        | Enable or disable Continuous ASR.                                                                                                              |
    | Continuous ASR Submit Digit | CognigyScript | Defines a special DTMF key that sends the accumulated recognitions to the Flow.                                                                |
    | Continuous ASR Timeout      | Number        | Defines the number of milliseconds of silence before the accumulated recognitions are sent to the Flow. The default and minimum value is 2000. |
  </Accordion>

  <Accordion title="Atmosphere Sounds">
    This feature is useful in scenarios where users interact with an AI Agent instead of a human when calling the contact center. Within the Atmosphere Sound section, you can configure the MP3 background track. This track may include office noises or other sounds that simulate human interaction, helping the caller feel they are speaking with a person rather than an AI Agent.
    Playing a background MP3 track during the conversation with AI Agents makes it more engaging and personalized.

    The track plays during the conversation with the AI Agent, continues when the call is transferred to a human agent, and stops once the human agent accepts the call.

    | Parameter | Type     | Description                                                                                                                                                                                                                                                                                               |
    | --------- | -------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Action    | Dropdown | Selects an action to play, silence, or remove the track: <ul><li>**play** — plays the track in the background.</li><li>**silence** — mutes the track.</li><li>**remove** — removes the track from the background completely.</li></ul>                                                                    |
    | URL       | Text     | Accepts URL links to MP3 tracks. The URL doesn't need to include the `.mp3` extension. For example, `https://abc.xyz/music.mp3` or `https://audio.jukehost.co.uk/N5pnlULbup8KabGRE7dsGwHTeIZAwWdr`.                                                                                                       |
    | Loop      | Toggle   | Turns on looping for the audio track.                                                                                                                                                                                                                                                                     |
    | Volume    | Number   | Adjusts the volume of the track. Can be set from -50 to +50 dB. The default value is 0, meaning that the track is played as-is, with no adjustments to its volume. Users may need to adjust the volume by testing the call and checking if the Atmosphere Sounds track is neither too loud nor too quiet. |
  </Accordion>

  <Accordion title="Silence Overlay">
    Silence Overlay enables you to play an MP3 file in the background during calls with an AI Agent. This feature is activated during prolonged periods of silence, which may result from the AI Agent's background activity. Playing the track informs the customer that the AI Agent is processing their query, which may take some time. The Silence Overlay track can simulate office sounds, for example, a human agent typing on a keyboard and clicking the mouse.

    When Silence Overlay is enabled in the [Set Session Config](./set-session-config) Node, the Silence Overlay track starts playing automatically once the AI Agent takes longer to respond, then stops the moment the AI Agent responds. You can adjust the delay before the Silence Overlay starts to make it sound more natural.

    If you enabled the [Call Recording](/voice-gateway/webapp/recent-calls#call-recordings) feature in the Voice Gateway Self-Service Portal, the Silence Overlay track is recorded together with the AI Agent's track and can be played back in the audio file.

    | Parameter                              | Type     | Description                                                                                                                                                                                                                                                                                                                                                                                             |
    | -------------------------------------- | -------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Action                                 | Dropdown | Defines an action to play or remove the track: <ul><li>**play** — plays the track in the background when prolonged silence occurs.</li><li>**remove** — removes the track from the conversation. Next time a prolonged silence occurs, the Silence Overlay doesn't play.</li></ul><br />Make sure to place the next Set Session Config Node before the Node that needs to have Silence Overlay removed. |
    | URL                                    | Text     | Accepts URL links to an MP3 track. The URL doesn't need to include the `.mp3` extension. For example, `https://abc.xyz/music.mp3` or `https://audio.jukehost.co.uk/N5pnlULbup8KabGRE7dsGwHTeIZAwWdr`. This parameter appears when the play action is selected.                                                                                                                                          |
    | Delay for starting the Silence Overlay | Number   | Defines the wait time before the MP3 track plays, simulating a humanlike response. For example, human agents often have a pause between speaking and typing. This parameter appears when the play action is selected.                                                                                                                                                                                   |
  </Accordion>

  <Accordion title="Advanced">
    | Parameter                     | Type | Description                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | ----------------------------- | ---- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Additional Session Parameters | JSON | Allows for configuring settings using JSON. If you have already made changes using the UI settings above, this field overwrites them. Also, you can specify additional parameters in the JSON, which are unavailable in the UI, such as vendor credentials. <br /><br />If you want to specify a custom TTS or STT provider in the **vendor** parameter, use the `custom:<provider-name>` format, for example, `"vendor": "custom:My Speech provider"`. |

    **List of Additional Session Parameters**

    | Feature                | Parameter               | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   | Type             | Example                                    |
    | ---------------------- | ----------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------- | ------------------------------------------ |
    | **User Input Timeout** | `user`                  | The User Input Timeout feature.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               | Object           | -                                          |
    |                        | `noInputMode`           | The mode when no input is detected. Possible values: `"event"`, `"speech"`, `"play"`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         | String           | `"speech"`                                 |
    |                        | `noInputTimeout`        | The timeout duration for no input (in milliseconds).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          | Number           | `5000` (5 seconds)                         |
    |                        | `noInputRetries`        | The number of retries allowed when no input is detected.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      | Number           | `3`                                        |
    |                        | `noInputSpeech`         | The speech to play when no input is detected.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 | String           | `"Please speak now."`                      |
    |                        | `noInputUrl`            | The URL to fetch audio when no input is detected.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             | String           | `"https://example.com/no-input-audio.mp3"` |
    | **Flow Input Timeout** | `flow`                  | The Flow Input Timeout feature.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               | Object           | `{}`                                       |
    |                        | `flowNoInputEnable`     | Enables no-input handling for Flows.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          | Boolean          | `true`                                     |
    |                        | `flowNoInputMode`       | The mode when no input is detected in a Flow. Possible values: `"speech"`, `"play"`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          | String           | `"play"`                                   |
    |                        | `flowNoInputTimeout`    | The timeout duration for no input in a Flow (in milliseconds).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                | Number           | `3000` (3 seconds)                         |
    |                        | `flowNoInputRetries`    | The number of retries for no input in a Flow.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 | Number           | `2`                                        |
    |                        | `flowNoInputSpeech`     | The speech to play for no input in a Flow.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    | String           | `"Please try again."`                      |
    |                        | `flowNoInputUrl`        | The URL to fetch audio for no input in a Flow.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                | String           | `"https://example.com/flow-no-input.mp3"`  |
    |                        | `flowNoInputFail`       | Indicates whether to fail the Flow on no input.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               | Boolean          | `true`                                     |
    | **Recognizer**         | `recognizer`            | The Recognizer (STT) feature.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 | Object           | -                                          |
    |                        | `vendor`                | The vendor for speech recognition. Possible values: `"aws"`, `"deepgram"`, `"google"`, `"microsoft"`, `"nuance"`, `"default"`, `"custom"`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    | String           | `"google"`                                 |
    |                        | `language`              | The language for speech recognition.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          | String           | `"de-DE"`                                  |
    |                        | `hints`                 | A list of hints to improve recognition accuracy.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              | Array of strings | `["help", "skip", "confirm"]`              |
    |                        | `hintsBoost`            | A value to boost the weight of hints in recognition.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          | Number           | `20`                                       |
    |                        | `profanityOption`       | The option for handling offensive, inappropriate, or abusive language in transcripts when using Microsoft Azure Speech Services as the STT vendor. Possible values:<ul><li>`"raw"` - profanity shows up in transcripts. This option is selected by default.</li><li>`"masked"` - replaces profanity with asterisks (`****`) in transcripts.</li><li>`"removed"` - removes profanity from transcripts.</li></ul>For more information, see the [Microsoft Azure Speech Services documentation on profanity filter](https://learn.microsoft.com/en-us/azure/ai-services/speech-service/display-text-format?pivots=programming-language-csharp#profanity-filter). | String           | `"masked"`                                 |
    | **Synthesizer**        | `synthesizer`           | The Synthesizer (TTS) feature.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                | Object           | -                                          |
    |                        | `vendor`                | The vendor for TTS synthesis. Possible values: `"aws"`, `"deepgram"`, `"elevenlabs"`, `"google"`, `"microsoft"`, `"nuance"`, `"default"`, `"custom"`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         | String           | `"microsoft"`                              |
    |                        | `language`              | The language for TTS synthesis.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               | String           | `"de-DE"`                                  |
    |                        | `voice`                 | The voice used for TTS synthesis.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             | String           | `"en-US-JennyNeural"`                      |
    | **DTMF**               | `dtmf`                  | The DTMF feature.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             | Object           | -                                          |
    |                        | `dtmfEnable`            | Enables DTMF (Dual-Tone Multi-Frequency) input.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               | Boolean          | `true`                                     |
    |                        | `dtmfInterDigitTimeout` | The timeout between digits during DTMF input (in milliseconds).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               | Number           | `1000` (1 second)                          |
    |                        | `dtmfMaxDigits`         | The maximum number of digits allowed for DTMF input.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          | Number           | `10`                                       |
    |                        | `dtmfMinDigits`         | The minimum number of digits required for DTMF input.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         | Number           | `3`                                        |
    |                        | `dtmfSubmitDigit`       | The digit used to submit the DTMF input.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      | String           | `"9"`                                      |
    | **Barge In**           | `bargeIn`               | The Barge In feature.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         | Object           | -                                          |
    |                        | `bargeInEnable`         | Enables Barge In functionality.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               | Boolean          | `true`                                     |
    |                        | `bargeInOnDtmf`         | Allows Barge In on DTMF input.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                | Boolean          | `true`                                     |
    |                        | `bargeInOnSpeech`       | Allows Barge In on speech input.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              | Boolean          | `false`                                    |
    |                        | `bargeInMinimumWords`   | The minimum words required to trigger Barge In.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               | Number           | `5`                                        |
    |                        | `bargeInSticky`         | Keeps Barge In active after initial input.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    | Boolean          | `true`                                     |
    | **Continuous ASR**     | `continuousAsr`         | The Continuous ASR feature.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   | Object           | -                                          |
    |                        | `asrEnabled`            | Enables Continuous Automatic Speech Recognition (ASR).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        | Boolean          | `true`                                     |
    |                        | `asrDigit`              | The specific digit for triggering ASR.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        | String           | `"5"`                                      |
    |                        | `asrTimeout`            | The timeout for ASR detection (in milliseconds).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              | Number           | `5000` (5 seconds)                         |

    JSON example:

    ```json theme={null}
    {
      "synthesizer": {
        "vendor": "microsoft",
        "language": "de-DE",
        "voice": "en-US-JennyNeural"
      },
      "recognizer": {
        "vendor": "google",
        "language": "de-DE",
        "hints": [
          "help",
          "skip",
          "confirm"
        ],
        "hintsBoost": 20
      },
      "user": {
        "noInputMode": "speech",
        "noInputTimeout": "{{context.user.noInputTimeout}}",
        "noInputRetries": "{{context.user.noInputRetries}}",
        "noInputSpeech": "{{context.user.noInputSpeech}}"
      },
      "continuousAsr": {
        "asrEnabled": true,
        "asrTimeout": "{{context.continuousAsr.asrTimeout}}"
      },
      "dtmf": {
        "dtmfEnable": true,
        "dtmfInterDigitTimeout": "{{context.dtmf.dtmfInterDigitTimeout}}"
      },
      "flow": {
        "flowNoInputEnable": true,
        "flowNoInputMode": "speech",
        "flowNoInputTimeout": 15,
        "flowNoInputRetries": 2,
        "flowNoInputSpeech": "I'm sorry, I didn't hear anything"
      }
    }
    ```
  </Accordion>
</AccordionGroup>


# Play
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/voice-gateway/play



<a href="/release-notes/4.64"><Badge>Updated in 4.64</Badge></a>

<Frame>
  <img alt="Voice Gateway Play Node configuration panel" />
</Frame>

## Description

The Play Node lets you stream a sound file into the call with the ability to loop it.

This Node is useful for playing advertisements, pre-recorded jingles, or other sound files for callers. It supports MP3 and WAV encoded files.

## Parameters

| Parameter               | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
| ----------------------- | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Media URL               | CognigyScript | Accepts URL links to MP3 and WAV tracks. The URL should point to the storage location where you want to host it and doesn't need to include the `.mp3` or `.wav` extension. For example, `https://abc.xyz/music.mp3` or `https://audio.jukehost.co.uk/N5pnlULbup8KabGRE7dsGwHTeIZAwWdr`. Make sure the file is publicly accessible.                                                                                                                                                                                        |
| Loop count              | Number        | The number of times the track will be looped.                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |
| Set Activity Parameters | Toggle        | Enables parameters which will be valid for this activity only. The available parameters are: <ul><li>[Recognizer (STT)](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details)</li><li>[Barge In](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details)</li><li>[DTMF](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details)</li><li>[Advanced (Additional Session Parameters)](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details)</li></ul> |

## More Information

* [Parameter Details](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details)
* [AudioCodes Play URL Node](/ai/agents/develop/node-reference/voice/audiocodes/play-url)
* [Generic Play URL Node](/ai/agents/develop/node-reference/voice/generic/play-url)


# Send Metadata
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/voice-gateway/send-metadata



<Frame>
  <img alt="Voice Gateway Send Metadata Node configuration panel" />
</Frame>

## Description

The Send Metadata Node sends metadata via [SIP INFO messages](https://datatracker.ietf.org/doc/html/rfc2976) to the connected SIP trunk.

## Parameters

| Parameter | Type | Description                                                                                                                        |
| --------- | ---- | ---------------------------------------------------------------------------------------------------------------------------------- |
| Metadata  | JSON | Sets the metadata to send via SIP INFO. For dynamic values, use [CognigyScript](/ai/platform-features/cognigyscript#json-editors). |

### Example

```json theme={null}
{
  "callType": "refund_call",
  "order": {
    "orderId": "{{context.orderId}}",
    "product": "{{context.product}}"
  }
}
```


# Set Session Config
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/voice-gateway/set-session-config



<Frame>
  <img alt="Voice Gateway Set Session Config Node configuration panel" />
</Frame>

## Description

The Set Session Config Node configures the Voice Gateway session. All parameters set here are valid for the remainder of the session unless overwritten, either directly or by activity parameters.

You can find more information about activity parameters on our [Parameter Details](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details) page.


# Transfer
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/voice/voice-gateway/transfer



<a href="/release-notes/2026.9"><Badge>Updated in 2026.9</Badge></a>

<Frame>
  <img alt="Voice Gateway Transfer Node configuration panel" />
</Frame>

## Description

The Transfer Node allows you to transfer an ongoing call to a Tel or SIP target.

## Parameters

| Parameter                          | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          | Transfer Type |
| ---------------------------------- | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------- |
| Transfer Type                      | Dropdown      | There are two transfer types: <ul><li>**Refer** — forwarding an existing call.</li><li>**Dial** — creating a new outgoing call. If you want to use this type and still have the old Node version, add a new Voice Gateway Transfer Node in the Flow Editor and manually transfer the required settings from the old Node.</li></ul>                                                                                                                                                                                  | -             |
| Reason                             | CognigyScript | The reason for the handover. It is shown in Voice Gateway logs.                                                                                                                                                                                                                                                                                                                                                                                                                                                      | All           |
| Target                             | CognigyScript | E.164 syntax or a SIP URI are supported.                                                                                                                                                                                                                                                                                                                                                                                                                                                                             | All           |
| Caller ID                          | Number        | The caller ID. Some carriers, like Twilio, require a registered number for outgoing calls.                                                                                                                                                                                                                                                                                                                                                                                                                           | Dial          |
| Dial Music                         | URL           | Custom audio or ring-back which plays to the caller while the outbound call is ringing. Only `.wav` or `.mp3` files are supported. The URL doesn't need to include the `.mp3` or `.wav` extension. For example, `https://abc.xyz/music.mp3` or `https://audio.jukehost.co.uk/N5pnlULbup8KabGRE7dsGwHTeIZAwWdr`.                                                                                                                                                                                                      | Dial          |
| Enable Duration Limit for Transfer | Toggle        | Automatically disconnects transferred calls after a set time to prevent excessive call durations. This setting is useful in the following cases:<ul><li>Limit call duration by automatically ending transferred calls that exceed the predefined time, helping manage system performance.</li><li>Free up resources by disconnecting calls if the caller forgets to hang up, making lines available for other callers.</li><li>Reduce costs by preventing unnecessary charges from long, unattended calls.</li></ul> | Dial          |
| Duration Limit                     | Number        | This parameter is active only when **Enable Duration Limit for Transfer** is selected. Set the maximum duration in seconds. Transferred calls end automatically after this time, even if the caller is on the line. The default value is `600`.                                                                                                                                                                                                                                                                      | Dial          |
| Timeout                            | Number        | The amount of time (in seconds) that the AI Agent will ring before a [no-answer](/voice-gateway/references/events/NO_ANSWER) timeout. The default value is 60 seconds.                                                                                                                                                                                                                                                                                                                                               | Dial          |
| Enable Copilot                     | Toggle        | Creates the UUIValue which will be sent to the Contact Center through SIP Headers. Will collect information from the Transcription Webhook field as well as from the VG Endpoint Agent Copilot Config field to create the UUIValue. This setting requires a configured [Voice Copilot](/ai/agents/deploy/endpoint-reference/voice-copilot) Endpoint.                                                                                                                                                                 | Dial          |
| Copilot Headers Key                | CognigyScript | Define the Headers Key for Agent Copilot. The key is provider-dependent. For example, for Genesys, the key is `User-to-User`. The generated value becomes accessible after engaging the Transfer Node, located in the input object as `{{input.UUIValue}}`. Note that you do not need to handle this value manually. It is automatically generated and added to the header.                                                                                                                                          | Dial          |

<AccordionGroup>
  <Accordion title="Transcribe">
    | Parameter                           | Type          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                | Transfer Type |
    | ----------------------------------- | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------- |
    | STT Vendor                          | Dropdown      | Defines the STT vendor. You can select a custom vendor.<br /><br />The **Default** option's behavior in the Set Session Config Node depends on whether you select this option in the first or in a subsequent Node of this type:<ul><li>**First Set Session Config Node** — switches the STT vendor to the default configuration set in [Voice Gateway Self-Service Portal](/voice-gateway/webapp/applications#add-additional-tts-and-stt-vendor)</li><li>**Subsequent Set Session Config Node** — keeps the TTS vendor set in the previous Set Session Config Node, allowing you to change other parameters without overwriting the STT vendor.</li></ul> The ElevenLabs STT vendor is hidden behind a feature flag. Activate ElevenLabs on the Cognigy side: <ul><li>If you have a SaaS installation, contact [Cognigy technical support](/help/get-help). </li><li>If you have an on-premises installation, specify the `FEATURE_ENABLE_ELEVENLABS_STT_VENDOR=true` in the `values.yaml`.</li></ul>The Speechmatics STT vendor is hidden behind a feature flag. Activate Speechmatics on the Cognigy and Voice Gateway sides:<ul><li>If you have a SaaS installation, contact [Cognigy technical support](/help/get-help).</li><li>If you have an on-premises installation, specify the [feature flags](/voice-gateway/webapp/speech-services#speechmatics) in `values.yaml`.</li></ul> | Dial          |
    | STT Language                        | Selector      | Select the desired STT Language. For custom languages, use the following format: de-DE, fr-FR, en-US.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      | Dial          |
    | STT Label                           | CognigyScript | The alternative name of the vendor is the one you [specify in the Voice Gateway Self-Service Portal](/voice-gateway/webapp/applications#add-additional-tts-and-stt-vendor). If you have created multiple speech services from the same vendor, use the label to specify which service to use.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              | Dial          |
    | Disable STT Punctuation             | Toggle        | This parameter is active only when Google or Deepgram is selected from the **STT Vendor** list.<br /><br /> Prevents the transcription returned to the AI Agent from including punctuation marks.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          | Dial          |
    | Deepgram Model                      | Selector      | This parameter is active only when Deepgram is selected from the **STT Vendor** list. <br /><br /> Choose a model for processing submitted audio. Each model is associated with a tier. Ensure that the selected tier is available for the chosen STT language. For detailed information about Deepgram models, refer to the [Deepgram documentation](https://developers.deepgram.com/docs/model).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         | Dial          |
    | Endpointing                         | Toggle        | This parameter is active only if Deepgram or Deepgram Flux is selected from the **STT Vendor** list: <ul><li>**Deepgram** – detects the end of speech by monitoring user pauses. When a pause is detected, Deepgram finalizes the transcript with `speech_final = true`. The duration for detecting the end of speech is preconfigured with the default value of 10 milliseconds. If you want to change this value, use the **Endpointing Time** parameter.</li><li>**Deepgram Flux** – detects the end of speech using a confidence threshold. When a user's speech ends or the optional timeout is reached, the transcript is finalized. This approach ensures accurate, responsive multiturn transcription.</li></ul> For detailed information about the endpointing configuration, refer to the documentation for [Deepgram](https://developers.deepgram.com/docs/endpointing) or [Deepgram Flux](https://developers.deepgram.com/docs/flux/configuration).                                                                                                                                                                                                                                                                                                                                                                                                                            | Dial          |
    | Endpointing Time                    | Number        | This parameter is active only when Deepgram is selected from the **STT Vendor** list and the **Endpointing** toggle is enabled. <br /><br /> Customize the duration (in milliseconds) for detecting the end of speech. The default is 10 milliseconds of silence. After detecting silence, the system waits until either the speaker resumes or the required silence duration is reached. Once one of these conditions is met, the transcript is sent with `speech_final` set to `true`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   | Dial          |
    | Smart Formatting                    | Toggle        | This parameter is active only when Deepgram is selected from the **STT Vendor** list. <br /><br /> Deepgram's Smart Format feature applies additional formatting to transcripts to optimize them for human readability. Smart Format capabilities vary between models. When Smart Formatting is turned on, Deepgram will always apply the best-available formatting for your chosen model, tier, and language combination. For detailed examples, refer to the [Deepgram documentation](https://developers.deepgram.com/docs/smart-format). <br /><br /> Note that when Smart Formatting is turned on, punctuation will be activated, even if you have the Disable STT Punctuation setting enabled.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        | Dial          |
    | Google Model                        | Dropdown      | This parameter is active only when Google is selected from the **STT Vendor** list.<br /><br />Uses one of Google Cloud Speech-to-Text transcription models. The default value is `latest_short`. For a detailed list of Google models, refer to the [Transcription models](https://cloud.google.com/speech-to-text/docs/transcription-model#transcription_models) section in the Google Documentation. Keep in mind that the `default` value is a Google model type that can be used if other models don't suit your use case.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            | Dial          |
    | End of Turn Threshold               | Slider        | This parameter is active only if Deepgram Flux is selected from the **STT Vendor** list. Defines the confidence level for detecting when a user's speech ends. Higher values improve accuracy but may increase latency. The default value is `0.7`, meaning the system waits until it is 70% confident that the user has finished speaking before finalizing the transcript.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               | Dial          |
    | End of Turn Timeout in Milliseconds | Number        | This parameter is active only if Deepgram Flux is selected from the **STT Vendor** list. Defines the maximum number of milliseconds to wait before finalizing a turn. The timeout acts as a backup if the user stops speaking unexpectedly. The default value is `5000` (5 seconds), meaning the turn will be closed after 5 seconds of silence.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           | Dial          |
    | Transcription Webhook               | CognigyScript | The webhook is triggered with an HTTP POST whenever an interim or final transcription is received. If the **STT Vendor** and **STT Language** fields are empty, the system will use the default STT from the [Set Session Config Node](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details) (if it exists) or from the [Voice Gateway Self-Service Portal](/voice-gateway/webapp/speech-services). The parameter supports [CognigyScript](/ai/platform-features/cognigyscript), allowing it to accept dynamic content. For example, you can specify the URL as follows: `https://test-hook.com?contact={{ci.contact_name}}`. <br /><br /> Note if the [Voice Copilot](/ai/agents/deploy/endpoint-reference/voice-copilot) Endpoint is inactive, you can use any Webhook URL to receive voice call transcripts. However, when the Voice Copilot Endpoint is enabled, ensure that the specified Webhook URL is associated with it for processing.                                                                                                                                                                                                                                                                                                                                                                                                                        | Dial          |
    | Audio Stream Selection              | Selector      | Select the source of the audio stream: <ul><li>**Caller/Called** — both the incoming and outgoing audio streams of the caller and the called party.</li><li>**Caller** — the incoming and outgoing audio stream of the caller.</li><li>**Called** — the incoming and outgoing audio stream of the called party.</li></ul><br /> Ensure that the selected audio stream matches the language specified for transcription. If no audio stream is provided, the system will use the one set in the beginning, which should also match the language specified for transcription.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                | Dial          |
  </Accordion>

  <Accordion title="Answering Machine Detection">
    Answering Machine Detection (AMD) determines whether an outgoing call has been answered by a human or a machine (such as a voicemail or answering machine). When AMD is enabled, events are generated in the Flow to indicate the detection result, allowing you to take specific actions based on the outcome.

    For example, if a call is transferred to a user and AMD detects a machine, you will receive an event indicating the machine detection.

    | Parameter                    | Type     | Description                                                                                                                                                                        | Transfer Type |
    | ---------------------------- | -------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------- |
    | Answering Machine Detection  | Toggle   | Indicates whether a call has been answered by a human agent or a machine.                                                                                                          | Dial          |
    | Redirect on machine detected | Selector | Redirects the call back to the Flow in case a machine was detected.                                                                                                                | Dial          |
    | Redirect text                | Text     | Text to say after machine was detected before going back to the Flow. For example, `Mailbox detected, your call will be redirected back to the AI Agent. Please stay in the call`. | Dial          |
  </Accordion>

  <Accordion title="Advanced">
    <Note>
      For Voice Gateway 2025.19 and later, the Anchor Media parameter has been renamed to Media Path and includes more options. The previously available options correspond to:

      * **Full Media** — the toggled on Anchor Media parameter in earlier versions.
      * **Partial Media** — the toggled off Anchor Media parameter in earlier versions.

      The Media Path parameter is hidden behind the feature flag `FEATURE_DISABLE_VG_MEDIA_PATH`, with a default value of `false`. It is available only in environments where the feature flag is set to `false`.
    </Note>

    | Parameter                   | Type     | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           | Transfer Type |
    | --------------------------- | -------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------- |
    | Media Path                  | Dropdown | Controls the routing of RTP traffic through a media platform, such as FreeSwitch, for monitoring, transcoding, and security purposes. You can also [change this value during an active call](#changing-the-media-path-during-an-active-call) using the `updateCall` API. Select one of the following options:<ul><li>**Full Media** — routes all RTP traffic through a media platform. This option is used when complete media capture is required, for example, for logging, monitoring, or other processing workflows.</li><li>**Partial Media** — this option is selected by default. Routes through a media platform only when necessary. For example, to enable transcriptions or other optional processing.</li><li>**No Media** – routes directly from the caller's SBC to the far SBC or SIP trunk, without passing any media through the media platform. This option is primarily used for compliance purposes. When the option is selected, the system is disconnected from the media stream entirely, ensuring that sensitive data, such as credit card information, can't be captured.</li></ul> When you select **Full Media** or **Partial Media**, media anchoring process changes the addresses and ports of incoming and outgoing data:<ul><li>Outgoing data — the source IP address and port become the IP address and port of Voice Gateway. The destination IP address and port become the IP address and port of the public network. </li><li>Incoming data — the source IP address and port become the IP address and port of Voice Gateway. The destination IP address and port become the IP address and port of the PBX system or local telephone.</li></ul> | Dial          |
    | Custom Transfer SIP Headers | Toggle   | Data that needs to be sent as SIP headers in the generated SIP message.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               | All           |
    | Transfer SIP Headers        | JSON     | An object containing SIP headers that should be sent to the transferee. Each header is a key-value pair: `{ "Header-Name": "Header-Value" }`. The following example shows a JSON object containing two key-value pairs: `{ "X-Caller-ID": "+1234567890", "X-Request-ID": "abc-123" }`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                | All           |
    | Referred By                 | String   | This parameter is optional.<br /><br /> This setting allows you to change the original Referred By value, which can be a SIP URI or a user identifier such as a phone number. To define the Referred By value, you can use the following patterns:<ul><li>**SIP URI** — `sip:[referred-by]@custom.domain.com`. In this case, the entire SIP URI will be sent as the Referred-By header. Example: `"Referred-by": "sip:CognigyOutbound@custom.domain.com"`.</li><li>**User Identifier** — `sip:[referred-by]@[SIP FROM Domain from carrier config]`. Example, `"Referred-By": "sip:CognigyOutbound@sip.cognigy.ai"`.</li></ul>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         | Refer         |

    ### Changing the Media Path during an Active Call

    Besides setting the media path in the Flow, you can also change the media path during an active transferred call using the Voice Gateway API.

    This approach is useful for PCI compliance. For example, when a caller shares sensitive payment information, Voice Gateway, as an intermediate system, must be removed from the media stream. Audio is then routed directly between the caller and the contact center, bypassing Voice Gateway, reducing PCI data exposure and helping protect sensitive data.

    To remove Voice Gateway from the media path, send the [PUT /Accounts//Calls/](https://api-vg-trial.cognigy.ai/docs/#/Accounts/updateCallPut) API request with `{ "media_path": "noMedia" }` from the contact center to Voice Gateway after the call has been transferred. With this setting, Voice Gateway stops processing the call's audio and DTMF entirely, including for call transcription. After Voice Gateway confirms the API request with a `200` or `202` response, you can't revert to processing media for the transferred call. A `400`, `404`, or `422` response means the request failed, which keeps the call on its previous media path, and Voice Gateway continues processing audio and DTMF.

    Capture the call SID before the transfer takes place. During a transfer, a new call SID is created for the new call leg, but the original call SID remains unchanged. Pass the original call SID along during the transfer, for example, in SIP headers. This way, the contact center can reference the original call when triggering the API request at the correct moment.
  </Accordion>
</AccordionGroup>

### Troubleshooting

#### Possible 403 Error After Dial Transfer

In case you receive a 403 error log message after using a Transfer Node with Dial, check the related [PCAP](/voice-gateway/webapp/recent-calls) file in the Voice Gateway Self-Service Portal first. There you should get more information about the reason for the 403 error.

Some SBCs like Twilio may reject calls from unknown numbers.
In this case, it might be necessary to configure the P-Asserted-Identity header for the Transfer.
To do this, turn on the Custom Transfer SIP Headers toggle in the Advanced settings of the Transfer Node,
and enter the following in the JSON field:

```json theme={null}
{
  "P-Asserted-Identity": "<sip:phonenumber@host>"
}
```

#### Possible 487 Error After Dial Transfer

In case you receive a 487 error log message after using a Transfer Node with Dial,
check the related [PCAP](/voice-gateway/webapp/recent-calls) file in the Voice Gateway Self-Service Portal first.
There you should see a `CANCEL` method followed by the `487 Request Terminated` error.

This error indicates that the callee did not answer within the defined timeout.
Within the Flow, you will receive the [NO\_ANSWER](/voice-gateway/references/events/NO_ANSWER) call event.


# xApp: Get Session PIN
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/xApp/get-xApp-session-PIN



<Frame>
  <img alt="xApp Get Session PIN Node configuration panel" />
</Frame>

## Description

This Node generates a new xApp session PIN. You can access it using the [xApp: Get Session PIN](/xApps/tokens#xapp-session-pin) token or via `input.apps.session.pin`.

The PIN code is only valid for 5 minutes.

With this PIN, users can log in via an xApp PIN webpage.

### xApp PIN Page URL

Share the xApp PIN Page URL with users, where they can enter the PIN. The URL can be sent in the Flow with an [xApp PIN Page URL](/xApps/tokens#xapp-pin-page-url) Cognigy token.

## More Information

* [xApp Tokens](/xApps/tokens)
* [xApp Nodes](/ai/agents/develop/node-reference/xApp/overview)
* [xApp: Init Session](/ai/agents/develop/node-reference/xApp/init-xApp-session)
* [xApp: Show HTML](/ai/agents/develop/node-reference/xApp/set-html-xApp-state)
* [xApp: Show Adaptive Card](/ai/agents/develop/node-reference/xApp/set-AdaptiveCard-xApp-state)


# xApp: Init Session
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/xApp/init-xApp-session



<a href="/release-notes/4.51"><Badge>Updated in 4.51</Badge></a>

<Frame>
  <img alt="xApp Init Session Node configuration panel" />
</Frame>

## Description

This Node initializes a new xApp session.

After the execution of this Node, you can access the xApp in two different ways:

* [Without entering PIN Codes](#without-pin)
* [By entering PIN Codes](#with-pin)

### Without PIN

To require only a URL to access an xApp session, you can share a Session URL. This URL is available via the [xApp Session URL](/xApps/tokens#xapp-session-url) token.

The **xApp Session URL** token provides a URL, which is also accessible via `input.apps.url` in the input object.

With the provided link, a user can access the xApp page without requiring a PIN.

### With PIN

To allow users to access the xApp session via a PIN, share both a PIN and a PIN Page URL with the users.

This functionality is available via the [xApp PIN Page URL](/xApps/tokens#xapp-pin-page-url) token and the [xApp: Get Session PIN](/ai/agents/develop/node-reference/xApp/get-xApp-session-PIN) Node.

The **Get Session PIN** Node generates a PIN, which is accessible via the [xApp Session PIN token](/xApps/tokens#xapp-session-pin) or `input.apps.session.pin` in the input object.

The **xApp PIN Page URL** token provides a URL, which is also accessible via `input.apps.baseUrl` in the input object.

With the provided link, a user can access the PIN page and enter their provided PIN to gain access to the xApp.

## Parameters

<AccordionGroup>
  <Accordion title="Style Customization">
    | Parameter          | Type          | Description                                                                                                                                                                                                      |
    | ------------------ | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Background Color   | CognigyScript | A color that is displayed in the background of all Shell Page Screens. You can use CSS background-color values such as `hex`, `hsl+a`, and `rgb+a`.                                                              |
    | Text Color         | CognigyScript | A color that is used for textual content on all Shell Page Screens. You can use CSS color values such as `hex`, `hsl+a`, and `rgb+a`. Make sure that the text color has high contrast with the background color. |
    | Logo               | List          | You can choose to show the default logo, not any logo, or a custom logo. If you select the custom logo option, specify the Custom Logo URL.                                                                      |
    | Custom Favicon URL | CognigyScript | An image URL for the custom favicon.                                                                                                                                                                             |
    | Page Title         | CognigyScript | A text shown as a tab or window title. The title is displayed on all Shell Page Screens.                                                                                                                         |
  </Accordion>

  <Accordion title="xApp Screens">
    | Parameter           | Type          | Description                                                                                |
    | ------------------- | ------------- | ------------------------------------------------------------------------------------------ |
    | 'Loading' Text      | CognigyScript | Specify messages displayed to the user when the xApp page is loading                       |
    | 'Launch Error' Text | CognigyScript | Specify messages displayed to the user when the xApp page fails for some reason            |
    | 'xApp Error' Text   | CognigyScript | Specify messages displayed to the user when the xApp page encounters an error or is halted |
  </Accordion>

  <Accordion title="Intermediate Screen">
    The intermediate screen appears on the xApps page while the xApp is inactive, for example, after you have submitted data in the xApp and no other xApp is triggered.

    | Parameter          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
    | ------------------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Customization Type | Sets whether and how to customize the . Select one of the following options:<ul><li>**No customization** — displays  on the . This option is selected by default.</li><li>**Customized Texts** — displays custom text on the .</li><li>**Replace with xApps** — displays a custom page on the , optionally, with custom data. This option provides flexibility because you can create a page using HTML, CSS, and JavaScript and handle custom data dynamically.</li></ul> |
    | Text Override      | This parameter appears when you select **Customized Texts** in the **Customization Type** field. Enter the custom text to be displayed on the .                                                                                                                                                                                                                                                                                                                            |
    | xApp Template      | This parameter appears when you select **Replace with xApps** in the **Customization Type** field. Enter the URL to the custom page to be displayed on the .                                                                                                                                                                                                                                                                                                               |
    | xApp Template Data | This parameter is optional. It appears when you select **Replace with xApps** in the **Customization Type** field. Enter custom data in JSON format to be sent to the custom page. For example, you can use a [CognigyScript expression](/ai/platform-features/cognigyscript) in a JSON object, such as `{"name": "{{profile.firstname}}"}`, to fetch data from the Profile object and use it on the  as a personalized message with the end user's name.                  |
  </Accordion>

  <Accordion title="Connection Screen">
    The connection screen appears while the browser is establishing the connection to the xApp page.

    | Parameter          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
    | ------------------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Customization Type | Sets whether and how to customize the . Select one of the following options:<ul><li>**No customization** — displays  on the . This option is selected by default.</li><li>**Customized Texts** — displays custom text on the .</li><li>**Replace with xApps** — displays a custom page on the , optionally, with custom data. This option provides flexibility because you can create a page using HTML, CSS, and JavaScript and handle custom data dynamically.</li></ul> |
    | Text Override      | This parameter appears when you select **Customized Texts** in the **Customization Type** field. Enter the custom text to be displayed on the .                                                                                                                                                                                                                                                                                                                            |
    | xApp Template      | This parameter appears when you select **Replace with xApps** in the **Customization Type** field. Enter the URL to the custom page to be displayed on the .                                                                                                                                                                                                                                                                                                               |
    | xApp Template Data | This parameter is optional. It appears when you select **Replace with xApps** in the **Customization Type** field. Enter custom data in JSON format to be sent to the custom page. For example, you can use a [CognigyScript expression](/ai/platform-features/cognigyscript) in a JSON object, such as `{"name": "{{profile.firstname}}"}`, to fetch data from the Profile object and use it on the  as a personalized message with the end user's name.                  |
  </Accordion>
</AccordionGroup>

## More Information

* [xApp Tokens](/xApps/tokens)
* [xApp Nodes](/ai/agents/develop/node-reference/xApp/overview)
* [xApp: Get Session PIN](/ai/agents/develop/node-reference/xApp/get-xApp-session-PIN)
* [xApp: Show HTML](/ai/agents/develop/node-reference/xApp/set-html-xApp-state)
* [xApp: Show Adaptive Card](/ai/agents/develop/node-reference/xApp/set-AdaptiveCard-xApp-state)
* [xApps](/xApps/overview)


# xApp Nodes
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/xApp/overview



<Note>
  The existing payment terms for xApps are outlined in the [Billing](https://docs.cognigy.com/ai/administer/billing/) article. Cognigy reserves the right to change the billing of xApps in the future.
</Note>

The xApp Flow Nodes allow creating xApps, flexible micro-web applications for customer self-service. You can build the actual xApp pages by yourself in various ways, for example, by providing HTML code or by specifying an Adaptive Card in JSON format.

Cognigy supports the following Nodes:

<CardGroup>
  <Card title="xApp: Init Session" href="/ai/agents/develop/node-reference/xApp/init-xApp-session">
    This Node initializes a new xApp session.
  </Card>

  <Card title="xApp: Get Session PIN" href="/ai/agents/develop/node-reference/xApp/get-xApp-session-PIN">
    This Node generates a new xApp session PIN. You can access it using the [xApp: Get Session PIN](/xApps/tokens#xapp-session-pin) token or via `input.apps.session.pin`.
  </Card>

  <Card title="xApp: Show Adaptive Card" href="/ai/agents/develop/node-reference/xApp/set-AdaptiveCard-xApp-state">
    This Node builds an xApp page by providing an Adaptive Card in JSON format.
  </Card>

  <Card title="xApp: Show HTML" href="/ai/agents/develop/node-reference/xApp/set-html-xApp-state">
    This Node builds an xApp page by providing custom HTML code.
  </Card>
</CardGroup>

## More Information

* [xApps](/xApps/overview)
* [xApp Tokens](/xApps/tokens)
* [xApp Flow API](/xApps/api)


# xApp: Show Adaptive Card
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/xApp/set-AdaptiveCard-xApp-state



<a href="/release-notes/4.75"><Badge>Updated in 4.75</Badge></a>

<Frame>
  <img alt="xApp Show Adaptive Card Node configuration panel" />
</Frame>

## Description

An xApp page can be built by providing an adaptive card in JSON format.

To submit data back to the Flow, include an action of type `Submit` within the adaptive card.

To view submitted data, use the Interaction panel. You can access submitted data in the `data` object by clicking **INFO > Input**, or in the chat by adding a Say Node and specifying <code>Submit payload: \{\{JSON.stringify(input.data)}}</code> or <code>Submit payload: \{\{JSON.stringify(ci.data)}}</code>.

As an alternative, for more customization, you can [create your own Node](https://docs.cognigy.com/xApps/build/custom-xApp-node/) by implementing and uploading an Extension.

## Parameters

<AccordionGroup>
  <Accordion title="AdaptiveCard Definition">
    The [Adaptive Card Designer](https://adaptivecards.io/designer) can help you define an adaptive card. Copy the JSON card payload and paste it into the form.
  </Accordion>

  <Accordion title="Style Overrides">
    To override the background and interaction element colors, use CSS color codes.

    <AccordionGroup>
      <Accordion title="Waiting Behavior">
        To wait for a user input, for example, when the user clicks **Submit** on the Ordering page, activate the **Waiting Behavior** option.
        You can store the input result in the Context and use the data in the other Nodes.
      </Accordion>

      <Accordion title="xApp Overlay">
        By default, an xApp page opens in a separate browser tab.
        If you are using Webchat v3,
        you can override this behavior to open your xApp pages directly in Webchat v3.
        The pages are presented in the format of screens.

        The table lists the parameters for configuring xApp screens in the Webchat v3 interface.

        | Parameter                      | Type          | Description                                                                                                                                                                                                                          |
        | ------------------------------ | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
        | Header Title                   | CognigyScript | This parameter shows a title of the screen. The title displayed at the top panel of the Webchat v3 interface. If no title is provided, the top panel remains empty.                                                                  |
        | Show Close Icon                | Toggle        | This parameter displays the `x` button in the upper-right corner of the screen. This parameter is enabled by default.                                                                                                                |
        | Send Event on Close Icon Click | Toggle        | This parameter records the event when the user closes the screen without submitting. This parameter is disabled by default.                                                                                                          |
        | Auto Open                      | Toggle        | This parameter automatically opens the screen within the Webchat v3 interface. This setting is enabled by default. If you want xApp pages to open in browser tabs instead of within the Webchat v3 interface, activate this setting. |
        | Close Overlay on Submit        | Toggle        | This parameter closes the screen after the user clicks the Submit button. This parameter is enabled by default.                                                                                                                      |
        | Feedback Message on Submit     | CognigyScript | This parameter sends a message to the chat after the user clicks the Submit button.                                                                                                                                                  |
      </Accordion>
    </AccordionGroup>
  </Accordion>
</AccordionGroup>

## Examples

### Display Dynamic Values in Adaptive Cards

In adaptive cards, you can use data from the [Input](/ai/agents/develop/ai-agent-memory/input), [Context](/ai/agents/develop/ai-agent-memory/context), and [Profile](/ai/agents/develop/ai-agent-memory/profile) objects.
To display values dynamically, use [CognigyScript](/ai/platform-features/cognigyscript) to reference them.

<AccordionGroup>
  <Accordion title="Example 1: Display a Single Value">
    To display the user's first name by retrieving it from the Profile object, include this JSON in your adaptive card:

    ```json theme={null}
    {
      "type": "TextBlock",
      "text": "Hello, {{cp.firstName}}!"
    }
    ```

    If the user's first name is Alice, the xApp will display:

    ```text theme={null}
    Hello, Alice!
    ```
  </Accordion>

  <Accordion title="Example 2: Display an Array">
    Assume that you have the following array stored under `cart` in the Context object:

    ```json theme={null}
    {
      "cart": [
        { "name": "Apple", "price": 1.2 },
        { "name": "Banana", "price": 0.8 },
        { "name": "Orange", "price": 1.5 }
      ]
    }
    ```

    <Tabs>
      <Tab title="The Entry from the Array">
        To access a specific entry from the array, use its index.

        Adaptive Card JSON:

        ```json theme={null}
        {
          "type": "TextBlock",
          "text": "Item: {{cc.cart[0].name}}, Price: {{cc.cart[0].price}}"
        }
        ```

        Output:

        ```text theme={null}
        Item: Apple, Price: 1.2
        ```
      </Tab>

      <Tab title="All Entries from the Array">
        In the Flow editor, place the Code Node above the xApp: Show Adaptive Card Node to format the array into a string:

        ```javascript theme={null}
        const cart = cc.cart || [];

        cc.cartFormatted = cart
          .map(item => `- ${item.name}: $${item.price}`)
          .join('\n');
        ```

        The string is stored under `cartFormatted` in the Context object.

        Adaptive card JSON:

        ```json theme={null}
        {
          "type": "AdaptiveCard",
          "version": "1.5",
          "body": [
            { "type": "TextBlock", "text": "Your shopping cart:" },
            { "type": "TextBlock", "wrap": true, "text": "{{cc.cartFormatted}}" }
          ]
        }
        ```

        Output:

        ```text theme={null}
        Your shopping cart:
        - Apple: $1.2
        - Banana: $0.8
        - Orange: $1.5
        ```
      </Tab>

      <Tab title="Quick Debug with JSON.stringify">
        If you want to view the array for debugging, use `JSON.stringify` in the adaptive card.

        Adaptive card JSON:

        ```json theme={null}
        {
          "type": "AdaptiveCard",
          "version": "1.5",
          "body": [
            { "type": "TextBlock", "text": "Cart contents:" },
            { "type": "TextBlock", "wrap": true, "text": "{{JSON.stringify(cc.cart)}}" }
          ]
        }
        ```

        Output:

        ```text theme={null}
        [
          {"name":"Apple","price":1.2},
          {"name":"Banana","price":0.8},
          {"name":"Orange","price":1.5}
        ]
        ```
      </Tab>
    </Tabs>
  </Accordion>
</AccordionGroup>

## More Information

* [xApps](/xApps/overview)
* [xApp Nodes](/ai/agents/develop/node-reference/xApp/overview)
* [xApp: Init Session](/ai/agents/develop/node-reference/xApp/init-xApp-session)
* [xApp: Get Session PIN](/ai/agents/develop/node-reference/xApp/get-xApp-session-PIN)
* [xApp: Show HTML](/ai/agents/develop/node-reference/xApp/set-html-xApp-state)


# xApp: Show HTML
Source: https://docs.cognigy.com/ai/agents/develop/node-reference/xApp/set-html-xApp-state



<a href="/release-notes/4.75"><Badge>Updated in 4.75</Badge></a>

<Frame>
  <img alt="xApp Show HTML Node configuration panel" />
</Frame>

## Description

The xApp page can be built by providing custom HTML code. Once this Node is executed, this xApp page is displayed in the xApp Session.

You can access the xApp page SDK through a global `SDK` variable.
To submit a result back to the Flow, call `SDK.submit({ name: "John" });`.
The submitted payloads need to be serializable JSON objects.

The following example renders a button that submits on click:

`<button type="button" onclick="SDK.submit({ variant: 'button1' })">button 1</button>`.

To view submitted data, use the Interaction panel. You can access submitted data in the `data` object by clicking **INFO > Input**, or in the chat by adding a Say Node and specifying <code>Submit payload: \{\{JSON.stringify(input.data)}}</code> or <code>Submit payload: \{\{JSON.stringify(ci.data)}}</code>.

To access session data from within the HTML Code, use CognigyScript `{{JSON.stringify(cc.someContextKey)}}`.

As an alternative, for more customization, you can [create your own Node](https://docs.cognigy.com/xApps/build/custom-xApp-node/) by implementing and uploading an Extension.

## Parameters

| Parameter | Type | Description                                                                 |
| --------- | ---- | --------------------------------------------------------------------------- |
| Content   | List | Select the content format:<ul><li>HTML document</li><li>HTML body</li></ul> |

<AccordionGroup>
  <Accordion title="Waiting Behavior">
    To wait for a user input, for example, when the user clicks **Submit** on the Ordering page, activate the **Waiting Behavior** option.
    You can store the input result in the Context and use the data in the other Nodes.
  </Accordion>

  <Accordion title="xApp Overlay">
    By default, an xApp page opens in a separate browser tab.
    If you are using Webchat v3,
    you can override this behavior to open your xApp pages directly in Webchat v3.
    The pages are presented in the format of screens.

    The table lists the parameters for configuring xApp screens in the Webchat v3 interface.

    | Parameter                      | Type          | Description                                                                                                                                                                                                                          |
    | ------------------------------ | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Header Title                   | CognigyScript | This parameter shows a title of the screen. The title displayed at the top panel of the Webchat v3 interface. If no title is provided, the top panel remains empty.                                                                  |
    | Show Close Icon                | Toggle        | This parameter displays the `x` button in the upper-right corner of the screen. This parameter is enabled by default.                                                                                                                |
    | Send Event on Close Icon Click | Toggle        | This parameter records the event when the user closes the screen without submitting. This parameter is disabled by default.                                                                                                          |
    | Auto Open                      | Toggle        | This parameter automatically opens the screen within the Webchat v3 interface. This setting is enabled by default. If you want xApp pages to open in browser tabs instead of within the Webchat v3 interface, activate this setting. |
    | Close Overlay on Submit        | Toggle        | This parameter closes the screen after the user clicks the Submit button. This parameter is enabled by default.                                                                                                                      |
    | Feedback Message on Submit     | CognigyScript | This parameter sends a message to the chat after the user clicks the Submit button.                                                                                                                                                  |
  </Accordion>

  <Accordion title="xApp Page SDK">
    When hosted within Cognigy, the xApp page or any generic HTML xApp can load the xApp page SDK using a fixed URL:
    `<script src="/sdk/app-page-sdk.js"></script>`.
    Loading this file exposes the xApp page SDK as a global instance called `SDK`.
    To access its methods, call `SDK.someExampleMethod()` after loading the SDK file. The SDK code isn't loaded asynchronously.

    | Function                                    | Usage                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         |
    | ------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | `SDK.getAppTemplateData():Promise<unknown>` | Returns the xApp Template Data associated with this xApp page.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
    | `SDK.submit(result: Object):Promise<void>`  | Submits the provided "result" object to the Flow. Has to be a JSON serializable object.                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
    | `SDK.getCustomReadyHandler():Function`      | Returns a custom ready state handler. By default, the xApp page is ready as soon as the document is fully loaded.<br />If you want to fetch additional content before showing the Page, you can override the `ready` event by getting a custom `ready handler` and calling it as soon as your xApp page is ready manually. <br /> The end user sees the xApp Loading Screen on the Shell Page.<br />Note: If loading the xApp page takes more than 10 seconds, the xApp is treated as `Failed to load` without the possibility of recovering. |
  </Accordion>
</AccordionGroup>

## More Information

* [xApps](/xApps/overview)
* [xApp Nodes](/ai/agents/develop/node-reference/xApp/overview)
* [xApp: Init Session](/ai/agents/develop/node-reference/xApp/init-xApp-session)
* [xApp: Get Session PIN](/ai/agents/develop/node-reference/xApp/get-xApp-session-PIN)
* [xApp: Show Adaptive Card](/ai/agents/develop/node-reference/xApp/set-AdaptiveCard-xApp-state)


# Channels
Source: https://docs.cognigy.com/ai/agents/develop/nodes/channels



<a href="/release-notes/4.48"><Badge>Updated in 4.48</Badge></a>

*Channels* allow you to use your Flow with multiple Endpoints, ensuring that output messages are correctly reproduced in both chat and voice conversations.

You can use multiple channels simultaneously and customize the AI Agent's messages for different voice and chat platforms. For example, if you deploy an AI Agent on Webchat v3, WhatsApp, and Voice Gateway, you can have 1 Flow with 3 corresponding channels and their corresponding Endpoints. This setup ensures that AI Agents deliver messages in the format supported by each channel.

You can configure channels in the [Say](/ai/agents/develop/node-reference/basic/say), [Question](/ai/agents/develop/node-reference/basic/question), and [Optional Question](/ai/agents/develop/node-reference/basic/optional-question) Nodes. Voice channels support Speech Synthesis Markup Language (SSML), and text-based channels can generate different content type outputs, such as:

* Text
* Image
* Audio
* Video
* Adaptive cards

## Available Channels

<AccordionGroup>
  <Accordion title="Default AI Channel">
    The default AI channel is a universal channel and its settings correspond to the parameters in the [Say Node](/ai/agents/develop/node-reference/basic/say). Use this channel to get more flexibility around different platforms.
  </Accordion>

  <Accordion title="Alexa">
    <Warning>
      * If the user triggers more than one Say Node in a Flow execution with the Alexa channel, the AI Agent concatenates each Say Nodes' SSML or text outputs as one large SSML statement.
      * If the Flow outputs more than one Card, the AI Agent sends only the last one.
      * If the Flow outputs more than one Display configuration, the AI Agent uses only the last one.
    </Warning>

    The Alexa channel defines what and how an Amazon Echo-enabled system provides answers.

    | Parameter                       | Description                                                                                                                                                                                                                                                                                                                            |
    | ------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Session Lifecycle               | This parameter defines if the AI Agent should end the session after the Node is executed.                                                                                                                                                                                                                                              |
    | Output Speech                   | The content and pronunciation of the output speech. This parameter supports SSML and JSON directives. For more details on JSON directives, read the [Amazon Alexa documentation](https://developer.amazon.com/de/docs/custom-skills/request-and-response-json-reference.html).                                                         |
    | Reprompt                        | The content and pronunciation of the reprompt speech in case the user doesn't respond after a certain time. This parameter supports SSML and JSON directives. For more details on JSON directives, read the [Amazon Alexa documentation](https://developer.amazon.com/de/docs/custom-skills/request-and-response-json-reference.html). |
    | Home Cards                      | A Card that provides additional visual information. You can send Cards if the user has a device with a screen that is connected to the Alexa application.                                                                                                                                                                              |
    | Display Templates for Echo Show | This parameter defines how Amazon Echo Show devices display content.                                                                                                                                                                                                                                                                   |
  </Accordion>

  <Accordion title="Messenger">
    The Messenger channel defines how AI Agents provide answers in Facebook Messenger.

    | Output Type   | Parameters                                                                                                  |
    | ------------- | ----------------------------------------------------------------------------------------------------------- |
    | Quick Replies | <ul><li>Message</li><li>Title</li><li>Payload</li><li>Image URL</li><li>Condition</li></ul>                 |
    | Buttons       | <ul><li>Button Type</li><li>Title</li><li>Payload</li><li>Condition</li></ul>                               |
    | Gallery       | <ul><li>Title</li><li>Subtitle</li><li>Image URL</li><li>Default action web URL</li><li>Condition</li></ul> |
    | Attachment    | <ul><li>Attachment Type</li><li>URL</li></ul>                                                               |

    For more details on how to configure the parameters, read the [Facebook Messenger documentation](https://developers.facebook.com/docs/messenger-platform/send-messages).

    <Accordion title="Use Code Nodes to Output Facebook Markup">
      You can use the [output action in Code Nodes](/ai/for-developers/code/api-functions) to send a JSON object directly to Facebook. To do so, set the following code as the data property:

      ```json theme={null}
      {
          "_cognigy": {
              "_facebook": {      
                  "message": {
                      // this contains your message to facebook
                  }
              }
          }
      }
      ```
    </Accordion>
  </Accordion>

  <Accordion title="Webchat">
    The Webchat channel defines how AI Agents provide answers through [Webchat v2](/webchat/v2/overview).

    <Note>
      The Webchat channel has similar configuration options to the Messenger channel. Activate the **Use Facebook Messenger channel** toggle to use the same configuration options in both channels.
    </Note>

    | Output Type   | Parameters                                                                                                                                                       |
    | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Quick Replies | <ul><li>Message</li><li>Title</li><li>Payload</li><li>Image URL</li><li>Condition</li></ul>                                                                      |
    | Buttons       | <ul><li>Button Type</li><li>Title</li><li>Payload</li><li>Condition</li></ul>                                                                                    |
    | Gallery       | <ul><li>Title</li><li>Subtitle</li><li>Image URL</li><li>Default action web URL</li><li>Condition</li></ul>                                                      |
    | Attachment    | <ul><li>Attachment Type</li><li>URL</li></ul>                                                                                                                    |
    | List          | <ul><li>Title</li><li>Subtitle</li><li>Image URL</li><li>Default action web URL</li></ul>                                                                        |
    | Adaptive Card | A `$schema` property for the adaptive card. You can find supported versions in the [Content Conversion](/ai/agents/deploy/endpoints/content-conversion) article. |
  </Accordion>

  <Accordion title="LINE">
    The LINE channel defines how AI Agents provide answers in LINE.

    | Output Type | Description                                                                                                                                          |
    | ----------- | ---------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Text        | A simple text message.                                                                                                                               |
    | Custom JSON | A valid LINE message object. See the [LINE documentation](https://developers.line.me/en/reference/messaging-api/) for further details and templates. |
  </Accordion>

  <Accordion title="Twilio">
    <Warning>
      Make sure the Twilio Markup Language (TwiML) provided in the Node editor is valid. If the AI Agent sends an invalid TwiML to Twilio, the call immediately fails or doesn't initiate.

      Make sure that the content of the TwiML is escaped XML.
    </Warning>

    The Twilio channel defines how AI Agents provide answers in Twilio.

    | Output Type | Description                                                                                                                                                                                           |
    | ----------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Text        | A simple text message.                                                                                                                                                                                |
    | TwiML       | A valid TwiML. See the [Twilio documentation](https://www.twilio.com/docs/voice/twiml?code-sample=code-say-hello-to-an-inbound-caller\&code-language=output-twiml) for further details and templates. |

    **Amazon Polly Voice**

    In the **Twilio Endpoint** settings, you can select the **Amazon Polly** voice. For more information about Amazon Polly features, read the [Twilio documentation](https://www.twilio.com/docs/voice/twiml/say/text-speech#amazon-polly).
  </Accordion>

  <Accordion title="Twilio SMS">
    <Warning>
      Make sure the Twilio Markup Language (TwiML) provided in the Node editor is valid. If the AI Agent sends an invalid TwiML to Twilio, the call immediately fails or doesn't initiate.

      Make sure that the content of the TwiML is escaped XML.
    </Warning>

    The Twilio SMS channel defines how AI Agents provide answers through Twilio SMS.

    | Output Type | Description                                                                                                                            |
    | ----------- | -------------------------------------------------------------------------------------------------------------------------------------- |
    | Text        | A simple text message.                                                                                                                 |
    | TwiML       | A valid TwiML object. See the [Twilio documentation](https://www.twilio.com/docs/sms/twiml/message) for further details and templates. |
  </Accordion>

  <Accordion title="Microsoft Teams">
    <Warning>
      If the user triggers more than one Say Node in one Flow execution with the Microsoft Teams channel, the AI Agent:

      * Concatenates each Say Nodes' text or Microsoft Teams text outputs as one message.
      * Uses only the JSON field output of the last Node triggered.
    </Warning>

    With the Microsoft Teams channel, you can send either text or adaptive cards. Refer to [Deployment a Microsoft Teams Endpoint](/ai/agents/deploy/endpoint-reference/azure-bot-services) page for information on how to send messages.

    | Output Type   | Description                                                                                                                                                                                                                                                        |
    | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Text          | A simple text message.                                                                                                                                                                                                                                             |
    | JSON          | A valid JSON in the Bot Framework/Microsoft Teams format. See the [Microsoft documentation](https://docs.microsoft.com/en-us/azure/bot-service/rest-api/bot-framework-rest-connector-add-rich-cards?view=azure-bot-service-3.0) for further details and templates. |
    | Adaptive Card | A `$schema` property for the adaptive card. You can find supported versions in [Content Conversion](/ai/agents/deploy/endpoints/content-conversion).                                                                                                               |
  </Accordion>

  <Accordion title="Voice Gateway">
    With the Voice Gateway channel, you can configure the content and pronunciation of the AI Agents voice output.

    You can also configure Voice Gateway settings using Activity Parameters. To do this, activate the **Set Activity Parameters** toggle. You can find a description of the parameter configurations in [Voice Gateway Parameter Details](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details).
  </Accordion>

  <Accordion title="8x8">
    The **8x8** channel defines how AI Agents provide answers in 8x8.

    | Output Type   | Description                                                                                                                |
    | ------------- | -------------------------------------------------------------------------------------------------------------------------- |
    | Text          | A simple text message.                                                                                                     |
    | Custom JSON   | A valid 8x8 message object.                                                                                                |
    | Adaptive Card | An adaptive card. You can find supported versions in [Content Conversion](/ai/agents/deploy/endpoints/content-conversion). |

    <Accordion title="Custom JSON example">
      ```json theme={null}
      {
          "authorType": "customer",
          "text": "Hello!",
          "attachments": [
              {
                  "id": "string"
              }    
          ],
          "cards": [
              {
                  "contentType": "application/vnd.microsoft.card.adaptive",
                  "content": {}
              }
          ]
      }
      ```
    </Accordion>
  </Accordion>

  <Accordion title="Genesys">
    The **Genesys** channel defines how AI Agents provide answers in Genesys.

    | Output Type | Description                                                                        |
    | ----------- | ---------------------------------------------------------------------------------- |
    | Text        | A simple text message.                                                             |
    | Custom JSON | A valid Genesys Cloud message object. Defines more complex messages and templates. |

    <Accordion title="Custom JSON example">
      ```json theme={null}
      {
          "replymessages": [],
          "parameters": {
              "command": "handover"
          }
      }
      ```
    </Accordion>

    This message suggests that the AI Agent working on the Cognigy side has finished its task. Also, the message means that the control is handed over to the logic within the [Genesys architect flow](https://help.mypurecloud.com/articles/use-architect-to-create-a-bot-for-an-inbound-flow/).
  </Accordion>

  <Accordion title="Bandwidth">
    The **Bandwidth** channel defines how AI Agents provide answers in Bandwidth.

    | Output Type | Description                                                                    |
    | ----------- | ------------------------------------------------------------------------------ |
    | Text        | A simple text message.                                                         |
    | Custom JSON | A valid Bandwidth message object. Defines more complex messages and templates. |

    <Accordion title="Custom JSON example">
      ```json theme={null}
      {
          "activities": [
              {
                  "id": "123456789",
                  "timestamp": "1999-12-31T23:59:59.999Z", 
                  "type": "message",
                  "text": "Hello. This is my message."
              }
          ]
      }
      ```
    </Accordion>
  </Accordion>
</AccordionGroup>

## More Information

* [Say Node](/ai/agents/develop/node-reference/basic/say)
* [Question Node](/ai/agents/develop/node-reference/basic/question)
* [Optional Question Node](/ai/agents/develop/node-reference/basic/optional-question)


# Node Search
Source: https://docs.cognigy.com/ai/agents/develop/nodes/node-search



<a href="/release-notes/2026.12"><Badge>Updated in 2026.12</Badge></a>

With the *Node Search* field, you can search through the Nodes in the [Flow editor](/ai/agents/develop/projects-and-flows/editor).

## Restrictions

* The Node Search field applies only to [specific content](#how-to-use) and doesn't match expressions in the following fields:

<Accordion title="Non-searchable fields">
  | Field                                                    | Node                                                                                                                      |
  | -------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------- |
  | **Operator**                                             | [Lookup Node](/ai/agents/develop/node-reference/logic/lookup), when **CognigyScript** is selected from the **Type** list. |
  | **Locale**                                               | [Switch Locale Node](/ai/agents/develop/node-reference/logic/switch-locale)                                               |
  | **System Prompt**                                        | [LLM Prompt Node](/ai/agents/develop/node-reference/service/llm-prompt)                                                   |
  | **Instructions and Context**                             | [AI Agent Node](/ai/agents/develop/node-reference/ai/ai-agent)                                                            |
  | **Headers** and **Parameters**                           | [HTTP Request Node](/ai/agents/develop/node-reference/service/http-request)                                               |
  | **Key** and **Value** in the **Webhook Headers** section | AI Agent and [LLM Prompt](/ai/agents/develop/node-reference/service/llm-prompt) Nodes                                     |
  | **Mock Code**                                            | All Nodes                                                                                                                 |
</Accordion>

## How to Use

To open the Node Search field, click <img alt="search" /> in the lower-left corner of the Flow editor.

<ExpandableTable />

### Search Matches in the Flow Editor

When you enter a text into the Node Search field, the Flow editor highlights matching Nodes in yellow and scrolls to the first match. Use <img alt="previous match" /> and <img alt="next match" /> to navigate between matches. Click <img alt="clear search" /> to reset the search.

<Accordion title="Toggle Search Results">
  To close the Node Search field but keep the matching Nodes highlighted, click
  <img alt="search" /> again.
</Accordion>

### Search Matches in the Node Editor

The Node search also matches text in the Node editor. When you open a matching Node, the field, section, or tab that matches the search text is highlighted.

<Frame>
  <img alt="Node editor with a highlighted field matching the Node search text" />
</Frame>

## More Information

* [Nodes](/ai/agents/develop/nodes/overview)
* [Flows](/ai/agents/develop/projects-and-flows/overview)
* [Node Editor](/ai/agents/develop/nodes/overview#working-with-nodes)


# Nodes
Source: https://docs.cognigy.com/ai/agents/develop/nodes/overview



<a href="/release-notes/2025.24"><Badge>Updated in 2025.24</Badge></a>

*Nodes* are the fundamental building blocks of [Flows](/ai/agents/develop/projects-and-flows/overview). Nodes define the logic of AI Agents and are used to create complex conversational experiences. Each Node represents a specific action, decision point, or interaction in the Flow. Users can configure and connect Nodes to design guided interactions between AI Agents and users.

Cognigy provides a wide range of [Node types](/ai/agents/develop/node-reference/overview) to allow diverse conversational experiences.

Some Node types, for example, [If Nodes](/ai/agents/develop/node-reference/logic/if) and [Lookup Nodes](/ai/agents/develop/node-reference/logic/lookup), include child Nodes to control the Flow execution logic. Child Nodes function as a decision branch in the Flow, allowing different outcomes depending on the user input.

## Working with Nodes

<Tabs>
  <Tab title="GUI">
    You can create and delete Nodes in the [Flow editor](/ai/agents/develop/projects-and-flows/editor). To edit a Node, open the Node editor by clicking the Node in the Flow editor. Node editors include settings specific to each Node type. In the Flow editor, you can also right-click a Node to copy, cut, and paste the Node as well as add comments to it.

    <Tip>
      To speed up building your AI Agent, use [keyboard shortcuts](/ai/administer/hotkeys).
    </Tip>
  </Tab>

  <Tab title="API">
    You can view, create, edit, and delete Nodes using the [Cognigy API](https://api-trial.cognigy.ai/openapi#tag--Charts-v2.0). You can also search and move Nodes and add Locales to them. For Nodes that support Generative AI output, you can generate this kind of output through the API as well.
  </Tab>
</Tabs>

### JSON Data Fields in the Node Editor

Some Nodes, such as the [Say Node](/ai/agents/develop/node-reference/basic/say) or [Question Node](/ai/agents/develop/node-reference/basic/question), include JSON fields to send a `data` object with the output.

<Note>
  The JSON fields enforce JSON formatting at the root level. If you directly enter an array into a JSON field, for example, `["name", "surname", "email"]`, the Node unpacks the array into an object. The result is an object at the root level, for example, `{"0": "name", "1": "surname", "2": "email"}`.
</Note>

### Set Nodes as Entrypoints

If your AI Agent has, for example, a [Go To Node](/ai/agents/develop/node-reference/logic/go-to) or an [Execute Flow Node](/ai/agents/develop/node-reference/logic/execute-flow) to switch between Flows, you can set a Node as an *Entrypoint*. An Entrypoint is the Node at which the Flow starts. The **Start** Node is the default Entrypoint. To set a Node as an Entrypoint:

* right-click the Node in the Flow editor and select **Set Entrypoint**.
* select <img alt="menu" /> **> Set Entrypoint** in the Node editor.

<Accordion title="Naming Entrypoints">
  Define unique names for each Entrypoint Node using the **Label** field in the Node editor's **Settings** section. This approach is helpful to easily differentiate a Node with an Entrypoint from other Nodes of the same type.
</Accordion>

### Create Flows from Node Branches

You can split complex Flows and create new ones from a child Node. To do so, right-click the child Node and select the **Create Flow from Branch** option in the context menu.

When you create a Flow from a child Node, you move the Nodes following the child Node to the new Flow. In the original Flow, these Nodes are replaced with an Execute Flow Node. The Execute Flow Node triggers the new Flow.

<Warning>
  NLU and Flow settings aren't copied to the new Flow. Check and adjust the **NLU** tab settings of the new Flow.
</Warning>

## Dynamic Content in Nodes

Most Nodes support [CognigyScript](/ai/platform-features/cognigyscript) and [Tokens](/ai/platform-features/tokens) for dynamic content use.

## Fallback Value in Node Fields

A *fallback value* is a default value automatically applied when required fields are empty due to updates in a Node. Fallback values keep existing Nodes working when new fields are added or existing ones are changed, so you don't need to reconfigure them immediately.

Fallback values are typically applied in the following cases:

* When a Node field is updated in a new release version.
* When a new Node field is added to the Node.
* When you upload a new version of the Node Extension.

Fields that use a fallback value are marked with the fallback icon <img alt="fallback" />. After you update and save the field, the icon disappears.

## More Information

* [CognigyScript](/ai/platform-features/cognigyscript)
* [Tokens](/ai/platform-features/tokens)
* [Flows](/ai/agents/develop/projects-and-flows/overview)


# Persona
Source: https://docs.cognigy.com/ai/agents/develop/persona



*Personas* are the digital identities of AI Agents. A persona gives the AI Agent a consistent personality, tone, and voice.

You can apply the same persona across different tasks, which makes the AI's interactions with users more predictable and flexible.

Key Benefits:

* **Consistent Branding**. Ensures a uniform personality and tone across multiple customer touchpoints.
* **Flexible Task Assignment**. Allows one persona to handle different jobs in various [Flows](/ai/agents/develop/projects-and-flows/overview).
* **Enhanced User Trust**. Provides a recognizable and reliable interaction style for users.

## Concepts

| **Concept**     | **Description**                                                                                                                                                           |
| --------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Persona         | Defines how the AI Agent interacts, including its personality, voice, tone, and basic knowledge. Acts as the how-to template for interaction: task-agnostic and reusable. |
| Job             | Defines what the AI Agent performs. Task-specific logic assigned to [AI Agent Nodes](/ai/agents/develop/node-reference/ai/ai-agent) within Flows.                         |
| Knowledge Store | Defines an external knowledge base, for example, FAQs, product details that the AI Agent may reference during execution.                                                  |

### Key Principles

* A persona can be reused across many Jobs.
* Jobs are defined at the Node level in Flows. A persona may power multiple Flows.

<Accordion title="Example: Persona Sophie">
  Sophie is an AI persona with a friendly and knowledgeable voice. She maintains this consistent personality even as her roles change.

  * **Flow A:** Restaurant Guide: When acting as a restaurant guide, Sophie uses her cheerful, helpful persona to recommend dining options, provide directions, and help users book reservations. Her voice is welcoming and enthusiastic, making the user feel like they're talking to a well-informed friend.
  * **Flow B:** Delivery Assistant: As a delivery assistant, Sophie's core voice remains the same, but her focus shifts to efficiency and clear communication. She uses her dependable, friendly persona to provide real-time updates on orders, confirm delivery details, and resolve any issues with a reassuring tone. Her goal is to make the delivery process as smooth and stress-free as possible.
</Accordion>

## Use Cases

* **Customer Support**. A persona named Sophie guides users through troubleshooting in one Flow and handles billing inquiries in another, maintaining her consistent tone.
* **Multichannel Assistance**. A persona named Sophie assists customers via chat in one Flow and voice interactions in another, ensuring a seamless brand experience.

## How to Use

<Steps>
  <Step title="Create a Persona">
    Create a persona by defining general settings, speaking style, instructions, knowledge, and data privacy.
    For detailed instructions, refer to [Manage AI Agents](/ai/agents/develop/manage-ai-agents).

    <Frame>
      <video title="Creating a new AI Agent persona in Cognigy" aria-label="The video shows a step-by-step configuration of an AI Agent persona." />
    </Frame>
  </Step>

  <Step title="Assign a Job">
    When you created a persona, Cognigy generates:

    * A new Flow.
    * An [AI Agent Node](/ai/agents/develop/node-reference/ai/ai-agent) with the persona you created.
    * Pre-filled job instructions.

    <Frame>
      <img alt="AI Agent job configuration with pre-filled instructions" />
    </Frame>
  </Step>

  <Step title="Test and Deploy">
    Test your AI Agent in the [Interaction Panel](/ai/agents/test/interaction-panel/overview), then deploy by selecting one of the Endpoints. You can also test it via [Demo Webchat](/webchat/demo).
  </Step>
</Steps>

## What's Next?

* Go to the [Job Market](/ai/agents/overview) to hire a skilled AI Agent to avoid creating it manually.
* Add [tools](/ai/agents/develop/node-reference/ai/ai-agent#child-nodes) and use the [MCP protocol](/ai/agents/develop/node-reference/ai/ai-agent#mcp-tool) to extend your AI Agent's capabilities.
* Explore how to [orchestrate interactions between different personas](/ai/agents/develop/node-reference/ai/ai-agent-handover).
* Learn how to effectively embed [knowledge into your AI Agent](/ai/agents/develop/knowledge-ai/overview) to improve responses.


# Collaborative Editing
Source: https://docs.cognigy.com/ai/agents/develop/projects-and-flows/collab-editing



<a href="/release-notes/2025.12"><Badge>Updated in 2025.12.0</Badge></a>

The *Collaborative Editing* feature allows multiple users to edit a Flow simultaneously. With Collaborative Editing, you can see the avatars of other users editing the Flow and Nodes. Their changes appear in the Flow and Node editors in real time.

<div>
  <video title="Collaborative editing function shown as multiple users editing a Flow simultaneously" aria-label="Collaborative editing function shown as multiple users editing a Flow simultaneously">
    <source />
  </video>
</div>

## Key Benefits

* **Visual Indication of User Presence**. You can see what users are working on in the Flow editor.
* **Easy Collaboration**. You can start discussions on Nodes to ask questions, make suggestions, or share ideas. Additionally, you can mention other users directly in discussion threads. Mentioning other users triggers an email notification to them.
* **Simultaneous Editing**. The Flow and Node editors show when users are working simultaneously on the same Flows and Nodes to prevent users from overwriting each other's changes.

## User Presence

The Flow editor shows the presence of other users with icons including initials:

* in the upper-right corner of the Flow editor when other users have the same Flow open.
* attached to the upper-right corner of Nodes when other users are currently editing them.

## Mouse Cursors

Mouse cursors show the position of other users editing the same Flow.

## Discussions

To start a discussion, right-click the Node and select **Start Discussion**. You can enter the first comment in the **Start Discussion** field.

Nodes with discussions display <img alt="active discussion" /> attached to them. In a discussion thread, you can add comments and mention other users. To mention other users, type `@` and select the user from the list of suggestions. Only Project members are included in the list of suggestions. Mentioning other users triggers an email notification to them.

You can start multiple discussion threads on a Node. This way, you can open multiple discussions on a Node and separate the topics if necessary. To start a new discussion thread, follow these steps:

1. Right-click the Node and select **Open Discussions**. The discussion menu opens.
2. Enter a comment in the **Start Discussion** field.

## Simultaneous Editing

### Flow Editor

The Flow editor shows when other users make changes to the Flow you are working on. Whenever other users make changes to the Flow, for example, by adding or editing a Node, the Flow editor displays the following:

* A refresh button in the upper-right corner. Click the refresh button to view the latest changes.
* The number of changes made by other users next to the refresh button.

### Node Editor

The Node editor shows the following:

* When multiple users open the same Node, the user icon appears in the upper-right corner.
* When another user makes changes, the Node editor is locked for all other users and displays <img alt="locked" />.
* When the changes are saved, the Node editor displays the <img alt="refresh" /> button. Then, you can:
  * Click the <img alt="refresh" /> button to refresh the Node editor and view the latest changes.
  * Select <img alt="menu" /> **> Edit** to unlock and keep editing the Node. The changes you save after unlocking the Node overwrite the changes other users previously made.

## More Information

* [Flows](/ai/agents/develop/projects-and-flows/overview)
* [Flow Editor](/ai/agents/develop/projects-and-flows/editor)
* [Nodes](/ai/agents/develop/nodes/overview)


# Flow Editor
Source: https://docs.cognigy.com/ai/agents/develop/projects-and-flows/editor



<a href="/release-notes/2025.12"><Badge>Updated in 2025.12.0</Badge></a>

The *Flow Editor* is a user interface (UI) where you can see a graphic representation of a [Flow](/ai/agents/develop/projects-and-flows/overview), the *Flow Chart*, and design AI Agents' conversations. In the Flow editor, the Flow consists of Nodes that represent various actions or decision points. The [Nodes](/ai/agents/develop/nodes/overview) are linked to one another by lines that indicate the conversation's path, from the **Start** Node to the **End** Node.

The Flow editor allows for [Collaborative Editing](/ai/agents/develop/projects-and-flows/collab-editing). This means that multiple users can edit the same Flow at the same time.

A green arrow next to a Node indicates the Flow Entrypoint. The Entrypoint is where the AI Agent expects a user entry. The Flow editor highlights the path of the conversation in green. The green path allows you to understand the Flow execution path at each step, depending on the input in the [Interaction Panel](/ai/agents/test/interaction-panel/overview).

## Limitations

* You can add up to 100 Nodes in the Flow editor.

## Exploring the Flow Editor

The following image shows an overview of the Flow editor:

<Frame>
  <img alt="Cognigy Flow editor interface overview" />
</Frame>

<AccordionGroup>
  <Accordion title="1. Top Bar">
    Besides the controls from the [Project Page](/ai/overview/user-interface/project-page) top bar, the Flow editor includes the following elements from left to right:

    | UI Element                             | Type | Description                                                                                                                                                                                                                                                                                                                    |
    | -------------------------------------- | ---- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Flow List                              | List | Search for Flows by name and switch between them.                                                                                                                                                                                                                                                                              |
    | Locale List                            | List | Switch between Locales.                                                                                                                                                                                                                                                                                                        |
    | Chart                                  | Tab  | Display the Flow chart.                                                                                                                                                                                                                                                                                                        |
    | NLU                                    | Tab  | Display the NLU settings, where you can configure [Intents](/ai/platform-features/nlu/intents/overview), States, [Slot Fillers](/ai/platform-features/nlu/slot-fillers), as well as attach [Flows](/ai/platform-features/nlu/attachments#attached-flows) and [Lexicons](/ai/platform-features/nlu/slots/user-defined/lexicon). |
    | Settings                               | Tab  | Display the [Flow Settings](/ai/agents/develop/projects-and-flows/overview#settings), [Default Context](/ai/agents/develop/ai-agent-memory/context), and [Locale Settings](/ai/platform-features/translation-and-localization/localization).                                                                                   |
    | <img alt="other users viewing Flow" /> | Icon | Shows other users who have the current Flow open and active.                                                                                                                                                                                                                                                                   |
  </Accordion>

  <Accordion title="2. Flow Chart">
    |                 Button                 | Description                                                                             |
    | :------------------------------------: | --------------------------------------------------------------------------------------- |
    |         <img alt="add node" />         | Click to add a new Node or [Extension Node](/ai/for-developers/extensions) to the Flow. |
    |          <img alt="comment" />         | Click to see the comments in the Node.                                                  |
    | <img alt="other users editing Flow" /> | Indicates other users who are editing the same Flow.                                    |
    | <img alt="other users editing Node" /> | Indicates that other users are editing the Node to which the icon is attached.          |
    |        <img alt="discussion" />        | Click to open the Discussion menu on the Node and access Discussion Threads.            |

    <Note>
      The Flow editor supports drag and drop. To copy a Node, simply hold the <kbd>Alt</kbd> or <kbd>Option</kbd> while dragging the Node to another position.
    </Note>
  </Accordion>

  <Accordion title="3. Chart Controls">
    |             Control            | Description                                                                                                                                                             |
    | :----------------------------: | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    |         **-** and **+**        | Zoom out (-) or in (+).                                                                                                                                                 |
    |  <img alt="flow-centralize" /> | Centers the chart.                                                                                                                                                      |
    |   <img alt="search nodes" />   | Search for Nodes in the Flow.                                                                                                                                           |
    | <img alt="save flow as PNG" /> | Take a screenshot of the entire Flow and save it in PNG format. In the top left corner of each screenshot, the time of its creation and the Cognigy logo are displayed. |
    |   <img alt="voice-preview" />  | Test voice outputs without executing the entire Flow.                                                                                                                   |
    |       <img alt="undo" />       | Undo the last step by going backward in history.                                                                                                                        |
    |       <img alt="redo" />       | Redo the last steps by going forward in history.                                                                                                                        |
  </Accordion>
</AccordionGroup>

## More Information

* [Flows](/ai/agents/develop/projects-and-flows/overview)
* [Nodes](/ai/agents/develop/nodes/overview)
* [Interaction Panel](/ai/agents/test/interaction-panel/overview)
* [Agents](/ai/overview/ai-workforce)


# Flows
Source: https://docs.cognigy.com/ai/agents/develop/projects-and-flows/overview



<a href="/release-notes/4.80"><Badge>Updated in 4.80</Badge></a>

A *Flow* is a conversational structure consisting of [Nodes](/ai/agents/develop/nodes/overview) that defines how AI Agents interact with users. With Flows, you can customize the behavior of AI Agents and design tailored and dynamic conversational experiences. AI Agents can access the [Input](/ai/agents/develop/ai-agent-memory/input), [Context](/ai/agents/develop/ai-agent-memory/context), and [Profile](/ai/agents/develop/ai-agent-memory/profile) objects from the Flow to provide context-aware responses.

Flows include:

* [Nodes](/ai/agents/develop/nodes/overview) — the Flow's building blocks.
* [Intents](/ai/platform-features/nlu/intents/overview) — user input mapping with natural language understanding (NLU).
* [Lexicons](/ai/platform-features/nlu/slots/user-defined/lexicon) — keyphrases to improve Intent recognition.
* [Locales](/ai/platform-features/translation-and-localization/localization) — language settings for a multilingual AI Agent.
* [Slot Fillers](/ai/platform-features/nlu/slot-fillers) — automatic information extraction from user input.

You can also connect different [Flows](/ai/platform-features/nlu/attachments) and switch between them with a [Go To Node](/ai/agents/develop/node-reference/logic/go-to) to create more complex conversational experiences.

## Working with Flows

<Tabs>
  <Tab title="GUI">
    You can create, clone, and delete Flows in **Build > Flows**. Also, you can export Flows as a Package, copy their Reference ID. To build and configure Flows, use the [Flow editor](/ai/agents/develop/projects-and-flows/editor).

    You can create Flows using [Generative AI](/ai/agents/develop/gen-ai-and-llms/generative-ai) to automatically generate Nodes with predefined configurations and content. For this approach, you need to add a [large language model](/ai/agents/develop/gen-ai-and-llms/generative-ai#prerequisites) (LLM).

    To test Flows, click <img alt="interaction panel" /> and use the [Interaction Panel](/ai/agents/test/interaction-panel/overview).
  </Tab>

  <Tab title="API">
    You can view, create, and edit Flows using the [Cognigy API](https://api-trial.cognigy.ai/openapi#tag--Charts-v2.0).
  </Tab>

  <Tab title="CLI">
    You can manage Flows using the [Cognigy CLI](https://github.com/Cognigy/Cognigy-CLI).
  </Tab>
</Tabs>

### Settings

Besides the Nodes, you can customize Flows further on the **Settings** tab of the [Flow editor](/ai/agents/develop/projects-and-flows/editor). For example, you can add Yes/No logic and personalized confirmation responses. Also, you can fine-tune the AI Agent's confidence in understanding Intents and activate recognition of specific details, including age, date, and email addresses, to ensure more accurate responses.

<AccordionGroup>
  <Accordion title="General Flow Logic">
    | Parameter                                 | Description                                                                                                                                                                                                                                                                  |
    | ----------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Yes/No logic                              | Add a specific [Yes/No logic](/ai/platform-features/nlu/intents/yes-no-intents) within your Flow to confirm and reject an [Intent](/ai/platform-features/nlu/intents/overview) or the **Yes/No** type in [Question Nodes](/ai/agents/develop/node-reference/basic/question). |
    | Positive confirmation words               | Add customized responses to confirm an [Intent](/ai/platform-features/nlu/intents/overview) and [Question Nodes](/ai/agents/develop/node-reference/basic/question).                                                                                                          |
    | Negative confirmation words               | Add customized responses to reject an [Intent](/ai/platform-features/nlu/intents/overview) and [Question Nodes](/ai/agents/develop/node-reference/basic/question).                                                                                                           |
    | Continue main Flow after attached Flow    | Enable this setting to continue the main Flow after an attached Flow has been executed. Disabled by default.                                                                                                                                                                 |
    | Continue Flow after Default reply         | Enable this setting to continue executing the Flow logic after a default reply has been delivered. Disabled by default.                                                                                                                                                      |
    | Continue Flow after negative confirmation | Enable this setting to continue the Flow logic after a negative response to an Intent confirmation sentence. Disabled by default.                                                                                                                                            |
    | Pass Default Reply into Flow              | Enable this setting to have the AI Agent store the default reply in the input instead of using it directly. This setting allows you to use the reply within the Flow logic. Disabled by default.                                                                             |
    | Include Default Reply in NLU Training     | Enable this setting to include default replies in NLU training. The default replies serve as example sentences for matching Intents. You can also configure this setting individually for each Intent. Enabled by default.                                                   |
  </Accordion>

  <Accordion title="Intent Mapper">
    | Parameter                             | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
    | ------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Attached Flow Intent Mapping Priority | <ul><li>**Jointly Map Main and Attached Flow** — treats Main and Attached Flow Intents as if they were a single Intent collection, finding the best matching Intent and executing the corresponding Flow.</li><li>**Map Main Flow first** — maps the Main Flow Intents separately. If no Intent is found in the Main Flow, then the Attached Flow is mapped.</li><li>**Map Attached Flow first** — maps the Attached Flow Intents separately. If no Intent is found in the Attached Flow, then the Main Flow is mapped.</li></ul>                                                                                          |
    | Implicit Slot Parsing                 | Determine whether Slots in NLU example sentences should be parsed implicitly or if only Slot annotations should be used for training:<ul><li>**Disabled — only use slot annotations** — use only Slot annotations.</li><li>**Full - Parse both System and Lexicon slots implicitly** — parse both system- and user-defined Slots implicitly.</li><li>**System Slots - No Lexicon Slots** — parse only system-defined Slots implicitly. All Slot annotations will be considered.</li><li>**Lexicon Slots - No System Slots** — Parse only user-defined Slots implicitly. All Slot annotations will be considered.</li></ul> |
  </Accordion>

  <Accordion title="Thresholds">
    | Parameter                | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
    | ------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Intent threshold         | Determine the level of confidence required for the AI Agent to understand what the user is trying to communicate. The level of confidence accepts values ranging from 0 to 1.                                                                                                                                                                                                                                                                                                              |
    | Reconfirmation Threshold | Determine a score from which an Intent is considered confirmed or marked for reconfirmation if a reconfirmation sentence is set. By default, the value is `0.2`. Note that the confidence threshold has **no effect unless the Intent uses confirmation sentences**. The Reconfirmation Threshold is the lower confidence threshold. You must set it in addition to the Confidence Threshold. Intent scores above the reconfirmation threshold are confirmed or marked for reconfirmation. |
    | Confidence Threshold     | Determine a score from which an Intent is considered confirmed if a confirmation sentence is set. By default, the value is `0.4`.                                                                                                                                                                                                                                                                                                                                                          |
  </Accordion>

  <Accordion title="Attached Flows specific settings">
    | Parameter                                                             | Description                                                                                                       |
    | --------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------- |
    | Use Thresholds of the attached Flows                                  | Determine whether the thresholds configured in the attached Flows are applied. Enabled by default.                |
    | Use "Continue Flow after Default Reply" setting of the attached Flows | Determine if the conversation continues after giving a default reply in the attached Flows. Enabled by default.   |
    | Use "Pass Default Reply into Flow" of the attached Flows              | Determine if the default reply from attached Flows affects the ongoing conversation. Enabled by default.          |
    | Use "Implicit Slot Parsing" setting of the attached flows             | Determine if the system should try to understand Slot values even if not directly mentioned. Disabled by default. |
  </Accordion>

  <Accordion title="Lexicon Slots">
    | Parameter                           | Description                                                                                                                                                                                                                                                                                                                                             |
    | ----------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Parse Lexicon Slots with submatches | Enable the parsing of Lexicon Slots with submatches. This setting allows more detailed understanding of user input by identifying not just the primary match but also submatches within the Lexicon. This feature can improve accuracy in language understanding tasks, especially when dealing with complex or ambiguous phrases. Disabled by default. |
  </Accordion>

  <Accordion title="System Slots">
    | Parameter                | Description                                                                                                                                          |
    | ------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Enable Age Slots         | Enable the recognition of age-related terms. For more information, read [System-Defined Slots](/ai/platform-features/nlu/slots/system-defined).      |
    | Enable Date Slots        | Enable the recognition of date-related terms. For more information, read [System-Defined Slots](/ai/platform-features/nlu/slots/system-defined).     |
    | Enable Distance Slots    | Enable the recognition of distance-related terms. For more information, read [System-Defined Slots](/ai/platform-features/nlu/slots/system-defined). |
    | Enable Duration Slots    | Enable the recognition of duration-related terms. For more information, read [System-Defined Slots](/ai/platform-features/nlu/slots/system-defined). |
    | Enable Email Slots       | Enable the recognition of email addresses. For more information, read [System-Defined Slots](/ai/platform-features/nlu/slots/system-defined).        |
    | Enable Money Slots       | Enable the recognition of monetary values. For more information, read [System-Defined Slots](/ai/platform-features/nlu/slots/system-defined).        |
    | Enable Number Slots      | Enable the recognition of numerical values. For more information, read [System-Defined Slots](/ai/platform-features/nlu/slots/system-defined).       |
    | Enable Percentage Slots  | Enable the recognition of percentage values. For more information, read [System-Defined Slots](/ai/platform-features/nlu/slots/system-defined).      |
    | Enable Temperature Slots | Enable the recognition of temperature values. For more information, read [System-Defined Slots](/ai/platform-features/nlu/slots/system-defined).     |
    | Enable URL Slots         | Enable the recognition of URLs. For more information, read [System-Defined Slots](/ai/platform-features/nlu/slots/system-defined).                   |
  </Accordion>
</AccordionGroup>

## More Information

* [Flow editor](/ai/agents/develop/projects-and-flows/editor)
* [Nodes](/ai/agents/develop/nodes/overview)
* [Interaction Panel](/ai/agents/test/interaction-panel/overview)
* [Agents](/ai/overview/ai-workforce)


# Projects
Source: https://docs.cognigy.com/ai/agents/develop/projects-and-flows/projects



<a href="/release-notes/2026.12"><Badge>Updated in 2026.12</Badge></a>

*Projects* are collections of Conversational AI resources to build and deploy [AI Agents](/ai/agents/overview). You can design AI Agents to tackle one specific or cover a few use cases and handle multiple user intents. You can deploy AI Agents on a wide range of channels within a Project. Multiple team members can work on a Project at the same time.

Projects include:

* [AI Agents](/ai/agents/overview)
* [Flows](/ai/agents/develop/projects-and-flows/overview)
* [Endpoints](/ai/agents/deploy/endpoints/overview)
* [Connections](/ai/platform-features/connections)

## Limitations

* The maximum number of invitations to Projects per user per day. For example, if the limit is set to 5, each user can send invitations to at most 5 different Projects in a single day. Any additional invitation attempts will be blocked until the next day. If you have an on-premises installation, you can change this limit by setting the `MAX_ADD_PROJECT_MEMBER_REQUESTS_PER_EMAIL_PER_DAY` environment variable.

## Working with Projects

<Tabs>
  <Tab title="GUI">
    To create a Project with the [Project Creation wizard](#project-creation-wizard), click the **+** button in the upper-left corner of the [Main page](/ai/overview/user-interface/main-page). On the **Main** page, you can also edit or delete a Project as well as copy a Project's Reference ID.
  </Tab>

  <Tab title="API">
    You can view, create, and manage Projects using the [Cognigy API](https://api-trial.cognigy.ai/openapi#get-/v2.0/projects).
  </Tab>

  <Tab title="CLI">
    You can access Project resources, for example, AI Agents, Flows, and Endpoints, using the [Cognigy CLI](https://github.com/Cognigy/Cognigy-CLI?tab=readme-ov-file#cognigy-command-line-interface-cli).
  </Tab>
</Tabs>

### Project Creation Wizard

In the Project Creation wizard, select a Project type based on the AI Agent you want to build:

<Tabs>
  <Tab title="Agentic AI">
    An autonomous AI Agent driven by large language models (LLMs). You can build an [Agentic AI Agent](/ai/agents/quickstart) or hire one from the [Job Market](/ai/agents/overview#job-market).
  </Tab>

  <Tab title="Intent-driven AI">
    An AI Agent that uses [Intents](/ai/platform-features/nlu/intents/overview) to understand user input. You can build an [AI Agent](/ai/agents/quickstart) and integrate it with [Knowledge AI](/ai/agents/develop/knowledge-ai/overview), a RAG-based solution that provides context-aware responses.

    After creating a Project, you can customize your AI Agent to be Intent-driven or Agentic according to your needs.
  </Tab>
</Tabs>

## Task Panel

*Tasks* are specific activities, such as training Intents, importing Snapshots, or creating a new Project. A Task is an asynchronous process that runs in the background and for which you get notifications. The Task Panel provides an overview of ongoing and completed Tasks. To open the Task Panel, click the <img alt="task-menu" /> button in the upper-right corner of the **Project** page.

<Accordion title="View Tasks">
  Note that only the most recent 100 tasks are visible on the Task Panel. To view tasks beyond the latest 100, use the [Get Tasks](https://api-trial.cognigy.ai/openapi#get-/v2.0/tasks) API request as follows:

  * To skip the first 100 Tasks that are already visible on the Task Panel, set the `skip` parameter to 100.
  * To view, for example, the next 100 Tasks, set the `limit` parameter to 100.
</Accordion>

## More Information

* [User Menu](/ai/administer/user-menu/overview)
* [Interaction Panel](/ai/agents/test/interaction-panel/overview)
* [Insights](/insights/overview)
* [Analytics](/ai/agents/analyze/overview)
* [Agentic AI Agent](/ai/agents/overview)
* [AI Workforce](/ai/overview/ai-workforce)


# Getting Started with Voice AI Agents
Source: https://docs.cognigy.com/ai/agents/develop/voice



In this tutorial, you will learn how to create a voice AI Agent that uses dial logic. Dial logic lets users select options by pressing the corresponding numbers on a dial pad. This way, the voice AI Agent can guide users through predefined paths based on users' choices.

## Prerequisites

* You need to have access to a Cognigy account.
* One of the following speech provider credentials is required for testing a voice AI Agent: Microsoft Azure Speech Services, Google Speech Services, or Amazon Polly.

## Add Speech Provider Credentials

To add credentials for a speech provider, follow these steps:

<AiVoiceProviders />

After adding credentials for the speech provider, you can create your first voice agent.

## Build a Voice AI Agent

To build a voice AI Agent, follow these steps:

<Steps>
  <Step title="Create a Flow">
    1. In the left-side menu of the Project, go to **Build > Flows**.
    2. Create a new Flow by clicking **+ New Flow** in the upper-right corner.
    3. In the **New Flow** window, do the following:
       * **Name** — add a unique name, for example, `Getting Started`.
       * **Description** — add a relevant description, for example, `Getting Started Guide for voice agent`.
    4. Click **Create**.
  </Step>

  <Step title="Add a Set Session Config Node">
    1. In the Flow editor, click **+** and select **Extensions > VG**.
    2. Create a new Node by selecting **Set Session Config** from the list. The Node appears in the Flow editor.
    3. Left-click the **Set Session Config** Node to open the Node editor.
    4. In the Node editor, go to the **DTMF** section.
    5. In the DTMF section, activate the **Capture DTMF Signals** setting and remove `#` from the **DTMF Submit Digit** field.
    6. Click **Save** Node.
  </Step>

  <Step title="Add a Question Node">
    1. Below the Set Session Config Node, add a Question Node.
    2. In the Question Node editor, configure the following settings:
       * From the **Question Type** list, select **Number**.
       * From the **Output Type** list, select **Text**.
       * In the **Text** field, enter the following sentences:`Welcome to our customer service line. To learn about payment options, press 1. To contact technical support, press 2.`
    3. Click **Save Node**.
  </Step>

  <Step title="Add a Lookup Node">
    1. Below the Question Node, add a Lookup Node.
    2. In the Lookup Node editor, select the **Text** type and click **Save Node**.
  </Step>

  <Step title="Add Case Nodes">
    1. Click the **Case** Node. In the **Value** field, specify `1`, then click **Save Node**.
    2. Below the **Case** Node, add a Say Node.
    3. In the Say Node editor, go to the Text field and in specify the following sentences: `We accept the following payment methods:Credit Card (Visa, Mastercard, American Express), PayPal for secure online payments, or Bank Transfer for direct transfers from your bank account. Thank you for contacting us! Have a great day! Goodbye.`
    4. Click **Save Node**.
    5. Click the second **Case** Node. In the **Value** field, specify `2`.
    6. Below the second **Case** Node, add a Say Node.
    7. In the Say Node editor, go to the Text field and in specify the following sentences: `Apologies, all support members are busy. Try to contact us later. Thank you for your patience.`
    8. Click **Save Node**.
  </Step>

  <Step title="Add a Hang Up Node">
    1. At the end of the Flow, add a Hang Up Node.
    2. Click **Save Node**.
  </Step>
</Steps>

Once the voice agent is created, you can test it using the Interaction Panel.

## Test your Voice AI Agent

To test your voice AI Agent, follow these steps:

1. In the existing Flow, open the Interaction Panel by clicking <img alt="interaction-panel" /> **Chat with your Agent** in the upper-right corner of the page. The Interaction panel appears on the right side of the Flow Editor.
2. Initiate a call by clicking <img alt="phone" /> **> Start Call** at the bottom of the Interaction Panel.
3. Listen to the message from the voice agent and then select either number 1 or 2 on the dial pad <img alt="dialpad" /> located in the bottom-right corner of the Interaction Panel.
4. The voice agent will respond with a voice message according to the number you selected, and then finish the conversation. If you want to listen to a specific voice output again, hover your cursor over the message in the Interaction Panel and click **Voice Preview**.

<Frame>
  <img alt="Voice AI Agent conversation in the Interaction Panel with dial pad" />
</Frame>

To deploy your Flow in production, you will need to create a Voice Gateway Endpoint and set up a SIP Trunk provider.
For these steps, refer to the [Voice Gateway](/voice-gateway/overview) documentation.

## What's Next?

* Familiarize yourself with the core concepts of building AI Agents, such as [Projects](/ai/agents/develop/projects-and-flows/projects), [Flows](/ai/agents/develop/projects-and-flows/overview), [Nodes](/ai/agents/develop/node-reference/overview), and [CognigyScript](/ai/platform-features/cognigyscript).
* Enhance your AI Agent's performance with AI features, such as [NLU](/ai/platform-features/nlu/overview), [LLMs](/ai/agents/develop/gen-ai-and-llms/llms) and [Generative AI](/ai/agents/develop/gen-ai-and-llms/generative-ai), as well as [Knowledge AI](/ai/agents/develop/knowledge-ai/overview).
* Explore [test](/ai/agents/test/interaction-panel/overview) and [deploy](/ai/agents/deploy/endpoints/overview) features.
* Connect your AI Agent to the [contact center](/ai/escalate/handovers).


# Agentic AI
Source: https://docs.cognigy.com/ai/agents/overview

Build autonomous, goal-oriented AI Agents that use humanlike reasoning and tools to deliver superior outcomes.

<a href="/release-notes/2026.13"><Badge>Updated in 2026.13</Badge></a>

*Agentic AI* represents a new generation of autonomous, goal-oriented AI Agents designed to boost operational efficiency, enhance customer satisfaction, and drive superior enterprise outcomes.

Key features of Agentic AI:

* **Humanlike Reasoning**. Independently analyzes requests to determine the best course of action.
* **Autonomous Decision-Making**. Dynamically navigates and utilizes tools to accomplish tasks successfully.
* **Collaborative AI**. Works seamlessly with other AI Agents and humans to deliver efficient and accurate solutions.
* **Hyper-Personalization**. Built-in memory enables context-aware interactions, creating uniquely tailored experiences for each customer.

## Prerequisites

Before using the Agentic AI feature, you need:

<AccordionGroup>
  <Accordion title="1. Create an account in one of the LLM Providers">
    * [OpenAI](https://platform.openai.com/). You need to have a paid account or be a member of an organization that provides you access. Open your OpenAI user profile, copy the existing API Key, or create a new one and copy it. You can also use an [OpenAI-compatible LLM provider](/ai/agents/develop/gen-ai-and-llms/providers/openai-compatible) that supports the OpenAI API.
    * [Azure OpenAI](https://azure.microsoft.com/en-us/products/cognitive-services/openai-service). You need to have a paid account or be a member of an organization that provides you access. Ask your Azure administrator to provide API Key, resource name, and deployment model name.
    * [Anthropic Claude](https://console.anthropic.com/docs). You need to be a [member of an organization](https://console.anthropic.com/docs/) that provides you access.
    * [Google Gemini](https://ai.google.dev/gemini-api/docs/models/gemini). You need to have a paid account or be a member of an organization that provides you access.
    * [Amazon Bedrock](https://docs.aws.amazon.com/bedrock/). You need to have a paid account or be a member of an organization that provides you access.
    * [Mistral AI](https://docs.mistral.ai/). You need to have an account or be a member of an organization that provides you access.
  </Accordion>

  <Accordion title="2. Add two models">
    * A primary model to control the AI Agent, selected from the following providers:
      * `gpt-5-4`, `gpt-5-4-mini`, `gpt-5-4-nano`, `gpt-5-2`, `gpt-5-1`, `gpt-5` series models, `gpt-4.1`, `gpt-4.1-mini`, `gpt-4.1-nano`, `gpt-4o`, or `gpt-4o-mini` powered by [OpenAI](/ai/agents/develop/gen-ai-and-llms/providers/openai) or [Azure OpenAI](/ai/agents/develop/gen-ai-and-llms/providers/microsoft-azure-openai). You can also use an OpenAI-compatible model.
      * `claude-sonnet-4-6`, `claude-sonnet-4-5`, or `claude-haiku-4-5` by [Anthropic](/ai/agents/develop/gen-ai-and-llms/providers/anthropic).
      * `gemini-2.5-pro` and `gemini-2.5-flash` by [Google Gemini](/ai/agents/develop/gen-ai-and-llms/providers/google-gemini).
      * `gemini-3.1-pro-preview` and `gemini-3-flash-preview` by [Google GenAI](/ai/agents/develop/gen-ai-and-llms/providers/google-genai).
      * `amazon.nova-lite-v1:0`, `amazon.nova-pro-v1:0`, or `anthropic.claude-3-5-sonnet-20240620-v1:0` by [Amazon Bedrock](/ai/agents/develop/gen-ai-and-llms/providers/amazon-bedrock). Note that the model will only work if your AWS administrator gives you [access to this model](https://docs.aws.amazon.com/bedrock/latest/userguide/model-access.html).
      * `mistral-large-latest` or `mistral-small-latest` by [Mistral AI](/ai/agents/develop/gen-ai-and-llms/providers/mistral).
    * If you want to enable the [Knowledge AI](/ai/agents/develop/knowledge-ai/overview) feature for your AI Agent, make sure to add an embedding model along with the primary model. Cognigy recommends using the `text-embedding-ada-002` model to enhance Agentic AI capabilities.
  </Accordion>
</AccordionGroup>

## Concepts

* [AI Agent Management](#ai-agent-management)
* [AI Agents — a New Conversational Paradigm](#ai-agents--a-new-conversational-paradigm)

### AI Agent Management

*AI Agent Management* refers to the workspace used to control AI Agents within a Project.
This workspace includes the creation, modification, and deletion of AI Agents.
With AI Agent Management, you can effectively build autonomous AI Agents for specific roles, tasks, and workflows.

#### AI Agent Creation

You can create AI Agents tailored to meet specific business needs.
Designing these AI Agents allows you to handle a wide range of tasks, such as providing customer support, driving sales, automating data analysis, and more.

#### Job Market

You can hire an AI Agent with a ready-made set of skills for a specific task and adapt it to your needs.

The table shows available example AI Agents for hire.

| AI Agent                                               | Job              | Job Description                                                                                                               | Job Package                                                                                                                                                                |
| ------------------------------------------------------ | ---------------- | ----------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| <img alt="John, Airline Support AI Agent avatar" />    | Airline Support  | The AI Agent that supports tasks such as flight status updates, rebooking, passenger FAQs, and more.                          | - Flow: Main <br />- Knowledge Stores: ACME Airline Brochure and ACME Airline FAQs                                                                                         |
| <img alt="Anna, BDR Assistance AI Agent avatar" />     | BDR Assistance   | The AI Agent that assists with business development tasks such as answering questions, scheduling callbacks, and other tasks. | - Flow: Main <br />- Knowledge Stores: ACME Brochure and ACME CX FAQs                                                                                                      |
| <img alt="Sophie, Restaurant Guide AI Agent avatar" /> | Restaurant Guide | The AI Agent that provides information on menu items, opening hours, reservations, and more.                                  | - Flows: Main - Dining Concierge, Place Order - Order Expert, and Reserve Table - Reservation Expert <br />- Knowledge Stores: ACME Cantina Brochure and ACME Cantina Menu |
| <img alt="Martin, Health Insurance AI Agent avatar" /> | Health Insurance | The AI Agent that assists with tasks such as adjusting policies and managing claims.                                          | - Flow: Main - Health Insurance <br />- Knowledge Store: FitDirekt FAQs                                                                                                    |

### AI Agents — a New Conversational Paradigm

AI Agents autonomously perform many tasks that in previous conversational AI deployments had to be manually developed, such as intent detection and management, decision-making, entity extraction, and more. All of these tasks are now configured as part of the AI Agent.

#### AI Agent Node

The [AI Agent Node](/ai/agents/develop/node-reference/ai/ai-agent) refers to the use of an AI Agent in the context of a specific job. The AI Agent Node references the AI Agent persona and defines the job name, job description, and job-related instructions and context.

##### Job Instructions

Within the AI Agent Node, you must give clear instructions that guide the AI Agents in performing specific tasks.
For example, you could include instructions for executing an API call, processing data for analysis, performing a sentiment analysis, triggering automated emails, or classifying user queries.

##### Tools

*Tool Actions* represent specific tasks that an AI Agent can perform. A [tool action](/ai/agents/develop/node-reference/ai/ai-agent#child-nodes) allows the AI Agent to perform a variety of tasks, for example, interact with external services such as an API or to execute several rules-based actions. The full power of Cognigy can be used to configure what happens when a tool action is called.

Along with tool actions, you can use [Model Context Protocol](/ai/agents/develop/node-reference/ai/ai-agent#mcp-tool) (MCP), which boosts tool actions by providing a standardized connection between AI Agents and external services. It allows smooth integration with remote systems like APIs and databases, enabling the AI Agent to perform more complex tasks in a secure and efficient way.

##### Tool Calls

*Tool Calls* occur when the AI Agent decides to use a tool action and executes the branch underneath the Tool Action Node.

#### Resolve Tool Action Node

The [Resolve Tool Action Node](/ai/agents/develop/node-reference/ai/resolve-tool-action) sends an answer as the response to a tool call to the AI Agent after completing a [tool's action](#tools).

When the AI Agent calls a tool to perform an action, the tool action is executed.
Once the action is completed, the Resolve Tool Action Node sends a response back to the AI Agent,
indicating the result of that action.
This approach helps the AI Agent process and make decisions based on the results of the tools it initiated.

#### AI Agent Handover Node

The [AI Agent Handover Node](/ai/agents/develop/node-reference/ai/ai-agent-handover) enables smooth AI Agent-to-AI Agent handovers, allowing multiple AI Agents to collaborate and take over tasks to efficiently resolve complex customer inquiries.

#### LLM Prompt Node

The [LLM Prompt Node](/ai/agents/develop/node-reference/service/llm-prompt) is an alternative to the AI Agent Node. The updated version of LLM Prompt Node adds support for tools, image generation, and conversation transcripts. Use it with the Load AI Agent Node to fully replicate AI Agent Node behavior.

#### Load AI Agent Node

The [Load AI Agent Node](/ai/agents/develop/node-reference/ai/load-ai-agent) lets you load AI Agent data, such as instructions, name, and knowledge, into the Context or Input object.
This Node allows you to reuse the AI Agent's setup in other Nodes without duplicating configuration steps.

#### Transcript Nodes

AI Agents use the conversation transcript to process conversations, allowing it to reference past interactions for improved decision-making and providing more accurate responses in conversations.

##### Get Transcript Node

The [Get Transcript Node](/ai/agents/develop/node-reference/service/get-transcript) lets you receive a real-time conversation transcript and store the result in the Input or Context object.

##### Add Transcript Step

The [Add Transcript Step Node](/ai/agents/develop/node-reference/service/add-transcript-step) lets you add a step for one of the conversation actors to simulate an input or output.

## More Information

* [Getting Started](/ai/agents/quickstart)
* [Manage AI Agents](/ai/agents/develop/manage-ai-agents)
* [LLMs](/ai/agents/develop/gen-ai-and-llms/llms)
* [Knowledge AI](/ai/agents/develop/knowledge-ai/overview)


# Getting Started with AI Agents
Source: https://docs.cognigy.com/ai/agents/quickstart

Follow this step-by-step guide to create your first AI Agent with persona, tools, and knowledge in Cognigy.

<Frame>
  <img alt="Build your AI Agent workforce with Cognigy" />
</Frame>

## What are Agentic AI Agents?

An [Agentic AI Agent](/ai/agents/overview) is advanced conversational intelligence powered by a Large Language Model
(LLM),
designed to operate autonomously while adapting dynamically in real time to context, persona, and instructions.
Unlike traditional intent- or rule-based systems,
these AI Agents go beyond scripted responses — demonstrating flexibility,
problem-solving capabilities, and proactive decision-making.

In this guide, you will create a Logistics AI Agent to answer general questions and assist with parcel inquiries.

<Steps>
  <Step title="Set Up Prerequisites">
    Configure the necessary LLMs for language and embeddings, for example, OpenAI GPT-4o and ada-002, to enable your AI Agent.
  </Step>

  <Step title="Create Your AI Agent Persona">
    Go to **Build > AI Agents**, set the name, for example, `Peter Parcel`, description, avatar, and speaking style. Optionally, assign knowledge and configure security. Click **Create** to test.

    <Frame>
      <video title="Creating a new AI Agent persona in Cognigy" aria-label="The video shows a step-by-step configuration of an AI Agent persona." />
    </Frame>
  </Step>

  <Step title="Assign a Job">
    On the **Jobs** tab, define the job title, for example, `Delivery Manager`, description, and instructions.

    <Frame>
      <img alt="AI Agent job configuration with title, description, and instructions" />
    </Frame>
  </Step>

  <Step title="Enhance with Knowledge">
    Upload documents in Knowledge AI, then assign the knowledge to your AI Agent in its configuration to improve company-specific responses.
  </Step>

  <Step title="Add Tools">
    In the **Tools** section, click **Add Tool > Regular Tool**. Set Tool ID (for example, `track_parcel`) and parameters (for example, `tracking_code`).

    <Frame>
      <img alt="AI Agent tool configuration with Tool ID and parameters" />
    </Frame>
  </Step>

  <Step title="Configure Memory">
    Configure memory settings to personalize interactions and retain user details.
    To configure short-term memory, in the **Memory Handling** section, set **Long-Term Memory Injection** to **Inject full Contact Profile**.

    <Frame>
      <img alt="AI Agent memory handling settings with long-term memory injection" />
    </Frame>
  </Step>

  <Step title="Enable Image Recognition">
    In the **Image Handling** section, activate **Process Images**. Enable **Attachment Upload** in the Webchat v3 Endpoint settings. Ensure a storage provider like AWS or Azure is set.
  </Step>

  <Step title="Explore the Flow">
    Open **Advanced Editor** to view and edit all Nodes controlling AI Agent responses.

    <Frame>
      <img alt="Flow editor showing Nodes that control AI Agent responses" />
    </Frame>
  </Step>

  <Step title="Deploy Your AI Agent">
    [Cognigy Endpoints](/ai/agents/deploy/endpoint-reference/overview) allow AI Agents to interact with external users via chat and voice channels.

    Deploy an AI Agent via Webchat:

    1. Go to **Deploy > Endpoints**.
    2. Add a new Endpoint and link it to the Flow.
    3. Click **Open Demo Webchat** to test the AI Agent.
  </Step>
</Steps>

## What's Next?

1. Try [creating AI Agents tailored to your business needs](/ai/agents/develop/manage-ai-agents#create-from-scratch) or [hire one from the Job Market](/ai/agents/develop/manage-ai-agents#hire-from-the-job-market).
2. [Add the AI Agent to your Flow](/ai/agents/develop/projects-and-flows/overview) and [create a set of tools](/ai/agents/develop/node-reference/ai/ai-agent) for your AI Agent to operate.
3. For a deeper dive into Agentic AI, visit the [Cognigy Help Center](https://support.cognigy.com/hc/en-us/articles/17346581840796-Getting-Started-with-AI-Agents).


# Assertions
Source: https://docs.cognigy.com/ai/agents/test/assertions



<a href="/release-notes/2025.19"><Badge>Updated in 2025.19</Badge></a>

*Assertions* are validation checks that verify the AI Agent's responses and different aspects of the Flow for each Step during a Playbook run.

## How to Use Assertions

Assertions include an input field where you can provide the text or data you want to validate during the Playbook run. Assertions are case-sensitive. The text, Intent names, Slot names, and data fields that you enter in the input field must match exactly what you want to validate. Specific assertion types include an option to validate only a portion of the provided text or data:

* **Use fuzzy search** — for text assertions.
* **Use partial comparison** — for data and context assertions.

The fuzzy search functionality handles case-insensitive matches, differences in spaces, tab stops, and newlines, and escaped characters.

<Accordion title="Fuzzy Search Examples">
  | Pattern    | Feature                   | Matches                                     |
  | ---------- | ------------------------- | ------------------------------------------- |
  | `My order` | **Case Insensitivity**    | `my order`, `MY ORDER`                      |
  |            | **Differences in spaces** | <pre><code>My   order</code></pre>          |
  |            | **Newlines**              | <pre><code>My<br /><br />order</code></pre> |
  |            | **Escaped Characters**    | `My\norder`, `My\torder`                    |
</Accordion>

You can also invert the assertion check with the **Invert assertion** option. This option switches the assertion check to validate the absence of the defined value. Inverted assertions must also match the exact text or data provided in the assertion. You can activate both **Invert assertion** and **Use fuzzy search** to validate the absence of a portion of the provided text or data, for example, only the `"status": "active"` JSON property or the word `"error"` in the AI Agent's response.

## Assertion Types

### Text Assertion

Validates text in the AI Agent's response against the expected text or pattern provided in the input field.

| Parameter        | Type        | Description                                                                                    |
| ---------------- | ----------- | ---------------------------------------------------------------------------------------------- |
| Invert assertion | Option      | Sets the assertion to validate the absence of the text or pattern provided in the input field. |
| Use fuzzy search | Option      | Allows partial matches in the AI Agent's response.                                             |
| Text             | Input field | Sets the exact text or pattern to validate.                                                    |

#### Examples

<AccordionGroup>
  <Accordion title="Exact Match">
    In this example, configure the assertion as follows:

    | Parameter        | Value                           |
    | ---------------- | ------------------------------- |
    | Invert assertion | Deactivated                     |
    | Use fuzzy search | Deactivated                     |
    | Text             | `Your order has been confirmed` |

    Validates that `Your order has been confirmed` is present in the AI Agent's response.
    Fails if `We have received your order` is in the AI Agent's response.
  </Accordion>

  <Accordion title="Fuzzy Search">
    In this example, configure the assertion as follows:

    | Parameter        | Value       |
    | ---------------- | ----------- |
    | Invert assertion | Deactivated |
    | Use fuzzy search | Activated   |
    | Text             | `"order"`   |

    Validates that `Your order has been confirmed` and `We have received your order` are in the AI Agent's response.
    Fails if `We are preparing your purchase` is in the AI Agent's response.
  </Accordion>

  <Accordion title="Inverted Assertion">
    In this example, configure the assertion as follows:

    | Parameter        | Value     |
    | ---------------- | --------- |
    | Invert assertion | Activated |
    | Use fuzzy search | Activated |
    | Text             | `"error"` |

    Validates that the AI Agent's response doesn't include `error`, for example, `Your order will be shipped shortly`.
    Fails if the AI Agent's response includes `error`, for example, `We found an error while processing your order`.
  </Accordion>
</AccordionGroup>

### Data Assertion

Validates the AI Agent message's data fields and JSON structures against the data provided in the input field. If the AI Agent's response doesn't include any data fields, the Assertion checks data fields in the user input.

| Parameter              | Type        | Description                                                                                                     |
| ---------------------- | ----------- | --------------------------------------------------------------------------------------------------------------- |
| Invert assertion       | Option      | Sets the assertion to validate the absence of the JSON structure provided in the input field or the user input. |
| Use partial comparison | Option      | Allows matches of a subset of the data structure to validate.                                                   |
| Data                   | Input field | Sets the data structure to validate in JSON format.                                                             |

#### Examples

<AccordionGroup>
  <Accordion title="Exact Match">
    In this example, configure the assertion as follows:

    <table>
      <thead>
        <tr>
          <th>Parameter</th>
          <th>Value</th>
        </tr>
      </thead>

      <tbody>
        <tr>
          <td>Invert assertion</td>
          <td>Deactivated</td>
        </tr>

        <tr>
          <td>Use partial comparison</td>
          <td>Deactivated</td>
        </tr>

        <tr>
          <td>Data</td>

          <td>
            ```json theme={null}
            {
            "userId": "12345",
            "status": "active",
            "timestamp": "2024-01-01T10:00:00Z"
            }
            ```
          </td>
        </tr>
      </tbody>
    </table>

    Validates only the exact JSON object provided in the input field.
  </Accordion>

  <Accordion title="Partial Comparison">
    In this example, configure the assertion as follows:

    <table>
      <thead>
        <tr>
          <th>Parameter</th>
          <th>Value</th>
        </tr>
      </thead>

      <tbody>
        <tr>
          <td>Invert assertion</td>
          <td>Deactivated</td>
        </tr>

        <tr>
          <td>Use partial comparison</td>
          <td>Activated</td>
        </tr>

        <tr>
          <td>Data</td>

          <td>
            ```json theme={null}
            {
            "userId": "12345",
            "status": "active"
            }
            ```
          </td>
        </tr>
      </tbody>
    </table>

    Validates any JSON structure that contains the JSON object provided in the input field, for example:

    ```json theme={null}
    {
    "userId": "12345", // part of the provided JSON object
    "status": "active", // part of the provided JSON object
    "timestamp": "2024-01-01T10:00:00Z" // not part of the provided JSON object
    }
    ```

    Fails if the JSON structure doesn't contain the JSON object provided in the input field, for example:

    ```json theme={null}
    {
    "userId": "12345", // subset of the provided JSON object
    "status": "failed", // not part of the provided JSON object
    "timestamp": "2024-01-01T10:00:00Z" // not part of the provided JSON object
    }
    ```
  </Accordion>

  <Accordion title="Inverted Assertion">
    In this example, configure the assertion as follows:

    <table>
      <thead>
        <tr>
          <th>Parameter</th>
          <th>Value</th>
        </tr>
      </thead>

      <tbody>
        <tr>
          <td>Invert assertion</td>
          <td>Activated</td>
        </tr>

        <tr>
          <td>Use partial comparison</td>
          <td>Activated</td>
        </tr>

        <tr>
          <td>Data</td>

          <td>
            ```json theme={null}
            {
            "status": "failed"
            }
            ```
          </td>
        </tr>
      </tbody>
    </table>

    Validates any JSON structure that doesn't contain the JSON object provided in the input field, for example:

    ```json theme={null}
    {
    "userId": "12345",
    "status": "active", // validates if status isn't failed
    "timestamp": "2024-01-01T10:00:00Z"
    }
    ```
  </Accordion>
</AccordionGroup>

### Intent Assertion

Validates that the [Intent](/ai/platform-features/nlu/intents/overview) provided in the input field has been recognized in the user input.

| Parameter        | Type        | Description                                                                           |
| ---------------- | ----------- | ------------------------------------------------------------------------------------- |
| Invert assertion | Option      | Sets the assertion to validate the absence of the Intent provided in the input field. |
| Intent           | Input field | Sets the Intent name to validate.                                                     |

#### Examples

<AccordionGroup>
  <Accordion title="Exact Match">
    In this example, configure the assertion as follows:

    | Parameter        | Value           |
    | ---------------- | --------------- |
    | Invert assertion | Deactivated     |
    | Intent           | `"greet.hello"` |

    Validates only that the `"greet.hello"` Intent has been recognized.
  </Accordion>

  <Accordion title="Inverted Match">
    In this example, configure the assertion as follows:

    | Parameter        | Value            |
    | ---------------- | ---------------- |
    | Invert assertion | Activated        |
    | Intent           | `"cancel.order"` |

    Validates that the `"cancel.order"` Intent hasn't been recognized.
  </Accordion>
</AccordionGroup>

### Slot Assertion

Validates that the [Slot](/ai/platform-features/nlu/slots/overview) provided in the input field has been recognized in the user input.

| Parameter        | Type        | Description                                                                         |
| ---------------- | ----------- | ----------------------------------------------------------------------------------- |
| Invert assertion | Option      | Sets the assertion to validate the absence of the Slot provided in the input field. |
| Slot             | Input field | Sets the Slot name to validate.                                                     |

#### Examples

<AccordionGroup>
  <Accordion title="Exact Match">
    In this example, configure the assertion as follows:

    | Parameter        | Value       |
    | ---------------- | ----------- |
    | Invert assertion | Deactivated |
    | Slot             | `username`  |

    Validates only that the `username` Slot has been recognized.
  </Accordion>

  <Accordion title="Inverted Assertion">
    In this example, configure the assertion as follows:

    | Parameter        | Value      |
    | ---------------- | ---------- |
    | Invert assertion | Activated  |
    | Slot             | `password` |

    Validates that the `password` Slot hasn't been recognized.
  </Accordion>
</AccordionGroup>

### Context Assertion

Validates the data in the [Context object](/ai/agents/develop/ai-agent-memory/context) against the data provided in the input field.

| Parameter              | Type        | Description                                                                                |
| ---------------------- | ----------- | ------------------------------------------------------------------------------------------ |
| Invert assertion       | Option      | Sets the assertion to validate the absence of the context data provided in the input field |
| Use partial comparison | Option      | Allows matches of a subset of the context data to validate.                                |
| Context data           | Input field | Sets the context data to validate in JSON format.                                          |

#### Examples

<AccordionGroup>
  <Accordion title="Exact Match">
    In this example, configure the assertion as follows:

    <table>
      <thead>
        <tr>
          <th>Parameter</th>
          <th>Value</th>
        </tr>
      </thead>

      <tbody>
        <tr>
          <td>Invert assertion</td>
          <td>Deactivated</td>
        </tr>

        <tr>
          <td>Use partial comparison</td>
          <td>Deactivated</td>
        </tr>

        <tr>
          <td>Context data</td>

          <td>
            ```json theme={null}
            {
            "authenticated": true,
            "userRole": "premium",
            "sessionId": "abc123"
            }
            ```
          </td>
        </tr>
      </tbody>
    </table>

    Validates only the exact JSON object provided in the input field.
  </Accordion>

  <Accordion title="Partial Match">
    In this example, configure the assertion as follows:

    <table>
      <thead>
        <tr>
          <th>Parameter</th>
          <th>Value</th>
        </tr>
      </thead>

      <tbody>
        <tr>
          <td>Invert assertion</td>
          <td>Deactivated</td>
        </tr>

        <tr>
          <td>Use partial comparison</td>
          <td>Activated</td>
        </tr>

        <tr>
          <td>Context data</td>

          <td>
            ```json theme={null}
            {
            "hasCreditCard": true,
            "activeCreditCard": true
            }
            ```
          </td>
        </tr>
      </tbody>
    </table>

    Validates any context data that contains the JSON object provided in the input field, for example:

    ```json theme={null}
    {
    "hasCreditCard": true, // part of the provided JSON object
    "activeCreditCard": true, // part of the provided JSON object
    "sessionId": "abc123" // not part of the provided JSON object
    }
    ```

    Fails if the JSON structure doesn't contain the JSON object provided in the input field, for example:

    ```json theme={null}
    {
    "hasCreditCard": true, // subset of the provided JSON object
    "activeCreditCard": false, // not part of the provided JSON object
    "sessionId": "abc123" // not part of the provided JSON object
    }
    ```
  </Accordion>

  <Accordion title="Inverted Assertion">
    In this example, configure the assertion as follows:

    <table>
      <thead>
        <tr>
          <th>Parameter</th>
          <th>Value</th>
        </tr>
      </thead>

      <tbody>
        <tr>
          <td>Invert assertion</td>
          <td>Activated</td>
        </tr>

        <tr>
          <td>Use partial comparison</td>
          <td>Activated</td>
        </tr>

        <tr>
          <td>Context data</td>

          <td>
            ```json theme={null}
            {
            "activeCreditCard": false
            }
            ```
          </td>
        </tr>
      </tbody>
    </table>

    Validates any JSON structure that doesn't contain the JSON object provided in the input field, for example:

    ```json theme={null}
    {
    "hasCreditCard": true,
    "activeCreditCard": true, // validates if activeCreditCard isn't false
    "creditCardNumber": "1234567890"
    }
    ```
  </Accordion>
</AccordionGroup>

### State Assertion

Validates the conversation state.

| Parameter        | Type        | Description                                                                          |
| ---------------- | ----------- | ------------------------------------------------------------------------------------ |
| Invert assertion | Option      | Sets the assertion to validate the absence of the State provided in the input field. |
| State            | Input field | Sets the State name to validate.                                                     |

#### Examples

<AccordionGroup>
  <Accordion title="Exact Match">
    In this example, configure the assertion as follows:

    | Parameter        | Value               |
    | ---------------- | ------------------- |
    | Invert assertion | Deactivated         |
    | State            | `waiting_for_input` |

    Validates only that the `waiting_for_input` State is active.
  </Accordion>

  <Accordion title="Inverted Assertion">
    In this example, configure the assertion as follows:

    | Parameter        | Value         |
    | ---------------- | ------------- |
    | Invert assertion | Activated     |
    | State            | `error_state` |

    Validates only that the `error_state` State isn't active.
  </Accordion>
</AccordionGroup>

## Best Practices

For testing best practices, see [Best Practices: Testing your Virtual Agents with Playbooks](https://support.cognigy.com/hc/en-us/articles/9585431937948-Best-Practices-Testing-your-Virtual-Agents-with-Playbooks#7-regression-test-nlu-models-0-6).

## More Information

* [Slots](/ai/platform-features/nlu/slots/overview)
* [Intents](/ai/platform-features/nlu/intents/overview)


# Demo Webchat
Source: https://docs.cognigy.com/ai/agents/test/demo-webchat



*Cognigy Demo Webchat* offers the opportunity to test Webchat integrations with various providers, including handover providers, and also allows you to preview the appearance of your webchat before deploying it in a production environment. This tool is valuable for testing and customization, ensuring that your Webchat functions according to your expectations and looks the way you want it to before launching.

For more information, read the [Demo Webchat](/webchat/demo) documentation.


# Chat
Source: https://docs.cognigy.com/ai/agents/test/interaction-panel/chat



<a href="/release-notes/4.84"><Badge>Updated in 4.84</Badge></a>

With *Chat* mode, you can chat with your AI Agent to test, [debug](#debug-mode), and improve it.

## How to Use Chat Mode

To open chat mode, click the <img alt="chat" /> tab in the [Interaction Panel](/ai/agents/test/interaction-panel/overview). Chat mode works as a common chat interface that simulates your [Flows](/ai/agents/develop/projects-and-flows/overview) based on the [Interaction Panel settings](/ai/agents/test/interaction-panel/overview). You can also click <img alt="record-audio" /> to activate your device's microphone and record voice input.

Additionally, you can use the [debug mode](#debug-mode) to analyze the [natural language understanding](/ai/platform-features/nlu/overview) (NLU) results, Flow transitions, and specific Node outputs.

### Debug Mode

Debug mode displays more information about the Flow execution than just user and AI Agent messages for better understanding of the conversation. Debug mode is activated by default. With this mode, you can check:

* Triggered [Intents](/ai/platform-features/nlu/intents/overview) and their [scores](/ai/platform-features/nlu/intents/intent-analyzer)
* Triggered [Yes/No Intents](/ai/platform-features/nlu/intents/yes-no-intents) and their [scores](/ai/platform-features/nlu/intents/intent-analyzer)
* Matched [Slots](/ai/platform-features/nlu/slots/overview)
* Found [Lexicon Keyphrases](/ai/platform-features/nlu/slots/user-defined/lexicon)
* Flow Transitions, for example, when executing a [Go To Node](/ai/agents/develop/node-reference/logic/go-to), an [Execute Flow Node](/ai/agents/develop/node-reference/logic/execute-flow), or triggering an [Attached Flow](/ai/platform-features/nlu/attachments)
* Node-specific output, for example, from an [HTTP Request Node](/ai/agents/develop/node-reference/service/http-request)

Some Nodes have settings that let you further configure their debug mode output. You can turn off node-specific outputs selectively in the [Interaction Panel settings](/ai/agents/test/interaction-panel/overview#configure-the-interaction-panel).

You can explicitly trigger debug messages either using the [Debug Message Node](/ai/agents/develop/node-reference/basic/debug-message) or [debug api functions](/ai/for-developers/code/api-functions) in the Code Node.

<Frame>
  <img alt="Interaction Panel Chat tab with debug mode enabled" />
</Frame>

## More Information

* [Interaction Panel](/ai/agents/test/interaction-panel/overview)
* [Intents](/ai/platform-features/nlu/intents/overview)
* [Slots](/ai/platform-features/nlu/slots/overview)
* [Lexicons](/ai/platform-features/nlu/slots/user-defined/lexicon)


# Live Follow
Source: https://docs.cognigy.com/ai/agents/test/interaction-panel/live-follow



With *Live Follow* mode, you can monitor conversations between AI Agents and users in real time to debug [Flows](/ai/agents/develop/projects-and-flows/overview) connected to an [Endpoint](/ai/agents/deploy/endpoints/overview).

## Prerequisites

* Configure an [Endpoint](/ai/agents/deploy/endpoints/overview) for the Flow.

## Restrictions

* Live Follow mode doesn't display human agent messages.
* Live Follow mode uses the [Endpoint](/ai/agents/deploy/endpoints/overview) from the monitored conversation and doesn't take the [Interaction Panel settings into account](/ai/agents/test/interaction-panel/overview#1-project-related-settings-and-action-menu).
* Live Follow mode doesn't support conversations carried out through an Endpoint that points to a [Snapshot](/ai/agents/deploy/snapshots).

## How to Use Live Follow Mode

To use and test Live Follow mode, you first need the user ID of the user you want to follow. To find the user ID, you have the following options:

* On the **Logs** page in **Test > Logs**, search for `userId` in logs with a `info Received message from user` log message.
* If you don't find `userId`, navigate to the chat you created via the Endpoint, for example, Webchat, Teams, or Slack. Send a message in this chat, then go back to the **Logs** page and search for `userId`.
* To get user ID for Demo Webchat, refer to the [Demo Webchat](/webchat/demo#manually-defining-the-user-id) documentation.

Once you have the user ID, follow these steps:

1. In the Interaction Panel, open live follow mode by clicking <img alt="live-follow" />.
2. Paste the user ID into the **User ID** field and click <img alt="start-live-follow" />.

You can now view and track the user's real-time conversation from the Interaction Panel. To stop live following, click <img alt="stop-live-follow" />. The following video showcases a voice use case in Live Follow mode:

<iframe title="Live Follow Mode" />

## More Information

* [Logs](/ai/agents/test/logs)


# Interaction Panel
Source: https://docs.cognigy.com/ai/agents/test/interaction-panel/overview



<a href="/release-notes/2026.1"><Badge>Updated in 2026.1</Badge></a>

The *Interaction Panel* offers a robust interface for testing chat and voice AI Agents. You can run different tests and analyze how AI Agents respond. This approach allows you to evaluate the AI Agents' performance and change their communication and service skills.

The Interaction Panel provides the following features:

* **Three Testing Modes**. Test your AI Agent across all Channels with the [chat](/ai/agents/test/interaction-panel/chat), [voice call](/ai/agents/test/interaction-panel/voice-call), and [live follow](/ai/agents/test/interaction-panel/live-follow) modes to simulate real-world interactions.
* **Memory Inspection and Extraction**. View and extract the data stored in the [AI Agent's memory](/ai/agents/develop/ai-agent-memory/overview) to analyze its behavior, as well as to identify potential areas for improvement.
* **Playbooks**. Use [Playbooks](/ai/agents/test/playbooks) to save time by automating repetitive tasks and catching errors efficiently.
* **Flexible Configuration**. Design custom test scenarios for in-depth analysis of your AI Agent's behavior.

## Prerequisites

* You must have a [Flow](/ai/agents/develop/projects-and-flows/overview) with configured [Nodes](/ai/agents/develop/nodes/overview) to test the AI Agent.

## Explore the Interaction Panel

To open the Interaction Panel, click <img alt="interaction-panel" /> in the upper-right corner. You can open the Interaction Panel from any page within your Project and set the Flow, [Locale](/ai/platform-features/translation-and-localization/localization), and [Snapshot](/ai/agents/deploy/snapshots) you want to test. However, when opened from the Flow Editor, the Interaction Panel uses the current Flow configuration and allows you to follow the Flow execution in real time.

<Frame>
  <img alt="Interaction Panel opened in the Cognigy interface" />
</Frame>

<AccordionGroup>
  <Accordion title="1. Project-Related Settings and Action Menu">
    | <span>UI Element</span> | Type        | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
    | ----------------------- | ----------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Snapshot                | List        | This setting is optional. If applicable, you can select the Snapshot of a Project to test your AI Agent.                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | Locale                  | List        | Define the language your AI Agent should use for testing.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
    | Flow                    | List        | Select the Flow you want to test with the AI Agent.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
    | <img alt="menu" />      | Action menu | Use this menu to: <ul><li>**Reset Session** — clear the chat history in the Interaction Panel.</li><li>**Create Playbook** — create a new Playbook for testing conversations. For more information, refer to [Playbooks](/ai/agents/test/playbooks).</li><li>**Create Playbook with Assertions** — create a new Playbook with [Assertions](/ai/agents/test/playbooks#working-with-playbooks) for in-depth testing.</li><li>**Enable/Disable Debug Mode** — toggle the [Debug mode](/ai/agents/test/interaction-panel/chat#debug-mode). Debug mode is enabled by default.</li></ul> |
  </Accordion>

  <Accordion title="2. Panel Tabs">
    | Tab      | Description                                                                                      |
    | -------- | ------------------------------------------------------------------------------------------------ |
    | Test     | Interact with your AI Agent.                                                                     |
    | Info     | Examine the data stored in the [AI Agent's memory](/ai/agents/develop/ai-agent-memory/overview). |
    | Settings | [Configure the Interaction Panel](#configure-the-interaction-panel).                             |
  </Accordion>

  <Accordion title="3. Mode Tabs">
    | Tab                       | Description                                                                                                   |
    | ------------------------- | ------------------------------------------------------------------------------------------------------------- |
    | <img alt="chat" />        | Test your chat AI Agent in the [Chat panel](/ai/agents/test/interaction-panel/chat).                          |
    | <img alt="voice-call" />  | Test your voice AI Agent in the [Voice Call panel](/ai/agents/test/interaction-panel/voice-call).             |
    | <img alt="live-follow" /> | Monitor your AI Agent in real time in the [Live Follow panel](/ai/agents/test/interaction-panel/live-follow). |
    | <img alt="simulation" />  | Run [simulated conversations](/ai/agents/test/interaction-panel/simulation) using LLMs.                       |
    | <img alt="playbook" />    | Test your AI Agent using a [Playbook](/ai/agents/test/playbooks).                                             |
  </Accordion>
</AccordionGroup>

## Configure the Interaction Panel

<Accordion title="Settings">
  | Setting                           | Type   | Description                                                                                                                                                                                                                                                                                                                                                                                                                  |
  | --------------------------------- | ------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
  | **General**                       |        |                                                                                                                                                                                                                                                                                                                                                                                                                              |
  | NLU Connector                     | List   | Allows you to select third-party NLU engines to use in the Interaction Panel. To install a new NLU engine in Cognigy, go to **Build > NLU Connectors** and click **+ New NLU Connector**. For more information on how to manage NLU connectors, refer to [NLU Connectors](/ai/platform-features/nlu/external/nlu-connectors/overview). The default is the Cognigy NLU engine.                                                |
  | Text to Speech                    | Toggle | Allows the AI Agent vocalizes the responses using your browser's built-in Text-to-Speech (TTS) functionality.                                                                                                                                                                                                                                                                                                                |
  | Channel Selection                 | List   | Select the Channels you want to display in the Interaction Panel, such as Webchat, Voice Gateway, Slack, and more. The **Show all** option is selected by default.                                                                                                                                                                                                                                                           |
  | **Debug Mode**                    |        |                                                                                                                                                                                                                                                                                                                                                                                                                              |
  | Debug mode                        | Toggle | Activates [Debug mode](/ai/agents/test/interaction-panel/chat#debug-mode).                                                                                                                                                                                                                                                                                                                                                   |
  | Node Filter                       | List   | Select the Node types to show in Debug mode.                                                                                                                                                                                                                                                                                                                                                                                 |
  | **Chat**                          |        |                                                                                                                                                                                                                                                                                                                                                                                                                              |
  | Show Data Input                   | Toggle | Displays the **Data Input** field below the standard text input field in the Interaction Panel, allowing you to enter JSON data with or without a text message. For example, `{"firstName": "Max", "lastName": "Müller"}`.                                                                                                                                                                                                   |
  | Render Markdown                   | Toggle | Enables Markdown rendering for AI Agent text outputs. Supports basic formatting like italics, bold, links, and headers, as well as advanced features such as tables, footnotes, and nested lists. Formulas aren't supported. This parameter is disabled by default and matches the behavior of the Webchat v3 [Render Markdown](/webchat/v3/configuration#webchat-behavior) option.                                          |
  | **Voice Call**                    |        |                                                                                                                                                                                                                                                                                                                                                                                                                              |
  | Language                          | List   | Select a language for the AI Agent. The supported languages depend on the [configured voice provider](/ai/agents/test/voice-preview). The **Custom** option lets you use a TTS language that isn't in the list. The format depends on the TTS Vendor, for example, `de-DE`, `fr-FR`, `en-US`.                                                                                                                                |
  | Voice                             | List   | Select a female or male speaker's voice. This setting applies to region-specific voices as well. The format depends on the TTS Vendor, for example, `de-DE-ConradNeural`.                                                                                                                                                                                                                                                    |
  | **Playbooks**                     |        |                                                                                                                                                                                                                                                                                                                                                                                                                              |
  | Repeat                            | Toggle | Allows Playbook runs repeatedly until you explicitly stop it. This feature is helpful for testing use cases that need consistent behavior over many interactions.                                                                                                                                                                                                                                                            |
  | Delay                             | Number | Set a waiting time between each Playbook Step execution in milliseconds (ms). The default value is 200 ms.                                                                                                                                                                                                                                                                                                                   |
  | **Advanced**                      |        |                                                                                                                                                                                                                                                                                                                                                                                                                              |
  | Enable Mocking                    | Toggle | Activates [mocking](/ai/agents/test/mocking) for chat and simulation mode in the Interaction Panel.                                                                                                                                                                                                                                                                                                                          |
  | Auto-Move Flow Editor             | Toggle | Allows the Flow editor view to automatically: <ul><li>Center on the last executed Node in the current Flow.</li><li>Open another Flow that is triggered during the Flow execution and centers on the triggered Node.</li></ul>The setting is activated by default.                                                                                                                                                           |
  | Auto-Switch Target Flow Selection | Toggle | Automatically switches between Flows in the Interaction Panel when another Flow is opened in the editor. The setting is disabled by default. Example: A project includes two Flows, Flow A and Flow B. The process starts according to Flow A and reaches a step that triggers Flow B. If this setting is activated, Flow B opens in the editor, and the Interaction Panel automatically switches to Flow B in its settings. |
</Accordion>

## More Information

* [Cognigy Script and Objects Lifespan](/ai/agents/develop/ai-agent-memory/overview#lifespan-of-cognigy-objects)
* [Testing your AI Agents with Playbooks](https://support.cognigy.com/hc/en-us/articles/9585431937948-Best-Practices-Testing-your-Virtual-Agents-with-Playbooks#1-design-comprehensive-playbooks-0-0)
* [Voice Gateway Parameter Details](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details)


# Simulation
Source: https://docs.cognigy.com/ai/agents/test/interaction-panel/simulation



<a href="/release-notes/2026.12"><Badge>Added in 2026.12</Badge></a>

With *Simulation* mode, you can simulate conversations for your LLM-based AI Agents using [predefined scenarios](/ai/agents/test/simulator/overview). This mode allows you to test and debug your AI Agent's behavior in various situations without needing real user input.

## How to Use Simulation Mode

To open simulation mode, click the <img alt="simulation" /> tab in the [Interaction Panel](/ai/agents/test/interaction-panel/overview), then select the scenario to run the simulation on.

The simulation of the selected scenario starts with the following predefined settings:

* The number of simulation runs is 1 and can't be changed.
* The simulation uses the [default LLM](/ai/agents/develop/gen-ai-and-llms/other-operations#set-a-model-as-default) selected on the **Large Language Models** page.

<div>
  <video title="Interaction Panel simulation mode demonstration" aria-label="The Interaction Panel showing exchanged messages during a simulation, then presenting the results.">
    <source />
  </video>
</div>

## More Information

* [Simulator](/ai/agents/test/simulator/overview)
* [Interaction Panel](/ai/agents/test/interaction-panel/overview)


# Voice Call
Source: https://docs.cognigy.com/ai/agents/test/interaction-panel/voice-call



With *Voice Call* mode, you can initiate a test call to hear how your voice AI Agent sounds.

## Prerequisites

* Activate voice call:
  * For SaaS installations, contact the support team to activate the feature.
  * For on-premises installations, activate the feature by adding `FEATURE_ENABLE_VOICECALL_WHITELIST` in `values.yaml` and assigning the corresponding organization IDs. For example, `FEATURE_ENABLE_VOICECALL_WHITELIST:<organization-1-id>,<organization-2-id>`. For all organizations, use `*`.
* Configure a voice provider in the Project's [Voice Preview Settings](/ai/agents/test/voice-preview#set-up-a-speech-provider) and the AI Agent's [voice and language](/ai/agents/test/interaction-panel/overview).
* Allow microphone access in your browser, and then in your device settings if needed.

## How to Use Voice Call Mode

To start a test call with the AI Agent, click the <img alt="voice-call" /> tab in the Interaction Panel and click <img alt="start-call" /> at the bottom. To end the call, click <img alt="end-call" />.

During a call, you can:

* **Monitor the call duration**. The timer displayed next to the <img alt="end-call" /> button shows how long the call has been active.
* **Send DTMF tones**. Click <img alt="dialpad" /> to open a dialpad and send touch-tone (DTMF) signals during the call. With the dialpad, you can enter codes or navigate menus during a call.
* **Follow the conversation**. View the real-time conversation transcript in the Interaction Panel to track the call.

Voice call mode has a timeout. If you don't provide any voice input for a period, the call ends automatically to prevent unnecessary connection time.

To view and analyze the payload of a voice input, click <img alt="expand" /> indicated in the message fields.

## More Information

* [Voice Gateway Nodes](/ai/agents/develop/node-reference/voice/voice-gateway/overview)
* [Voice Preview](/ai/agents/test/voice-preview)
* [Voice Gateway](/voice-gateway/overview)


# Logs
Source: https://docs.cognigy.com/ai/agents/test/logs



*Logs* are system records that contain information about input and output messages that AI Agents receive and send. The primary purpose of the logs is to debug workflows by observing if and where errors are occurring.

<Frame>
  <img alt="Cognigy Logs page showing input and output message records" />
</Frame>

You can track logs in real time or load the log history to review outputs that occurred in the past.

## Limitations

* Log entries on the Logs page are retained for up to 24 hours. By default, the Logs page displays the logs of the last 2 hours. To change this value, contact [Cognigy technical support](/help/get-help). Alternatively, use the [GET Log Entries](/api-reference-logs/logs-v21/get-log-entries) API method to view all retained logs.

### Log entries

The following table provides an overview of the log entries.

| Field       | Description                                                                                                                                                                                                                                                                                                       |
| ----------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `timestamp` | The date and time of the log entry. The format is `YYYY-MM-DD HH:MM:SS`.                                                                                                                                                                                                                                          |
| `log type`  | One of the following log levels: <ul><li>`info` — contains general information about the normal operation of a Flow.</li><li>`error` — contains information about errors that occur during Flow execution.</li><li>`debug` — contains detailed information about Flow execution for debugging purposes.</li></ul> |
| `message`   | Describes the event of the log entry.                                                                                                                                                                                                                                                                             |
| `metadata`  | Additional data associated with the log entry, for example, `organisationID`, `flowID`, `text`, and other.                                                                                                                                                                                                        |

## Working with Logs

### View Logs

<Tabs>
  <Tab title="GUI">
    <Note>
      To improve browser performance, the log list uses virtualization. This means your browser loads only the log entries visible on your screen at any given time. As a result, you can't copy more lines than what's currently visible.
    </Note>

    In **Test > Logs**, you can view and filter Project logs as well as search them by user ID and Flow name.

    **Settings**

    To configure the log entries displayed in the log list, use the following settings:

    | Setting Name             | Function                                                                                      |
    | ------------------------ | --------------------------------------------------------------------------------------------- |
    | Live Logging             | Activates real-time log updates. This setting is activated by default.                        |
    | Show Timestamp           | Includes the timestamp of the log entries. This setting is activated by default.              |
    | Show Additional Metadata | Includes metadata related to input and output messages. This setting is activated by default. |
    | Info                     | Includes log entries of the `info` type. This setting is activated by default.                |
    | Error                    | Includes log entries of the `error` type. This setting is activated by default.               |
    | Debug                    | Includes log entries of the `debug` type. This setting is deactivated by default.             |
  </Tab>

  <Tab title="API">
    You can view the logs of a Project with the [Cognigy Logs API](/api-reference-logs/).
  </Tab>
</Tabs>

### Write Logs

Use a [Code Node](/ai/for-developers/code/overview) to write logs with [api.log](/ai/for-developers/code/api-functions#api-log):

```javascript theme={null}
api.log("info", "This is an informational log message.");
api.log("error", "An error occurred during execution.");
api.log("debug", "Detailed debug information.");
```

The recorded logs are available on the [Logs page and via API](#view-logs).


# Mocking
Source: https://docs.cognigy.com/ai/agents/test/mocking



<a href="/release-notes/2026.10"><Badge>Added in 2026.10</Badge></a>

*Mocking* lets you change the behavior of a [Node](/ai/agents/develop/nodes/overview) by replacing its logic with mock code without changing the original Node settings. This approach allows you to keep the original Flow logic and mock Node behaviors to test specific Flow conditions in the [Simulator](/ai/agents/test/simulator/overview), the [Interaction Panel](/ai/agents/test/interaction-panel/overview), and [Endpoints](/ai/agents/deploy/endpoints/overview) without reconfiguring Nodes.

## Key Benefits

* **Testing Flexibility**. Replace any Node's output or function without changing Flow paths or production logic. To change the Node's behavior, you can just activate mocking and add mock code.
* **Cost Savings**. Run simulations with lower costs or less time by mocking data to avoid sending HTTP or LLM requests.
* **Consistent Testing Interactions**. Control interactions for AI Agent testing by mocking LLM responses with deterministic data.

## Restrictions

* If you activate mocking for a Node that has child Nodes, for example, the Then and Else Nodes of an [If Node](/ai/agents/develop/node-reference/logic/if), the child Nodes aren't executed. However, you can use mocking on child Nodes.

## How Mocking Works

Mocking provides the **Mock Code** field in every Node's settings in the Node editor. This field supports JavaScript code to access the [Input](/ai/agents/develop/ai-agent-memory/input), [Context](/ai/agents/develop/ai-agent-memory/context), and [Profile](/ai/agents/develop/ai-agent-memory/profile) objects, and `api.say` [function](/ai/for-developers/code/api-functions#api-output-api-say). The mock code entirely overrides the configured Node's execution logic at run-time. With the **Mock Code** field, you can mock:

* The AI Agent output, for example, `api.say("Are there any requirements for the drilling machine?")`.
* User input and data, and third-party requests, by storing data in:
  * the Input object, for example, `input.text = "I would like to buy a drilling machine"`.
  * the Context object, for example, `context.selectedProduct = "Drilling machine"`.
  * the Profile object, for example, `profile.name = "John Doe"`.

<Warning>
  If you activate mocking for Nodes that control Flow paths, for example, the [If](/ai/agents/develop/node-reference/logic/if) and [Go To Nodes](/ai/agents/develop/node-reference/logic/go-to), the Flow execution is affected. This approach can lead to tests and simulations that don't reflect your use case, for example, one of the branches of the If Node can't be executed.
</Warning>

## How to Use Mocking

To use mocking, you need to activate it at both levels:

* the [Node level](/ai/agents/test/mocking#at-the-node-level). Enter the mock code to replace the Node's logic.
* the [interface level](/ai/agents/test/mocking#at-the-interface-level). The mock code replaces the Node's logic only in the interface you activate mocking for. You can use the following interfaces to run tests with mock code:

  * [Interaction Panel](/ai/agents/test/interaction-panel/overview)
  * [Simulator](/ai/agents/test/simulator/overview)
  * [individual Endpoints](/ai/agents/deploy/endpoints/overview)

By selecting the interface for mocking, you can limit where mocking applies. For example, you can have interactions using the original Flow logic through one Endpoint and test conversations with the mock code logic through another Endpoint or in the Simulator.

### Activate Mocking

<AccordionGroup>
  <Accordion title="At the Node Level">
    1. Go to **Build > Flows**, select the Flow you want to test.
    2. Right-click the Node you want to mock a behavior for and select **Enable Mocking**. The <img alt="mocking" /> icon appears next to the Node to indicate that the Node logic is replaced by the mock code.
    3. Click the Node. The Node editor opens.
    4. In the **Mocking** section, enter the mock code in the **Mock Code** field, for example, `api.say("Are there any requirements for the drilling machine?")`. Save the Node.
  </Accordion>

  <Accordion title="At the Interface Level">
    <Tabs>
      <Tab title="Interaction Panel">
        1. In the upper-right corner, click <img alt="interaction-panel" /> to open the Interaction Panel.
        2. Click the **Settings** tab and toggle on the **Enable Mocking** setting in the **Advanced** section.

        Now, you can run the Flow with the mock code in the Interaction Panel, both in [chat](/ai/agents/test/interaction-panel/chat) and [simulation](/ai/agents/test/interaction-panel/simulation) modes.
      </Tab>

      <Tab title="Simulator">
        1. In **Test > Simulator**, select the scenario you want to run with the mock code.
        2. In the **Start Simulation** panel, toggle on the **Enable Mocking** setting.
      </Tab>

      <Tab title="Endpoints">
        1. In **Deploy > Endpoints**, select the Endpoint you want to test with the mock code.
        2. In the **Advanced Settings** section of the **Test Endpoint** panel, toggle on the **Enable Mocking** setting.
      </Tab>
    </Tabs>
  </Accordion>
</AccordionGroup>

### Deactivate Mocking

To fully deactivate mocking, deactivate it at Node and interface levels.

However, you can selectively deactivate mocking:

* Deactivate mocking only at the Node level — switches only the Node's execution logic back to the original Node settings. If you have other Nodes with active mocking, their mock code is executed instead of the original Node settings.
* Deactivate mocking only at the interface level — switches the Node execution logic of all Nodes with mock code back to the original Node settings for that interface. If you have active mocking for other interfaces, the mock code is executed instead of the original Node settings.

This approach provides flexibility to easily switch between mocked and original behavior.

<AccordionGroup>
  <Accordion title="Deactivate at the Node Level">
    1. Go to **Build > Flows**, select the relevant Flow.
    2. Right-click the Node you want to deactivate mocking for and select **Disable Mocking**.
  </Accordion>

  <Accordion title="Deactivate at the Interface Level">
    <Tabs>
      <Tab title="Interaction Panel">
        1. In the upper-right corner, click <img alt="interaction-panel" /> to open the Interaction Panel.
        2. Click the **Settings** tab and toggle off the **Enable Mocking** setting in the **Advanced** section.
      </Tab>

      <Tab title="Simulator">
        1. In **Test > Simulator**, select the scenario you want to run with the mock code.
        2. In the **Advanced** section of the **Start Simulation** panel, toggle off the **Enable Mocking** setting.
      </Tab>

      <Tab title="Endpoints">
        1. In **Deploy > Endpoints**, select the Endpoint you want to test with the mock code.
        2. In the **Advanced Settings** section of the **Test Endpoint** panel, toggle off the **Enable Mocking** setting.
      </Tab>
    </Tabs>
  </Accordion>
</AccordionGroup>

## Examples

The following examples show how to use mock code in different cases.

<AccordionGroup>
  <Accordion title="Mock a User Response">
    To test how your AI Agent performs when it receives a predefined response from the user to a [Question Node](/ai/agents/develop/node-reference/basic/question), you can mock the response. Consider that you have a Question Node that stores the user's response in `input.result`. Follow these steps:

    1. Right-click the Question Node you want to mock a behavior for and select **Enable Mocking**.
    2. Open the Node editor and enter the code to mock the user's response, for example:
       ```javascript theme={null}
       input.result = "I would like to order a drilling machine"
       ```
  </Accordion>

  <Accordion title="Mock an AI Agent Response">
    To test a predefined response instead of an LLM-generated response, for example, from an [LLM Prompt Node](/ai/agents/develop/node-reference/service/llm-prompt), follow these steps:

    1. Right-click the LLM Prompt Node you want to mock a behavior for and select **Enable Mocking**.
    2. Open the Node editor and enter the code to mock the AI Agent's response, for example:
       ```javascript theme={null}
       api.say("Are there any requirements for the drilling machine?")
       ```
  </Accordion>

  <Accordion title="Replace Costly Requests">
    Your Flow may contain HTTP or LLM requests that may incur costs or take a long time to complete. These requests could cause issues if you test your AI Agent with frequent simulations. To avoid these issues, you can mock a request during the simulations, for example, in an [HTTP Request Node](/ai/agents/develop/node-reference/service/http-request). Consider an HTTP Request Node that retrieves information about a movie from the [OMDB API database](https://www.omdbapi.com/) and stores it in the Input object under `input.movie`. You can mock the request of the HTTP Request Node with a predefined response. To do so, follow these steps:

    1. Right-click the HTTP Request Node you want to mock a behavior for and select **Enable Mocking**.
    2. Open the Node editor and enter the mock output in the **Mock Code** field, for example:
       ```javascript theme={null}
       input.movie = {
         "Title": "Inception",
         "Year": "2010",
         "Rated": "PG-13",
         "Released": "16 Jul 2010",
         "Runtime": "148 min",
         "Genre": "Action, Adventure, Sci-Fi",
         "Director": "Christopher Nolan",
         "Writer": "Christopher Nolan",
         "Actors": "Leonardo DiCaprio, Joseph Gordon-Levitt, Elliot Page",
         "Plot": "A thief who steals corporate secrets through the use of dream-sharing technology is given the inverse task of planting an idea into the mind of a CEO.",
         "Language": "English, Japanese, French",
         "Country": "USA, UK",
         "Awards": "Won 4 Oscars. Another 152 wins & 218 nominations."
       }
       ```
  </Accordion>
</AccordionGroup>

## More Information

* [Simulator](/ai/agents/test/simulator/overview)
* [Interaction Panel](/ai/agents/test/interaction-panel/overview)
* [Endpoints](/ai/agents/deploy/endpoints/overview)


# Playbooks
Source: https://docs.cognigy.com/ai/agents/test/playbooks



*Playbooks* are conversation scripts that you can run automatically to test your Flows. When running a Playbook, you can set the Flow, Locale, and Snapshot you want to test.

Playbooks include:

* **Steps** — simulate a user input and can include an input text and input data.
* [Assertions](/ai/agents/test/assertions) — check the AI Agent's responses and different aspects of the Flow for each Step against a defined value. You can also invert the Assertion check, which means the Assertion checks if the defined value is missing.

<Accordion title="Assertion Types">
  | Type                   | Description                                                | API Value       |
  | ---------------------- | ---------------------------------------------------------- | --------------- |
  | Message text assertion | Checks the text content of the AI Agent's response.        | `assertText`    |
  | Message data assertion | Checks the data the AI Agent sent.                         | `assertData`    |
  | State assertion        | Checks the current State of the Flow.                      | `assertState`   |
  | Context assertion      | Checks the current data in the Context object of the Flow. | `assertContext` |
  | Slots assertion        | Checks if specific Slots have been recognized.             | `assertSlots`   |
  | Intent assertion       | Checks if specific Intents have been recognized.           | `assertIntent`  |
</Accordion>

## Limitations

* You can run up to 10 Playbooks simultaneously. You can change this limit for on-premises installations by specifying the `MAX_CONCURRENT_PLAYBOOK_EXECUTIONS` variable in the `values.yaml` file.
* A Playbook can contain up to 50 Steps.

## Restrictions

* Playbooks don't allow dots `(.)` in JSON keys. Referencing subkeys like `user.name` can cause runs to be hidden or [Snapshots](/ai/agents/deploy/snapshots) to fail. Instead, validate the full JSON object:

  <CodeGroup>
    ```json Valid Example title theme={null}
    {
       "user": {
           "name": "John"
       }
    }
    ```

    ```json Invalid Example title theme={null}
    {
        "user.name": "John"
    }
    ```
  </CodeGroup>

### Working with Playbooks

<Tabs>
  <Tab title="GUI">
    In **Test > Playbooks**, you can create, delete, and run Playbooks as well as export them as a Package. To define Steps and Assertions in the Playbook editor, select a Playbook from the list. Playbook runs are [Tasks](/ai/agents/develop/projects-and-flows/projects#task-panel) and take place in the background. The results are available in the Playbook editor.

    Alternatively, you can also run a Playbook and view the results from the [Interaction Panel](/ai/agents/test/interaction-panel/overview), as well as create Playbooks from a conversation with the AI Agent.

    **Playbook Editor**

    In the Playbook editor, configure the following settings for all Steps:

    | Setting              | Description                                                                                                                                                                            |
    | -------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Set timeout per step | The timeout for each step in the Playbook in milliseconds. If a step takes longer than the set timeout, an error is displayed in the Playbook player. The default timeout is 20000 ms. |
    | Abort on error       | Select whether to continue or abort the Playbook after a failed step.                                                                                                                  |

    For each Step, you can enter input text and data to send to the AI Agent, and add Assertions to check the AI Agent's response to the input.

    **Runs Tab**

    On the **Runs** tab in the Playbook editor, you can run Playbooks and view the results of the Playbook runs. You can also filter the runs by status and creator.

    After the Playbook run is completed, click it in the list to see the results. In the pane that opens on the right, you can hover over the Assertion checks for more details.

    **Interaction Panel**

    To run a Playbook in the Interaction Panel, select and run a Playbook from the <img alt="playbook" /> tab. You can hover over the Assertion checks for more details.

    To create Playbooks from a conversation, follow these steps:

    1. Chat with the AI Agent on the <img alt="chat" /> tab.
    2. Click <img alt="menu" /> above the chat window and select **Create Playbook**.

    You are redirected to the Playbook editor. You can then add Assertions and additional Steps.
  </Tab>

  <Tab title="CLI">
    You can run Playbooks using the [Cognigy CLI](https://github.com/Cognigy/Cognigy-CLI?tab=readme-ov-file#command-run). Schedule multiple Playbook runs simultaneously using the `run` command.
  </Tab>

  <Tab title="API">
    You can create, edit, and delete Playbooks using the [Cognigy API](https://api-trial.cognigy.ai/openapi#tag--Playbooks-v2.0) as well as run Playbooks.
  </Tab>
</Tabs>

## Examples

The following video provides examples of how to create and run a Playbook using the Interaction Panel.

<iframe title="YouTube video player" />


# Simulator Dashboard
Source: https://docs.cognigy.com/ai/agents/test/simulator/dashboard



<a href="/release-notes/2026.3"><Badge>Added in 2026.3</Badge></a>

The *Simulator Dashboard* provides a quick overview of the key metrics of simulations through charts, indicators, and tables. It includes performance trends, recent simulations, and general information.

The summary of key metrics is represented in indicators and a bar chart. The tables display the scheduled and recent simulations.

<Frame>
  <img alt="Simulator Dashboard showing key metrics with indicators, bar charts, and tables of scheduled and recent simulations" />
</Frame>

## Prerequisites

* You need to have run or scheduled at least one simulation to see the Simulator dashboard.

## Filters

In the upper-right corner of the dashboard, you can select the period and the scenarios for which you want to display the data:

| Filter    | Description                                                                                                                                                                                    |
| --------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Period    | Select the period for which you want to display the data. You have the following options:<ul><li>Last 7 days. This is the default option.</li><li>Last 30 days</li><li>Last 3 months</li></ul> |
| Scenarios | Select the scenarios for which you want to display the data. You can enter the name of the scenario in the search field at the top of the list.                                                |

<Tip>
  Use filters to focus on the data of a group of related scenarios and to view how their performance develops over time.
</Tip>

## Indicators

### Total Scenarios

The **Total Scenarios** indicator displays the number of scenarios in the Project.

### Total Simulations

The **Total Simulations** indicator displays the number of simulation runs in the selected period.

### Avg Success Rate

The **Avg Success Rate** indicator displays the average success rate of the simulation runs in the selected period.

## Chart

### Success Rate Trend

The **Success Rate Trend** chart displays the proportion of succeeded and failed simulation runs during the selected period. The horizontal axis represents the days or weeks within the selected period. The vertical axis represents the split between succeeded and failed simulation runs, where the lower portion portrays the succeeded simulation runs and the upper portion portrays the failed simulation runs. Hover over the bars in the chart to view more details about the success rate trend, for example, percentage of succeeded and failed simulation runs.

## Tables

### Upcoming Scheduled Runs

The **Upcoming Scheduled Runs** table displays the next scheduled simulations. To view more details about the schedule, click a simulation, and you will be redirected to the **Schedule Configuration** panel on the **Scenarios** tab.

### Recent Simulations

The **Recent Simulations** table displays the latest completed simulations. Click a simulation to access the simulation results overview.

## Simulator Dashboard API

You can use the [Cognigy Simulator API](/api-reference-simulator/) to access the Simulator dashboard data.

## More Information

* [Simulator Overview](/ai/agents/test/simulator/overview)


# Simulator
Source: https://docs.cognigy.com/ai/agents/test/simulator/overview



<a href="/release-notes/2026.15"><Badge>Updated in 2026.15</Badge></a>

*Simulator* is an LLM-powered tool in Cognigy that allows users to model and test full conversational scenarios for LLM-based AI Agents.

Use the Simulator to mimic user interactions in controlled environments and evaluate the performance, behavior, and reliability of AI Agents through metrics displayed on the Simulator Dashboard. In contrast to Playbooks, which follow predefined test paths, the Simulator uses Generative AI to create dynamic, realistic conversations.

The Simulator contains the following components:

* **Scenarios** — define personas, missions, and success criteria for testing AI Agents.
* **Simulation Runs** — simulate scenarios to evaluate the performance of AI Agents.
* **Simulations** — the compilation of multiple simulation runs displaying different AI Agent behaviors in the same scenario. Each simulation run provides unique test outcomes for analysis. Additionally, you can automate testing by scheduling periodic simulations, and ensure continuous validation of your AI Agent's performance and reliability.

The Simulator is primarily designed for testing LLM-based AI Agents, but it can also be used with any Flow, with some restrictions.

<Frame>
  <img alt="Simulator overview interface showing Scenarios, Simulation Runs, and Simulations components for testing LLM-based AI Agents" />
</Frame>

## Key Benefits

* **Realistic End-to-End Testing**. Simulate full conversations, from greetings to resolutions. Use multiple personas to test how your AI Agent handles different user personas and Scenarios.
* **Simulations Across Multiple Interfaces**. Run simulations via the Simulator or the Interaction Panel and view all simulated conversations in the Conversations Explorer, Simulator, or Interaction Panel for quick outcome review.
* **Early Issue Detection and Optimization**. Detect gaps, edge cases, and quality issues before deployment to improve reliability and deliver a better user experience.
* **Faster, Automated QA**. Automatically test AI Agents across environments, ensuring high coverage and consistent quality.
* **Scheduled Simulations**. Schedule simulations to automate periodic tests at a specified frequency. Regular testing helps ensure that new deployments work as expected.

## Prerequisites

* To start a simulation, you need an LLM that supports the [AI Agent Node feature](/ai/agents/develop/node-reference/ai/ai-agent).
* To generate scenarios automatically, you need to:
  * Configure an LLM for [design-time features](/ai/agents/develop/gen-ai-and-llms/generative-ai#design-time-generative-ai-features). Select the platform-provided LLM or another model that [supports design-time features](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature).
  * Ensure that your Project includes a Flow with an [AI Agent Node](/ai/agents/develop/node-reference/ai/ai-agent).

## Restrictions

* You can't run simulations using Snapshots in the Interaction Panel.
* The following features aren't supported:
  * Voice interactions. The Simulator only supports text-based conversations.
  * Handovers to human agents. The Interaction Panel will show the message `Agent Handover is not supported in the Interaction Panel, please use the Webchat or another channel`.

## Limitations

* The maximum number of scenarios per Project is 100. You can change this limit as follows:
  * For SaaS installations, contact [Cognigy technical support](/help/get-help).
  * For on-premises installations, assign a new limit to the `MAX_SIMULATIONS_PER_PROJECT` environment variable in `values.yaml`.
* The maximum number of runs per simulation is 150.
* The maximum number of success criteria per scenario is 10.
* The maximum number of CSV transcripts you can upload at a time is 10.
* You can set only 1 schedule per scenario.
* Results from simulations are deleted after 90 days.

## How the Simulator Works

<Frame>
  <img alt="Simulator workflow from scenario creation to simulation to evaluation" />
</Frame>

1. **Create a scenario**. Set the persona, mission, and success criteria.
2. **Simulate**. Start the simulation and let the AI Agent interact with the persona you created. Run the simulation multiple times to gather more accurate insights.
3. **Evaluate**. View key metrics on the Simulation dashboard, such as complete success rate, average sentiment, and number of turns. You can then adjust the Flow logic, training data, or system prompts based on the results. Rerun simulations to verify improvements.

## Working with Simulations

<Tabs>
  <Tab title="GUI">
    <AccordionGroup>
      <Accordion title="1. Create a Scenario">
        In the Simulator, you can create a scenario automatically or manually. Alternatively, you can create a scenario from a transcript in the [Conversations Explorer](/insights/explorers/conversations).

        To create a scenario, follow these steps:

        <Tabs>
          <Tab title="Automatically">
            1. In your Project, go to **Test > Simulator** and click **+ New Scenario**.
            2. In the **Generate from AI Agent** section, click **Generate**.
            3. Fill in the following fields:
               * **Flow** – select a Flow containing an AI Agent to base your scenario on. Ensure the Flow is already configured.
               * **AI Agent Job Node** – select the AI Agent Node to be analyzed for scenario creation. The AI Agent Job Node must be already configured in the Flow.
            4. Click **Generate**. Cognigy generates a scenario based on the AI Agent Job Node configuration. Review the generated scenario and adjust it if necessary.
            5. Select the generated persona description, mission, and success criteria. If necessary, edit them manually or automatically by clicking the **Regenerate text** button on the right-hand side of each field.
          </Tab>

          <Tab title="Manually">
            1. In your Project, go to **Test > Simulator** and click **+ New Scenario**.
            2. Fill in the following fields:
               * **Name** – enter the name of the scenario. The name should reflect the test scenario and focus area. Example: `Menu Exploration – Exotic Dish Focus`.
               * **Persona Name** – enter the name of the persona to simulate. The name should reflect the user's role and context. Example: `Alex – The Curious Foodie`.
               * **Persona Description** – enter the description of the persona. To simulate user behavior, define key attributes: customer segment, demographics, familiarity with technology and AI, and interaction style. Example: `An adventurous diner who loves trying new and exotic dishes. Asks detailed questions about ingredients, preparation methods, and flavor profiles`.
               * **Mission** – enter the description of the persona's goal, which should be achieved within the simulation. This goal can involve testing the AI's ability to handle specific customer inquiries or tasks. Example: `Explore exotic menu options and get detailed recommendations for trying new dishes based on preferences`.
            3. In the **Success Criteria** section, add metrics to evaluate the simulation's success:
               * **Add Goal** – select a goal from the list of available goals. If you don't have any goals, you can [create a new one](/ai/agents/analyze/goals-and-tasks/goals). The goal should reflect the simulation's desired outcome. Example: `Receive personalized recommendations for exotic dishes`.
               * **Add AI Judge** – generate or enter the action-oriented criteria to help determine whether the simulation was successful. They can include numeric results, for example, 90% accuracy, or specific objectives the AI Agent should achieve. Examples: `The Agent recommends exotic dishes based on user preferences, achieving 90% accuracy in matching user preferences.`, `Answers questions about ingredients`.
            4. In the **Maximum Turns** slider, set the maximum number of back-and-forth messages (turns) allowed in a simulated conversation. Each turn consists of one user input and one AI Agent response. The default value is 15.
          </Tab>

          <Tab title="From Transcripts">
            1. In your Project, go to **Test > Simulator** and click **+ New Scenario**.
            2. In the **Create from Transcript** section, click **Generate**.
            3. In the **Upload Conversation Transcripts** window, upload a conversation transcript in CSV format.
               Before uploading the CSV file, ensure that it meets the following requirements:

               <Accordion title="CSV Format Requirements">
                 The CSV must contain three columns, in this order:

                 1. `speaker` — the name of the person or system sending the message. Use consistent names such as `customer` for the end user and `bot` for the AI Agent.
                 2. `message` — the text of the message, enclosed in double quotes (`" "`).
                 3. `created_at` — the timestamp in ISO 8601 UTC format (`YYYY-MM-DDTHH:mm:ssZ`).

                 **Example:**

                 ```csv theme={null}
                 speaker,message,created_at
                 customer,"Hello, I need help with my order.",2026-03-12T09:15:00Z
                 bot,"Good morning! Can you provide your order number?",2026-03-12T09:15:05Z
                 ```
               </Accordion>

               <Accordion title="Content Requirements">
                 * Each transcript must contain at least 100 words, counting only the text in the `message` column. Words in the `speaker` and `created_at` columns aren't counted.
                 * The CSV file must include messages from both actors: customer and bot.
                 * The `created_at` column must follow chronological order.
                 * Each row must represent one message.
                 * Each message must be a complete sentence or meaningful utterance.
                 * Messages can include punctuation, emojis, or line breaks.
               </Accordion>

            You can also create a scenario from an existing conversation transcript in the [Conversations Explorer](/insights/explorers/conversations#scenarios).
          </Tab>
        </Tabs>
      </Accordion>

      <Accordion title="2. Simulate">
        After creating a scenario, you can start a simulation manually or [schedule regular simulations](#schedule-simulations).

        To start a simulation, follow these steps:

        1. Click the scenario from the list to open the **Start Simulation** window.
        2. In the **Start Simulation** window, configure the simulation:
           * **Simulation Name** – enter the name of the simulation. The name should reflect the test scenario and focus area. Example: `Menu Exploration – Exotic Dish Focus`.
           * **Snapshot** – select the [Snapshot](#schedule-simulations) for the simulation. Use this option to test a Flow from a specific version rather than the current one.
           * **Flow** – select the Flow for the simulation.
           * **Locale** – select the [Locale](/ai/platform-features/translation-and-localization/localization) that defines the language of the AI Agent and simulated user.
           * **LLM** – select the model to manage the simulated conversation.
           * **Number of Simulation Runs** – set how many times the simulation runs. The default value is 10. Use multiple simulation runs to test conversation variations.
        3. *(Optional)* In the **Advanced Settings** section, configure additional parameters:
           * **AI Agent Output Timeout** – set the maximum time the AI Agent has to respond to each user input during the simulation. The default value is 60,000 milliseconds. The minimum value is 20,000 milliseconds.
           * **Custom Data Payload** – define a JSON object that is injected into every input message during a simulation. This parameter enables context-aware testing, for example, testing different user types, A/B scenarios, or personalized Flows, without changing the Flow itself. Examples of custom data payloads:
           ```json theme={null}
           { "userId": "u1001", "role": "admin" }      // different user types
           { "testGroup": "A", "featureFlag": true }   // A/B testing
           { "userName": "Emma", "preferredLanguage": "fr" } // personalized greeting
           ```
        4. Click **Start Simulation**. Once the simulation starts, you'll be redirected to the simulation overview to view the real-time results. After creating the Project's first simulation, the Simulator Dashboard is added to the Simulator interface.
      </Accordion>

      <Accordion title="3. Evaluate">
        The simulation overview provides the following key metrics:

        | **Chart**             | **Description**                                                                                                                                                   |
        | --------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------- |
        | Complete Success Rate | Shows the percentage of simulation runs that met the success criteria. A higher success rate indicates better overall performance.                                |
        | Success Criteria Met  | Shows the percentage of success criteria met across all simulation runs.                                                                                          |
        | Average Sentiment     | Shows the average sentiment across all simulation runs.                                                                                                           |
        | Average Turns         | Shows the average number of back-and-forth messages (turns) it took to complete a simulation run. Each turn consists of one user input and one AI Agent response. |
        | Turn Distribution     | Shows the frequency of turn counts within ranges. The horizontal axis represents turn count ranges, and the vertical axis shows the frequency for each range.     |
        | Success Criteria      | Shows how often a success criterion was met across all simulation runs in percentage.                                                                             |
        | Simulation Results    | Shows how many simulation runs succeeded and failed. A simulation succeeds if all success criteria are met.                                                       |

        In the **Runs** table, find the simulation runs marked as `Failed` in the **Complete Success** column. Click a simulation run to open the simulated conversation in a panel on the right. In the right-side panel, hover over the `Failed` status next to the success criterion to see a tooltip explaining why the simulation run failed to meet the success criterion.

        To filter simulation runs, click a bar or a slice on the charts that represents the data you want to investigate.

        You can also view the simulated conversation in the [Conversations Explorer](/insights/explorers/conversations) by toggling on **Interaction Panel / Playbooks / Simulations** in the Endpoint filters.

        Additionally, you can view key metrics, performance trends, and scheduled simulations across simulations in the [Simulator dashboard](/ai/agents/test/simulator/dashboard).
      </Accordion>
    </AccordionGroup>
  </Tab>

  <Tab title="API">
    You can use the [Cognigy Simulator API](/api-reference-simulator/) to create scenarios, run simulations, and retrieve results from simulations and simulation runs.
  </Tab>
</Tabs>

### Schedule Simulations

After you've [created a scenario](#1-create-a-scenario), you can schedule simulations to run automatically at a specified frequency.

<Tabs>
  <Tab title="GUI">
    To schedule a simulation, follow these steps:

    1. Hover over the scenario for which you want to schedule a simulation and select <img alt="menu" /> **> Schedule** on the right. The **Schedule Configuration** panel opens.
    2. Configure the simulation and schedule settings on the tabs in the **Schedule Configuration** panel:

    <AccordionGroup>
      <Accordion title="On the Run Config Tab">
        1) Configure the following:
           * **Simulation Name** — enter the name of the simulation. The name should reflect the test scenario and focus area. Example: `Menu Exploration – Exotic Dish Focus`. The time and date are appended to the name after the simulation is finished. The full simulation name is displayed on the **Simulations** tab and Simulator dashboard.
           * **Snapshot** — select the Snapshot for the simulation.
           * **Flow** — select the Flow for the simulation.
           * **LLM** — select the model to manage the simulated conversation.
        2) *(Optional)* In the **Advanced Settings** section, configure additional parameters:
           * **AI Agent Output Timeout** – set the maximum time the AI Agent has to respond to each user input during the simulation. The default value is 60,000 milliseconds. The minimum value is 20,000 milliseconds.
           * **Custom Data Payload** – define a JSON object that is injected into every input message during a simulation. This parameter enables context-aware testing, for example, testing different user types, A/B scenarios, or personalized Flows, without changing the Flow itself. Examples of custom data payloads:
           ```json theme={null}
           { "userId": "u1001", "role": "admin" }      // different user types
           { "testGroup": "A", "featureFlag": true }   // A/B testing
           { "userName": "Emma", "preferredLanguage": "fr" } // personalized greeting
           ```
      </Accordion>

      <Accordion title="On the Schedule Tab">
        1. Toggle on **Enable Schedule** to activate the schedule.
        2. Configure the following:

        * **Frequency** — select from the following options:
          * **Daily** — this option is selected by default.
          * **Every 3 days**
          * **Weekly**
          * **Bi-weekly**
          * **Monthly**
        * **Time** — select the time when the simulation starts.
        * **Next Scheduled Run** — shows the date and time of the next simulation, based on the **Frequency** and **Time** parameters.
        * **Number of Simulation Runs** — set how many times the simulation runs. The default value is 10. Use multiple simulation runs to test conversation variations.
        * **Email Notifications** — enter the email addresses to receive notifications about the simulation runs and press <kbd>Enter</kbd>.
      </Accordion>
    </AccordionGroup>
  </Tab>

  <Tab title="API">
    To schedule a simulation, use the [`POST /simulations/{simulationReference}/schedules`](https://api-trial.cognigy.ai/testing/openapi#post-/simulations/-simulationReference-/schedules) request.
  </Tab>
</Tabs>

After saving the configuration, the schedule icon <img alt="schedule" /> appears next to the scenario name on the **Scenarios** tab. Hover over the icon to view the next scheduled simulation.

## Example

This example shows how to create scenarios and run simulations using an AI Agent Flow. We'll work with Sophie, the Restaurant Guide from the [Job Market](/ai/agents/overview#job-market). Sophie assists users with dinner planning and restaurant recommendations.

1. Set up a Flow that uses Sophie as the AI Agent.

2. Set up three simulations with different personas to test various use cases:

   <AccordionGroup>
     <Accordion title="Persona 1. 🧑‍🍳 Alex – The Curious Foodie">
       | **Parameter** | **Value**                                                                                                                                                                                                               |
       | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
       | Scenario Name | Menu Exploration – Exotic Dish Focus                                                                                                                                                                                    |
       | Persona Name  | Alex – The Curious Foodie                                                                                                                                                                                               |
       | Persona       | An adventurous foodie in their late 20s who loves discovering new and exotic dishes. Tech-savvy and curious, Alex uses mobile apps and AI chat tools to explore unique menu items and expand their culinary experience. |
       | Mission       | Test the AI Agent's ability to recommend exotic dishes and assist Alex in placing an order or booking a table.                                                                                                          |
       | AI Judge      | <ul><li>Accurately handle free-text menu queries</li><li>Suggest complementary drinks after exotic dishes are added</li><li>Recognize and add at least one manually named exotic dish</li></ul>                         |
     </Accordion>

     <Accordion title="Persona 2. 👨 Jordan – The Corporate Guest">
       | **Parameter** | **Value**                                                                                                                                                                                                          |
       | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
       | Scenario Name | Quick Reservation – Dietary Notes                                                                                                                                                                                  |
       | Persona Name  | Jordan – The Corporate Guest                                                                                                                                                                                       |
       | Persona       | Marketing manager at a nearby company, 35 years old, with short lunch breaks. Tech-savvy, uses AI tools frequently, and expects fast, concise communication. Prefers professional tone and minimal back-and-forth. |
       | Mission       | Test how well the AI Agent handles quick lunch queries and reservation requests under time constraints.                                                                                                            |
       | AI Judge      | <ul><li>Reservation completed with dietary requirements noted</li><li>Minimal back-and-forth (1–2 exchanges)</li><li>Confirmation provided in under 30 seconds</li></ul>                                           |
     </Accordion>

     <Accordion title="Persona 3. 👩 Maya – The Tourist Planner">
       | **Parameter** | **Value**                                                                                                                                                                                                                                        |
       | ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
       | Scenario Name | Family-Friendly Group Dinner                                                                                                                                                                                                                     |
       | Persona Name  | Maya – The Tourist Planner                                                                                                                                                                                                                       |
       | Persona       | A woman in her early 40s visiting from out of town, organizing a dinner for a group of 6, including kids. Familiar with basic chat tools, occasionally uses AI assistants. Communicates politely and needs detailed answers to plan confidently. |
       | Mission       | Evaluate how well the AI Agent handles group reservations and provides family-friendly recommendations.                                                                                                                                          |
       | AI Judge      | <ul><li>Group reservation successfully created</li><li>At least one child-friendly menu item suggested</li><li>Additional info (for example, parking, high chairs) proactively offered</li></ul>                                                 |
     </Accordion>
   </AccordionGroup>

3. Run simulations with the following settings:

   | **Parameter**             | **Value**                                                                                         |
   | ------------------------- | ------------------------------------------------------------------------------------------------- |
   | Snapshot                  | No Snapshot                                                                                       |
   | Flow                      | Main-Dining Concierge                                                                             |
   | Locale                    | en-US                                                                                             |
   | LLM                       | Default. Make sure the default model supports Generative AI capabilities such as text completion. |
   | Number of Simulation Runs | 10                                                                                                |

4. Check the results on the Simulation dashboard. Open each simulation run to review the outcomes. Pay attention to failed runs. For example, look at the simulation for the persona **Maya – The Tourist Planner**.

| Metric                | Value    | What It Means                                              | Action Steps                                                                                                                    |
| --------------------- | -------- | ---------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------- |
| Complete Success Rate | 40%      | 40% of the simulation runs met all success criteria.       | Check the Flow for misunderstandings of user requests or missed information.                                                    |
| Success Criteria Met  | 67%      | Simulation runs met 67% of the success criteria in total.  | Review which specific success criteria failed, then adjust the Flow logic or refine the success criteria definitions as needed. |
| Average Sentiment     | Positive | Users show positive feelings.                              | No action is needed.                                                                                                            |
| Average Turns         | 9        | Average number of exchanges in the simulated conversation. | Track conversation length and ensure each turn effectively advances the conversation.                                           |

## Troubleshooting

<AccordionGroup>
  <Accordion title="Common Issues">
    | **Issue**                      | **Description**                                                  | **How to Resolve**                                                                                                                 |
    | ------------------------------ | ---------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------- |
    | Simulation doesn't run         | The simulation fails to start or shows an error.                 | Ensure all required fields (Flow, Locale, LLM, Number of Simulation Runs) are completed. Check system status on the **Logs** page. |
    | Inaccurate result              | The output doesn't match the expected behavior or test scenario. | Check your Flow version and Snapshot. Use the latest and most relevant setup for testing.                                          |
    | AI Agent responses not aligned | The AI Agent gives irrelevant, vague, or incorrect replies.      | Review persona and mission definitions. Ensure the LLM supports the expected output style.                                         |
  </Accordion>

  <Accordion title="Azure Content Filter Error">
    When you use an Azure OpenAI LLM for simulations, you may receive the following error message:

    `The response was filtered due to the prompt triggering Azure's content management policy.`

    This error isn't caused by Cognigy, it occurs because [Azure's content filter](https://learn.microsoft.com/en-us/azure/ai-foundry/openai/concepts/content-filter?view=foundry-classic) blocks the LLM prompt, also called system message, according to your subscription's safety policy. The LLM prompt includes your simulation settings, for example, personas, missions, and success criteria.

    **Best Practices**

    Azure's content filter blocks any content that it considers unsafe or ambiguous, including LLM prompts. The exact rules, scoring methods, or trigger terms used for the content filter aren't disclosed. However, they include best practices in their [system message documentation](https://learn.microsoft.com/en-us/azure/ai-foundry/openai/concepts/system-message?view=foundry-classic\&tabs=top-techniques#summary-of-best-practices). The following table shows issues and best practices while configuring the scenario.

    | Issues                                                         | Best Practice                                                | Examples and Recommendations                                                                       |
    | -------------------------------------------------------------- | ------------------------------------------------------------ | -------------------------------------------------------------------------------------------------- |
    | The prompt lacks sufficient context.                           | Provide clear context and purpose to guide the model safely. | Enter persona description with at least 50 words.                                                  |
    | References to personal or financial situations.                | Keep a neutral, business-appropriate tone.                   | Use `Customer exploring labor options`, instead of `unemployed`.                                   |
    | Words implying personal finance.                               | Focus on professional skills, goals, and scenarios.          | Use `payment upon delivery`, instead of `cash`.                                                    |
    | Including self-referring instructions in the scenario content. | Use detailed, descriptive personas.                          | Use `They are comfortable navigating digital menus`, instead of `They often use this application`. |
    | Personal framing without professional context.                 | Frame content around business use cases.                     | Use `A customer seeking assistance...`, instead of `A customer struggling with...`.                |

    **How to Resolve**

    1. Review the simulation's persona, mission, and success criteria. Make sure they comply with [Azure's best practices](https://learn.microsoft.com/en-us/azure/ai-foundry/openai/concepts/system-message?view=foundry-classic\&tabs=top-techniques#summary-of-best-practices).

    2. Ask your Azure OpenAI administrator to review the content filter settings for the environment you're using.

    **Examples**

    In the **Persona** parameter, you can influence the context the most. The examples in the following table show how to make sure the LLM prompt includes enough context.

    | <div>Accepted Persona</div>                                                                                                                                                                                                                                                                                                                                                                                           | <div>Blocked Persona</div>                                                            |
    | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------- |
    | `A customer who frequently uses online ordering platforms and is comfortable navigating digital menus and checkout processes. They communicate clearly when placing orders and prefer efficient interactions. This customer understands standard workflows, including item selection, delivery info entry, and payment options. They prefer payment at delivery and appreciate straightforward ordering experiences.` | `A cash-preferring customer who is unemployed and needs help with their application.` |
  </Accordion>
</AccordionGroup>

## More Information

* [Simulations in the Interaction Panel](/ai/agents/test/interaction-panel/simulation)
* [Conversations Explorer](/insights/explorers/conversations)
* [Contact Profiles](/ai/agents/analyze/contact-profiles)
* [Agentic AI](/ai/agents/overview)
* [Azure OpenAI Content Filtering](https://learn.microsoft.com/en-us/azure/ai-foundry/openai/concepts/content-filter?view=foundry-classic)
* [Azure OpenAI System Message Best Practices](https://learn.microsoft.com/en-us/azure/ai-foundry/openai/concepts/system-message?view=foundry-classic\&tabs=top-techniques#summary-of-best-practices)


# Testing your AI Agents
Source: https://docs.cognigy.com/ai/agents/test/testing-your-ai-agents

Test AI Agents and ensure reliable performance before deployment.

Simulate, monitor, and validate AI Agent conversations to ensure smooth and reliable performance before deployment.

<CardGroup>
  <Card title="Simulator" href="/ai/agents/test/simulator/overview">
    <br />Generate and test complete conversational scenarios for LLM-based AI Agents.
  </Card>

  <Card title="Interaction Panel" href="/ai/agents/test/interaction-panel/overview">
    <br />Access a full-featured panel to test, debug, and refine your AI Agent conversations.
  </Card>

  <Card title="Mocking" href="/ai/agents/test/mocking">
    <br />Change the behavior of a Node by overriding its logic with mock code execution without changing the original Node settings.
  </Card>

  <Card title="Playbooks" href="/ai/agents/test/playbooks">
    <br />Design and run automated test cases to verify AI Agent behavior across different scenarios.
  </Card>

  <Card title="Demo Webchat" href="/ai/agents/test/demo-webchat">
    <br />Use a ready-made Webchat to interact with your AI Agent like a real end user would.
  </Card>

  <Card title="Logs" href="/ai/agents/test/logs">
    <br />Review detailed logs and diagnostics to troubleshoot and optimize your AI Agent's performance.
  </Card>

  <Card title="Voice Preview" href="/ai/agents/test/voice-preview">
    <br />Test and listen to how your AI Agent sounds during voice interactions.
  </Card>
</CardGroup>


# Voice Preview
Source: https://docs.cognigy.com/ai/agents/test/voice-preview

Use the Voice Preview feature to quickly test voice output.

<a href="/release-notes/2026.16"><Badge>Updated in 2026.16</Badge></a>

The *Voice Preview* feature lets users quickly test their AI Agents' voice output in the [Flow editor](/ai/agents/develop/projects-and-flows/editor) or [Interaction Panel](/ai/agents/test/interaction-panel/overview) without going through the entire [Flow](/ai/agents/develop/projects-and-flows/overview) with a mock call. This feature supports cloud speech-to-text (STT) and text-to-speech (TTS) services. To use the Voice Preview, you need to enter text or SSML syntax, and select the language and voice.

## Restrictions

* The Voice Preview feature supports only the Google, Microsoft, Amazon AWS, Deepgram, Speechmatics, and ElevenLabs speech providers. When testing a voice Flow through the Interaction Panel, check the STT and TTS settings of the voice Nodes to ensure that one of these providers is specified. If any other provider is specified in the Nodes, the Interaction Panel won't support the voice conversation.
* To use SSML syntax in the Voice Preview, the speech provider must support SSML syntax. For example, ElevenLabs doesn't support SSML. Check the documentation of your speech provider to see if it supports SSML.

## Access STT and TTS Providers

If your environment has network connection restrictions or requires strict security rules, make sure that both STT and TTS providers are accessible. Add the IP address ranges of your STT and TTS providers to the firewall's whitelist. This approach ensures the Cognigy server can interact smoothly with both providers, guaranteeing the expected functionality of the Voice Preview feature.

## How to Use Voice Preview

### Set up a Speech Provider

To set up a speech provider and test your voice agents, follow these steps:

1. In the left-side menu of the Project, go to **Manage > Settings**.
2. Go to the **Voice Preview Settings** section and select one of the following providers from the **Speech Provider** list:

<Tabs>
  <Tab title="Microsoft">
    1) Next to the **Speech Connection** field, click <img alt="add speech connection" /> to enter credentials.<br />
    2) In the **New Connection** window, fill in the following fields:<br />
       * **Connection** — specify a unique connection name.<br />
       * **API Key** — specify an API key for Microsoft Azure Speech Services. Log in to the Azure portal, navigate to your Speech Services resource, and copy the API key labeled as **Key1** from the **Keys and Endpoint** section. For more information, read the [Microsoft Azure AI documentation](https://learn.microsoft.com/en-us/azure/ai-services/openai/text-to-speech-quickstart?tabs=command-line#set-up).<br />
       * **Region** — this parameter is optional. Enter a specific region if necessary. The supported voices vary by region. Verify which voices are available in the provided region in the [Azure AI Speech regions documentation](https://learn.microsoft.com/en-us/azure/ai-services/speech-service/regions?tabs=tts).
  </Tab>

  <Tab title="Google">
    1. Next to the **Speech Connection** field, click <img alt="add speech connection" /> to enter credentials.<br />
    2. In the **New Connection** window, specify a unique name for your connection in the **Connection** field.<br />
    3. Click **Upload JSON File** and upload the JSON file that you received from Google Speech Services. To obtain a JSON key file for accessing the Google Speech Services, first create a service account in the Google Cloud Console under **IAM & Admin**. Assign the appropriate roles, generate a JSON key file for the service account, and download this key.<br />
  </Tab>

  <Tab title="Amazon">
    1. Next to the **Speech Connection** field, click <img alt="add speech connection" /> to enter credentials.<br />
    2. In the **New Connection** window, fill in the following fields:<br />
       * **Access Key ID** — specify an Access Key ID. Log in to the AWS Management Console, go to the IAM dashboard, select **Users**, and choose the IAM user. Navigate to the **Security credentials** tab, and under **Access keys**, create a new access key if one hasn't been created. Copy the Access Key ID provided after creation.<br />
       * **Secret Access Key** — specify a Secret Access Key. After creating the access key, you'll be prompted to download a file containing the Access Key ID and the Secret Access Key. Alternatively, you can retrieve the Secret Access Key by navigating to the IAM dashboard, selecting the user, going to the **Security credentials** tab, and clicking **Show** next to the Access Key ID to reveal and copy the Secret Access Key.<br />
       * **Session Token** — this parameter is optional. If you use temporary security credentials, obtain the token when using AWS STS (Security Token Service) to assume a role or federate users.<br />
       * **Region** — this parameter is optional. Enter the AWS region where your Amazon Polly resources are located, for example, `us-east-1` for the US East (N. Virginia) region.<br />
  </Tab>

  <Tab title="Deepgram">
    1. Next to the **Speech Connection** field, click <img alt="add speech connection" /> to enter credentials.
    2. In the **New Connection** window, fill in the following fields:
       * **Connection** — specify a unique connection name.
       * **API Key** — specify an API key for Deepgram. To obtain the API key, log in to the Deepgram Console, navigate to the Settings of your Deepgram Project, and select **API Keys**. For more information, read the [Deepgram documentation](https://developers.deepgram.com/docs/create-additional-api-keys).
       * **Region** — select one of the following regions to host the data processed by Deepgram:
         * **United States** — points to the endpoint for Deepgram services located in the U.S.
         * **Europe** — points to the endpoint for Deepgram services located in the EU.
  </Tab>

  <Tab title="ElevenLabs">
    <Note>
      The feature is hidden behind a feature flag. Activate ElevenLabs Speech Services on the Cognigy side:

      * If you have a SaaS installation, contact [Cognigy technical support](/help/get-help).
      * If you have an on-premises installation, specify `FEATURE_ENABLE_ELEVENLABS_STT_VENDOR=true` in `values.yaml`.
    </Note>

    1. Next to the **Speech Connection** field, click <img alt="add speech connection" /> to enter credentials.
    2. In the **New Connection** window, fill in the following fields:
       * **Connection** — specify a unique connection name.
       * **API Key** — specify an API key for ElevenLabs. To get the API key, log in to the ElevenLabs platform, navigate to your account settings, and find the API key section. For more information, read the [ElevenLabs documentation](https://docs.elevenlabs.io/).
       * **Region** — select one of the following regions:
         * **Global** — for all global ElevenLabs services.
         * **Europe** — for ElevenLabs services hosted in the European Union.
       * **TTS model** — select the TTS model to use for text-to-speech conversion.
       * **STT model** — select the STT model to use for speech-to-text conversion.
  </Tab>

  <Tab title="Speechmatics">
    <Note>
      Support for Speechmatics is hidden behind a feature flag. To activate Speechmatics:

      * For SaaS installations, contact [Cognigy technical support](/help/get-help).
      * For on-premises installations, specify the [feature flags](/voice-gateway/webapp/speech-services#speechmatics) in `values.yaml`.
    </Note>

    1. Next to the **Speech Connection** field, click <img alt="add speech connection" /> to enter credentials.
    2. In the **New Connection** window, fill in the following fields:
       * **Connection** — specify a unique connection name.
       * **API Key** — specify an API key for Speechmatics. For more information on Speechmatics API keys, read the [Speechmatics Authentication guide](https://docs.speechmatics.com/get-started/authentication#create-an-api-key).
       * **Region** — select one of the following regions to host the data processed by Speechmatics:
         * **EU (EU2 - On Demand)** — points to the on-demand endpoint for Speechmatics services located in the EU.
         * **EU (EU1 - Enterprise)** — points to the enterprise endpoint for Speechmatics services located in the EU.
         * **US (US1 - Enterprise)** — points to the enterprise endpoint for Speechmatics services located in the U.S.
  </Tab>
</Tabs>

3. Click **Create**.
4. To check the connection, click **Test**.

You can set up multiple speech providers. By default, the speech provider selected in the **Speech Provider** list is applied to your Flows and the Interaction Panel. You can change the voice and speech provider for your AI Agent, for example, in the [AI Agent Node](/ai/agents/develop/node-reference/ai/ai-agent) or [Set Session Config Node](/ai/agents/develop/node-reference/voice/voice-gateway/set-session-config) without switching the default speech provider.

### Use Voice Preview

You can access the Voice Preview feature in two different ways:

* In the Flow editor, click <img alt="voice-preview" />.
* In the Interaction Panel, hover over the AI Agent message in the chat and click <img alt="voice-preview" /> in the interactive tooltip. The Voice Preview button is displayed only for certain types of output, including regular text output, fallback text, and text or SSML output from channels that support voice.

## More Information

* [Interaction Panel](/ai/agents/test/interaction-panel/overview)
* [Voice Gateway](/voice-gateway/index)
* [Voice Nodes](/ai/agents/develop/node-reference/voice/overview)


# Agent Copilot
Source: https://docs.cognigy.com/ai/escalate/agent-copilot



Agent Copilot is a Cognigy product designed to empower human agents with context-specific information, next-step suggestions, and more during conversations. With Agent Copilot, you also have a fully customizable workspace that integrates into your contact center platform to expand its capabilities and improve your interactions with customers. You can control and customize Agent Copilot's behavior entirely from the Flow in Cognigy.

Additionally, you can combine [Agent Copilot with Voice Gateway](/agent-copilot/getting-started/voice) to handle voice calls while still profiting from Agent Copilot's support.

## Core Components

### Flow

Agent Copilot uses specific Nodes from Flows to entirely control how the Agent Copilot workspace looks and behaves. Your Agent Copilot Flow serves only to define the Agent Copilot tiles and provides a separate environment from other Flows. In Agent Copilot Flows, you can perform knowledge searches, display customer information to human agents, and many other actions.

<Frame>
  <img alt="Agent Copilot Flow with Knowledge Tile and other Copilot Nodes" />
</Frame>

For more information, read the [Agent Copilot Nodes](/ai/agents/develop/node-reference/agent-copilot/overview) documentation.

### Backend

The Agent Copilot backend listens to the Agent Copilot Flow and handles the logic execution, API calls, and dynamic updates.

### Frontend

The frontend interface is a grid-like workspace that displays the configured tiles to human agents and updates in real time during conversations. You can run the frontend as a standalone app or embedded in your contact center platform.

#### Standalone

A separate app that you can open in your browser and that shows the full Agent Copilot workspace. The human agent receives a link to Agent Copilot in the conversation.

<Frame>
  <img alt="Agent Copilot standalone workspace in a browser" />
</Frame>

#### Embedded

The Agent Copilot tiles are displayed directly in your contact center platform via an iFrame.

<Frame>
  <img alt="Agent Copilot tiles embedded in a contact center platform" />
</Frame>

## Use Cases

| Use Case                                      | Agent Copilot Behavior                                                         |
| --------------------------------------------- | ------------------------------------------------------------------------------ |
| A customer mentions a billing problem.        | Retrieves the mentioned invoice and suggests resolution steps.                 |
| An agent needs to verify identity.            | Retrieves customer details from connected systems and displays to human agent. |
| A customer indicates frustration in the chat. | Displays customer history and user sentiment information.                      |
| A voice call is in progress.                  | Displays live transcript and next-step actions in real time.                   |

For more information, read the [Agent Copilot](/agent-copilot/overview) documentation.


# Handover Status
Source: https://docs.cognigy.com/ai/escalate/handover-features/handover-status



*Handover Status* lets you manage and respond to the state of conversations handed over to human agents. It provides real-time updates on handover progress, enabling you to design dynamic, context-aware conversational experiences.

When a conversation is handed over to a human agent, your AI Agent can react to the current handover status — guiding the Flow based on whether the conversation is queued, actively handled, completed, or has encountered an issue.

This capability is particularly valuable when building experiences with Agent Copilot Flows, as it allows you to deliver contextual messages and implement fallback handling tailored to the current handover state.

## Key Benefits

Using Handover Status in your Flows provides several benefits:

* **Improved user experience**. Keep end users informed about their status, estimated wait time, and position in the queue, while preventing confusion from duplicate chats.
* **Smarter automation**. Automatically route conversations based on real-time status, trigger fallback handling, or escalate when errors occur.
* **Context-aware communication**. Send tailored messages depending on whether a conversation is queued, active, completed, or in progress.
* **Flexible integration**. Compatible with multiple platforms, including Cognigy Live Agent, Genesys Cloud, and RingCentral Engage.

## How it Works

To react to handover statuses, use a Lookup Node directly after the Handover to Human Agent Node.
A Lookup Node creates different Flow paths based on a chosen operand.

The example shows how to structure your handover logic using a Lookup Node with Handover Status:

<Frame>
  <img alt="Lookup Node with Handover Status after a Handover to Human Agent Node" />
</Frame>

## Supported Statuses

The table lists all handover statuses you can configure in the Case Node following a Lookup Node.

| Status                       | Description                                                                                                                                                                                                                                                                                                                                                                                                                           | Restrictions                                                                                                                                                                                                       |
| ---------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| Completed                    | Conversations are resolved successfully. Agent assistance is provided.                                                                                                                                                                                                                                                                                                                                                                | –                                                                                                                                                                                                                  |
| Cancelled                    | Conversations are ended by users, indicating termination before resolution.                                                                                                                                                                                                                                                                                                                                                           | –                                                                                                                                                                                                                  |
| Error                        | Conversations are encountering issues. This could be technical, communicative, or due to misinformation. Human agents can step in to resolve.                                                                                                                                                                                                                                                                                         | –                                                                                                                                                                                                                  |
| Events                       | Conversations are triggered by specific events, such as important updates or time-sensitive actions that guide the flow forward.                                                                                                                                                                                                                                                                                                      | –                                                                                                                                                                                                                  |
| Active                       | Conversations are actively managed by a human agent.                                                                                                                                                                                                                                                                                                                                                                                  | –                                                                                                                                                                                                                  |
| Queue                        | Conversations are on hold, waiting for an agent. You can inform the user with a Say Node, for example: `Estimated time: {{input.data.request.estimatedWaitTime / (1000 * 60)}} seconds` and `Position: {{input.data.request.position}}`. If you use Agent Copilot, add a Handover Node, Lookup Node, and a Say Node in your Copilot Flow. In the Main Flow, add a Go To Node after the Queue Case Node to lead into the Copilot Flow. | Only available for:<ul><li>RingCentral Engage (with **Get Queue Updates** active)</li><li>Cognigy Live Agent</li><li>Salesforce MIAW (with **Get Queue Position** or **Get Estimated Wait Time** active)</li></ul> |
| Handover Already in Progress | Triggered when the same user ID opens multiple conversations. The older conversation becomes inactive and is linked to the newer one in the same agent chat window. Add a Say Node to notify the user that the older chat is now inactive.                                                                                                                                                                                            | Only available for **Genesys Cloud Messaging**                                                                                                                                                                     |
| genericHandoverUpdate        | Used for messages from the Genesys Bot. Normally routed to agents, but you can configure it to also show to end users.                                                                                                                                                                                                                                                                                                                | Only available for **Genesys Cloud Messaging**                                                                                                                                                                     |


# Real-Time Queue Notifications
Source: https://docs.cognigy.com/ai/escalate/handover-features/real-time-queue-notifications

Real-Time Queue Notifications provide end users with live updates on their queue position and estimated wait time, enhancing transparency and reducing anxiety during handovers to human agents.

<a href="/release-notes/2026.16"><Badge>Updated in 2026.16.0</Badge></a>

*Real-Time Queue Notifications* keeps end users informed about queue updates by providing real-time information on their position and estimated wait time. This feature helps reduce anxiety and creates transparency, making end users feel valued and respected.

By analyzing the data collected from the queue, you can identify patterns and trends for further analysis. For example, you might identify peak times when the queue is longest or determine the typical wait time for a response from a human agent.

## Key Benefits

* **Reduced User Anxiety**. End users receive continuous updates about their queue position and estimated wait time.
* **Increased Transparency**. Real-time visibility into queue status builds trust and sets clear expectations.
* **Alternative Messaging**. Automatically sends a customized message when queue conditions exceed defined thresholds.
* **Pattern Analysis**. Collected queue data helps identify peak times and typical response times.

## Prerequisites

* A [handover provider](/ai/escalate/handover-reference/overview) is configured. Real-time queue notifications are supported by [Cognigy Live Agent](/ai/escalate/handover-reference/live-agent), [Salesforce MIAW](/ai/escalate/handover-reference/salesforce-miaw), and [Genesys Cloud Open Messaging](/ai/escalate/handover-reference/genesys-cloud-open-messaging).
* For Genesys Cloud Open Messaging, an active [Amazon EventBridge Source integration](/ai/escalate/handover-reference/genesys-cloud-open-messaging-with-amazon-eventbridge) and admin access to Genesys Cloud to configure use and integration settings are required.
* The Endpoint that supports the selected handover provider is configured. For testing purposes, you can use the [Webchat v3 Endpoint](/webchat/v3/overview) with [Demo Webchat](/webchat/demo).

## How It Works

To set up Real-Time Queue Notifications, follow these steps:

1. Configure the following Nodes in your Flow:
   1. [Handover to Human Agent](#configure-settings-for-a-handover-to-human-agent-node) — enables queue position and wait time tracking.
   2. [Lookup](#add-a-lookup-node) — routes the Flow based on handover status.
   3. [If](#add-an-if-node) — evaluates queue conditions.
   4. [Say](#add-say-nodes) — displays queue updates or alternative messages to the user.
   5. [Go To](#add-a-go-to-node) — loops back to continue monitoring the queue.
2. Configure the provider-side settings to enable queue notifications for [Salesforce MIAW](#configure-settings-in-salesforce-miaw) or [Genesys Cloud Open Messaging](#configure-settings-in-genesys-cloud-open-messaging).
3. *(Optional)* Configure the [Agent Copilot workspace](/live-agent/assistants/agent-copilot).

## Configure Settings for a Handover to Human Agent Node

1. In the Flow editor, add the [Handover to Human Agent](/ai/agents/develop/node-reference/service/handover-to-agent) Node or use an existing one if you already added it earlier.
2. Open the Node editor by clicking the Handover to Human Agent Node.
3. From the **Handover Provider** list, select Cognigy Live Agent, Salesforce MIAW, or Genesys Cloud Open Messaging.
4. For the queue position, configure the following settings:
   * **Get Queue Position** — activate the setting to retrieve the end user's current queue position. This setting isn't available for Genesys Cloud Open Messaging, which doesn't provide a queue position. If you use Genesys Cloud Open Messaging, skip this step and go to step 5.
   * **Update Interval** — specify how often (in seconds) to receive updates on the queue status. For example, set the interval to `30` seconds to receive updates every half-minute. This setting is applicable only to Live Agent.
5. For estimated wait time, configure the following settings:
   * **Get Estimated Wait Time** — activate the setting to retrieve the estimated wait time for the end user.
   * **Update Interval** — specify how often (in seconds) to receive updates on the wait time. For example, set the interval to `30` seconds to receive updates every half-minute. This setting isn't available for Salesforce MIAW.
6. Configure alternative update options:
   * **Alternative Update** — activate the alternative update options that trigger when certain conditions are met, such as reaching the maximum queue position or maximum estimated wait time.
   * **Maximum Queue Position** — specify the maximum queue position. If you set this to `5`, the system will send an alternative message when the user's position exceeds 5. This setting isn't available for Genesys Cloud Open Messaging, which doesn't provide a queue position.
   * **Maximum Estimated Wait Time** — specify the maximum estimated wait time in milliseconds. Set this to `300000` milliseconds (5 minutes). If the estimated wait time exceeds this time, the system will send an alternative message from the **Alternative Text** field.
   * **Alternative Text** — specify the message that should be sent when either the maximum queue position or maximum wait time is reached. For example, `Thank you for your patience. We are experiencing higher than usual wait times. Please hold on, and an agent will assist you shortly.`
7. Click **Save Node**.

When the settings are configured, you can test this configuration by creating a handover via [Demo Webchat](/webchat/demo).
In the chat, as an end user, you will receive the message in the following format:

```json theme={null}
{
  "request": {
    "position": 3,
    "estimatedWaitTime": 300000
  }
}
```

To change the appearance of this message and make it more accessible, as well as to configure the **Alternative Text** message output, follow the instructions below.

## Add a Lookup Node

1. Below the Handover to Human Agent Node, add a [Lookup](/ai/agents/develop/node-reference/logic/lookup) Node.
2. Open the Node editor by clicking the Lookup Node.
3. From the **Type** list, select **Handover Status**.
4. Click **Save Node**.
5. In the Case Node of the Lookup Node, select **Queue** from the **Handover Status** list and click **Select**.
6. Click **Save Node**.
7. Remove the second Case Node.

## Add an If Node

1. Below the **Case** Node, add an [If](/ai/agents/develop/node-reference/logic/if) Node.

2. Open the If Node editor and add the following condition:

   ```txt theme={null}
   {{input.data.request.position}} < 2 && {{input.data.request.estimatedWaitTime / (1000 * 60)}} < 200
   ```

   where:

   * `{{input.data.request.position}} < 2` — checks if the customer's position in the queue is less than 2. If true, it means the customer is either in first place or second place in the queue.
   * `{{input.data.request.estimatedWaitTime / (1000 * 60)}} < 200` — calculates the estimated wait time in minutes. The estimated wait time is provided in milliseconds, dividing by `(1000 * 60)` converts it to minutes. It then checks if this wait time is less than 200 minutes. If true, it means the estimated wait time is manageable.

3. Click **Save Node**.

## Add Say Nodes

1. Below the condition **Then**, add a [Say](/ai/agents/develop/node-reference/basic/say) Node.

2. Open the Say Node editor and paste the following text into the **Text** field:

   ```txt theme={null}
   Position: {{input.data.request.position}}

   Estimated time: {{input.data.request.estimatedWaitTime / (1000 * 60)}} minutes
   ```

   where:

   * `input.data.request.position` — the user's current position in the queue.
   * `input.data.request.estimatedWaitTime / (1000 * 60)` minutes — the estimated wait time in minutes. The `estimatedWaitTime` is initially provided in milliseconds. It is divided by 1000 to convert it to seconds, and then by 60 to convert it to minutes.

3. Click **Save Node**.

4. Below the condition **Else**, add another Say Node and paste the following text into the **Text** field to display the text from the **Alternative Text** field of the Handover to Human Agent Node that [you configured earlier](#configure-settings-for-a-handover-to-human-agent-node):

   ```txt theme={null}
   {{input.handover.data.alternativeText}}
   ```

5. Click **Save Node**.

## Add a Go To Node

1. Below the Say Nodes, add a [Go To](/ai/agents/develop/node-reference/logic/go-to) Node.
2. In the Go To Node, select the current Flow from the **Select Flow** list and **Lookup** Node from the **Select Node** list.
3. In the **Advanced** section, select **Go to Node and wait for Input** from the **Execution Mode** list.
4. Click **Save Node**.

<Frame>
  <img alt="Flow configuration for real-time queue notifications with Say and Go To Nodes" />
</Frame>

## Configure Settings in Salesforce MIAW

If you use [Salesforce MIAW](/ai/escalate/handover-reference/salesforce-miaw) as your handover provider, enable queue notifications in the Salesforce settings:

1. In Salesforce, go to the **Setup** page, enter `Messaging Settings` in the **Quick Find** box, and click **Messaging Settings**. Alternatively, go to **Feature Settings > Service > Messaging > Messaging Settings** in the navigation menu.
2. In the **Channels** section, find the channel you want to configure, click the arrow-down icon, and select **Edit**.
3. Go to the **Queue Wait Time and Position** section.
4. From the **Notification** list, select one of the following options:
   * **Estimated Wait Time** — sends the estimated wait time to the end user.
   * **Queue Position** — sends the user's current position in the queue.
5. Click **Save**.

## Configure Settings in Genesys Cloud Open Messaging

If you use [Genesys Cloud Open Messaging](/ai/escalate/handover-reference/genesys-cloud-open-messaging) as your handover provider, real-time queue notifications rely on the [Amazon EventBridge integration](/ai/escalate/handover-reference/genesys-cloud-open-messaging-with-amazon-eventbridge). Complete the following configuration in Genesys Cloud.

<Note>
  Genesys Cloud Open Messaging doesn't provide a queue position. Estimated wait time is the only parameter available and is calculated internally by Genesys Cloud Open Messaging.
</Note>

<AccordionGroup>
  <Accordion title="1. Add the `user.start` Topic Filter to the Integration">
    Cognigy needs to receive the event indicating when a human agent accepts (answers) a conversation. This event is delivered through the Amazon EventBridge Source's event stream.

    1. In Genesys Cloud, go to **IT and Integrations > Integrations**.
    2. On the **Integrations** page, select the **Amazon EventBridge Source** that you created previously.
    3. Open the **Configuration** tab.
    4. Go to **Properties > Topic Filtering**.
    5. Select `v2.detail.events.conversation.{id}.user.start`.
    6. Click **Save and Activate**.

    For more information, see the [UserStartEvent documentation](https://developer.genesys.cloud/analyticsdatamanagement/analytics/detail/analytics-detail-events#userstartevent) in the Genesys Cloud Developer Center.
  </Accordion>

  <Accordion title="2. Set the Maximum Capacity">
    Estimated wait time is only calculated once an agent has reached maximum capacity and additional conversations are placed on hold.

    1. In Genesys Cloud, go to **Digital and Telephony > Interaction Assets > Utilization**.
    2. On the **Organization** tab, go to **Channels**.
    3. In the **Message** row, click **Edit** in the **Actions** column.
    4. Set **Maximum Capacity** to `1`, or the value needed for your setup.

    When a human agent reaches this capacity, any additional conversation is placed on hold, and Genesys begins polling and returning an estimated wait time for it.

    <Warning>
      A direct transfer between human agents overrides maximum capacity. Skipping the hold and estimated-wait-time behavior is a native Genesys Cloud Open Messaging capability and can't be changed in Cognigy
    </Warning>
  </Accordion>
</AccordionGroup>

## (Optional) Agent Copilot

If you use the [Agent Copilot workspace](/live-agent/assistants/agent-copilot) and Agent Copilot is active in your contact center, the Agent Copilot Flow needs additional configuration to send queue update messages.

To configure the Agent Copilot Flow for sending queue update messages, complete the following steps:

1. Add a Handover to Human Agent Node to the Agent Copilot Flow. Configure this Node the same way as the [Handover to Human Agent Node](#configure-settings-for-a-handover-to-human-agent-node) in your Main Flow.
2. Add a Lookup Node for the Queue status, followed by an If Node, Say Nodes, and a Go To Node. Mirror the logic you already configured in the Main Flow:
   * [Add a Lookup Node](#add-a-lookup-node)
   * [Add an If Node](#add-an-if-node)
   * [Add Say Nodes](#add-say-nodes)
   * [Add a Go To Node](#add-a-go-to-node). In the Agent Copilot Flow, point the Go To Node to the Agent Copilot Flow's own Lookup Node so the queue keeps updating.
3. In the Main Flow, below the [Case Node that checks the Queue status](#add-a-lookup-node), add a [Go To Node](/ai/agents/develop/node-reference/logic/go-to) that points to the Agent Copilot Flow's Lookup Node.

## More Information

* [Handover Status](/ai/escalate/handover-features/handover-status)
* [User Inactivity Detection and Handling](/ai/escalate/handover-features/user-inactivity-detection)
* [Lookup](/ai/agents/develop/node-reference/logic/lookup)
* [Agent Handover](/ai/agents/develop/node-reference/service/handover-to-agent)


# User Inactivity Detection and Handling
Source: https://docs.cognigy.com/ai/escalate/handover-features/user-inactivity-detection



<a href="/release-notes/4.53"><Badge>Updated in 4.53</Badge></a>

The *Inactivity Detection and Handling* feature detects user inactivity during handovers
and enables automated responses,
handover closures, or custom actions to optimize human agent efficiency.

## Key Benefits

* **Efficient Resource Use**. Automatically frees human agents from inactive conversations.
* **Improved User Experience**. Sends timely prompts or closes stalled handovers smoothly.

## Prerequisites

* A [handover provider](/ai/escalate/handover-reference/overview), for example, [Cognigy Live Agent](/ai/escalate/handover-reference/live-agent), is configured.
* The Endpoint that supports a selected handover provider is configured. For testing purposes, you can use the [Webchat v3 Endpoint](/webchat/v3/overview) with [Demo Webchat](/webchat/demo).

## Restrictions

* This feature triggers Flows via inactivity timeouts only when the handover status is `queue` or `active`. User inputs won't trigger Flows after a handover begins, except for [Agent Copilot](/agent-copilot/index) Flows.

## How to Use

To use this feature, add Nodes to your Flow, then test your Flow via Demo Webchat and any handover providers, for example, Cognigy Live Agent.

### How to Set Up

<AccordionGroup>
  <Accordion title="1. Add Handover to Human Agent Node">
    To ensure proper detection and handling of inactivity during queue time or active handovers, follow these steps:

    1. In the Flow editor, add the **Handover to Human Agent** Node.
    2. Open a Node editor by clicking the **Handover to Human Agent** Node, then go to the **Event Settings** section.
    3. Toggle on the **Send Queue Event** and **Send Active Event** settings:
       * **Send Queue Event** — triggers an event when the user enters the queue. This option allows tracking inactivity while waiting for a human agent.
       * **Send Active Event** — triggers an event when the human agent sends the first message. This option allows tracking inactivity during an active conversation.
    4. Save changes.

    <Frame>
      <img alt="Handover to Human Agent Node event settings with queue and active events" />
    </Frame>
  </Accordion>

  <Accordion title="2. Configure Lookup Node">
    Set up the Lookup Node with handover statuses:

    1. In the Flow editor, below the **Handover to Human Agent** Node, add the **Lookup** Node.
    2. Open a Node editor by clicking the Node, then go to the **Type** setting, select **Handover Status** from the list. Save changes.
       <Frame>
         <img alt="Lookup Node configured with Handover Status type" />
       </Frame>
    3. Go to the child Nodes:<br />
       1. Click the **Case** Node.<br />
       2. In the **Value** field, specify the `active` handover status if you want to detect and handle inactivity while the end users are in the queue. Save changes.<br />
       <Frame>
         <img alt="Case Node with active handover status value" />
       </Frame>
       3. Click the second **Case** Node.<br />
       4. In the **Value** field, specify the `queue` handover status if you want to detect and handle inactivity while the end users are in the queue.<br />
       5. Click **Save Node**.<br />
       <Frame>
         <img alt="Case Node with queue handover status value" />
       </Frame>
    4. Add **completed** and **error** child Nodes by analogy.
  </Accordion>

  <Accordion title="3. Set Inactivity Time">
    If you want to detect inactivity during the queue time, the first timer should be started once the end user enters the queue.

    To do this, add the Once Node to start the inactivity timer when you receive the queue event:

    1. Below the **queue** child Node, add **Once** Node.
    2. Below the **On First Time** child Node, add the **Set Handover Inactivity** Node.
    3. In the **Set Handover Inactivity** Node, specify the inactivity timeout in minutes, for example, `1`. On each incoming user input, the timer is automatically reset and starts from the beginning. Once the timer expires, an inactivity event is sent to the Flow. Save changes.
    4. Repeat the same steps for the **active** child Node in case you want to detect and handle inactivity during active handovers.

       <Frame>
         <img alt="Set Handover Inactivity Node with timeout configuration" />
       </Frame>

    The timer is set in minutes and may vary by up to one minute. If there is no end user input, the timer will expire after the specified time. If an end user input is received before the timeout, the timer will automatically reset and start from the beginning.
  </Accordion>

  <Accordion title="4. Loop Back">
    Set the **Go To** Node to return to the Lookup Node.
    This moves the entrypoint to the Lookup Node, allowing the Flow to execute the correct Lookup child branch based on incoming events.

    In the **Go To** Node editor, select **Lookup** from the **Select Node** list and **Go To Node and wait for Input** from the **Execution Mode** list in the **Advanced** section.

    <Frame>
      <img alt="Go To Node configured to return to the Lookup Node" />
    </Frame>
  </Accordion>

  <Accordion title="5. Handle Inactivity">
    An example of how detected inactivity could be handled is explained here:

    1. Add the **If** Node at the end of the Flow.
    2. In the Node editor, specify `input.handover.inactivityCount >= 2`. This allows handling the second detected inactivity differently from the first one. Note that the `input.handover.inactivityCount` is reset to 0 once the end user sends a message.
    3. Below the **If** Node and its **Else** child Node, do the following to ask the end user, based on detected inactivity, if they are still there.<br />
       1. Add the **Say** Node.<br />
       2. In the **Say** Node editor, in the text field, specify `Are you still there? The conversation will be automatically closed soon`. <br />
       3. In **Handover Settings > Handover Output Destination**, select **User** to only show the message to the user and not the human agent. Save changes.<br />
       4. Add the **Set Handover Inactivity** Node to start a subsequent inactivity timer.<br /> If the inactivity count is greater than two, it triggers the close handover process, which can close the handover on both the user and human agent sides.
    4. Below the **If** Node and its **Then** child Node, do the following to auto-close the handover after the inactivity timer expired the second time in direct succession:<br />
       1. Add the **Close Handover** Node to auto-close the handover.<br />
       2. Add the **Say** Node.<br />
       3. In the **Say** Node editor, in the text field, specify `You are back with the bot now!`.<br />
       4. In **Handover Settings > Handover Output Destination**, select **User and Agent**. Save changes.<br /> If the inactivity count is less than two, it prompts the user to respond to prevent the conversation from being automatically closed. The Set Handover Inactivity Node will set the time period for inactivity after which the conversation will be closed or restart a Flow from the beginning.

    <Frame>
      <img alt="If Node handling inactivity with Say and Close Handover Nodes" />
    </Frame>
  </Accordion>

  <Accordion title="6. Loop Back">
    Finally, add the **Go To** Node to return to the **Lookup** Node as you did before. In the **Go To** Node editor,
    select **Lookup** from the **Select Node** list
    and **Go To Node and wait for Input** from the **Execution Mode** list in the **Advanced** section.
  </Accordion>
</AccordionGroup>

### How to Test

1. Use Demo Webchat to initiate a handover.
2. Wait past the set timeout, for example, 1 minute, without responding.
3. Check for the `Are you still there?` message after the first timeout and handover closure after the second timeout.

## Example

Imagine a user in a queue stops responding.
After 1 minute,
they see
`Are you still there? The conversation will close soon`. If they remain inactive for another minute, the handover closes, and they get `You're back with the bot!`
— freeing the human agent and keeping the chat flowing.

## More Information

* [Set Handover Inactivity](/ai/agents/develop/node-reference/service/set-handover-inactivity)
* [Close Handover](/ai/agents/develop/node-reference/service/close-handover)
* [Handover Providers](/ai/escalate/handover-reference/overview)
* [Agent Handover](/ai/agents/develop/node-reference/service/handover-to-agent)
* [Real-Time Queue Notifications](/ai/escalate/handover-features/real-time-queue-notifications)


# 8x8
Source: https://docs.cognigy.com/ai/escalate/handover-reference/8x8



<a href="/release-notes/4.97"><Badge>Updated in 4.97</Badge></a>

<Frame>
  <img alt="8x8 handover provider logo" />
</Frame>

## Description

The 8x8 handover provider bridges Cognigy and [8x8](https://www.8x8.com/),
enabling end users to connect with human agents working in a contact center that uses 8x8.

## Prerequisites

* Access to [8x8](https://connect.8x8.com/login).
* Created Endpoint [compatible with 8x8](/ai/escalate/handovers#endpoints-compatible-with-handover-providers).

## Configuration on the Handover Provider Side

<Accordion title="Configure 8x8">
  1. Log in to the 8x8 Contact Center for admins.
  2. On the **My Applications** page, select **Configuration Manager for Contact Center**.
  3. In the left-side menu, click **Integration**, then go to the **Webhooks** tab.
  4. On the **Webhooks List** tab, click **+ Add webhook**.
  5. Fill in the following fields:
     * **Name** — specify a unique name.
     * **URL** — paste the URL in the following pattern: `https://endpoint-your-environment/handover/eightByEight`. For instance, in a Cognigy trial environment, the URL would be `https://endpoint-trial.cognigy.ai/handover/eightByEight`.
     * **Chat API Version** — select **Chat Gateway v1.0**.
  6. To check if the connection works, click **Test Connection** and save changes.
  7. In the left-side menu, click **Queues/Skills**.
  8. On the **Queue List** tab, create a queue with a chat type by clicking **+ Chat**.
  9. Fill in the following fields:
     * **Queue name** — specify a unique name.
     * **Queue description** — specify a unique description for the queue to differentiate it from others.
       Save changes.
  10. On the **Members** tab, assign and enable human agents who will operate within your queue to assist end users. Save changes.
  11. In the left-side menu, click **Channels**.
  12. On the **Chat API** tab, click **+ New Channel**.
  13. Fill in the following fields:
      * **Name** — specify a unique name.
      * **Description** — specify a description of your channel.
      * **Webhook** — select the webhook that you recently created.
      * **Queue** — select the queue that you recently created.
        Save changes.
</Accordion>

## Configuration on the Cognigy Side

<AccordionGroup>
  <Accordion title="1. Create a Handover Provider">
    1. Go to **Deploy > Handover Providers**.
    2. Click **+ New Handover Provider** and select **8x8** from the list.
    3. Fill in the following fields:
       * **Base URL** — the URL of the 8x8 Server.
       * **API Key** — the header value required for authenticating API calls to the 8x8 system.
       * **8x8-Tenant** — the Tenant ID, the field is mandatory if a customer has multiple tenants in the organization.
  </Accordion>

  <Accordion title="2. Configure Handover Settings">
    In the [Handover to Human Agent](/ai/agents/develop/node-reference/service/handover-to-agent) Node, configure the following settings:

    * **8x8 Channel Id** — specify the channel ID that you created during the 8x8 configuration. You can find this ID in the 8x8 admin console. Navigate to **Channels > Chat API**. In the channel list, find the channel that you created. Copy the **ID** from the channel row.
    * **8x8 Queue Id** — specify the header value required for authenticating API calls to the 8x8 system. You can find this ID in the 8x8 admin console. Navigate to **Channels > Chat API**. In the channel list, find the channel that you created. Copy the **Queue ID** from the channel row.
    * **8x8 JSON properties** — extend 8x8 properties with custom JSON.

    To check the handover, go to the installation that your server administrator has deployed.
  </Accordion>
</AccordionGroup>

## Additional Configuration

<Accordion title="Agent Copilot Workspace">
  Within 8x8 integration, you can use the [Agent Copilot workspace](/agent-copilot/overview) as an assistant to human agents. Agent Copilot will be available as a standalone application via the following link:

  `https://${AgentCopilotBaseUrl}/?userId=${userId}&sessionId=${sessionId}&URLToken=${URLToken}`

  Where:

  * `AICopilotBaseUrl` — the base URL used to access the Agent Copilot workspace.
  * `sessionId` — the Cognigy session ID. A session represents the interaction between the human agent and the Agent Copilot workspace.
  * `userId` — the Cognigy user ID. It helps Agent Copilot associate the interaction with a specific user.
  * `URLToken` — the Endpoint token from Cognigy. The Endpoint must be linked to the Handover Flow. To find this token, go to the Endpoint that you created previously. In the **Endpoint URL** field, copy the token after `https://endpoint-trial.cognigy.ai/`. For example, in the URL `https://endpoint-trial.cognigy.ai/f38791ae20d4961acf0e97d9f377c4fe3df92894e1eff1c7a774a8ed089a4590`, the token is `f38791ae20d4961acf0e97d9f377c4fe3df92894e1eff1c7a774a8ed089a4590`. Only authorized users can access the Agent Copilot system through this URL.

  Example URL:

  ```txt wrap theme={null}
  https://ai-copilot-trial.cognigy.ai/?sessionId=session123&userId=user123&URLToken=a4d5c86c98f27730311591f28d194510e05ffed30ca148e3344970defd418e7d
  ```
</Accordion>

## More Information

* [8x8 Endpoint](/ai/agents/deploy/endpoint-reference/8x8)
* [Agent Copilot Workspace](/agent-copilot/overview)
* [8x8 Documentation: Configuring the Agent Handover from Cognigy to 8x8](https://support.8x8.com/contact-center/integrations/Intelligent_Customer_Assistant/Configuring_the_Agent_Handover_from_Cognigy_to_8x8)


# Chatwoot
Source: https://docs.cognigy.com/ai/escalate/handover-reference/chatwoot



<a href="/release-notes/4.97"><Badge>Updated in 4.97</Badge></a>

<Frame>
  <img alt="Chatwoot handover provider logo" />
</Frame>

## Description

The Chatwoot handover provider bridges Cognigy and [Chatwoot](https://www.chatwoot.com/),
enabling end users to connect with human agents working in a contact center that uses Chatwoot.

Chatwoot integration allows users to receive text messages, files, videos (MP4), images (JPG, PNG), and GIFs.

Cognigy supports Chatwoot API v1.19.0, custom agent bot names and descriptions, full conversation history.

## Prerequisites

* Access to [Chatwoot](https://www.chatwoot.com/). You can also run Chatwoot on an [on-premises environment](https://www.chatwoot.com/docs/self-hosted).
* Created Endpoint [compatible with Chatwoot](/ai/escalate/handovers#endpoints-compatible-with-handover-providers).
* You need two IDs that are listed in the Chatwoot address line:
  * **Account ID** — a unique identifier assigned to you as the user.
  * **Inbox ID** — not unique and is created again for each channel.<br />
    Both IDs are required to use the new Inbox channel. You can find them in your Chatwoot URL. For example, `https://app.chatwoot.com/app/account/6607/inbox/5446`, where the Account ID is `6607` and the Inbox ID is `5446`.

## Configuration on the Handover Provider Side

<AccordionGroup>
  <Accordion title="1. Add Agents in Chatwoot">
    To enable agents to handle handovers, add them in your Chatwoot instance:

    1. Go to the Chatwoot home page.
    2. In the left-side **Settings** menu, select **Agents**.
    3. Add a new agent as needed.
  </Accordion>

  <Accordion title="2. Set Up a Webhook for Inbox">
    To configure Chatwoot to receive incoming messages via a webhook:

    1. On the Chatwoot main page, go to the left-side menu and click **Inboxes**.
    2. Click **Add Inbox** to start creating a new inbox.
    3. Select **API channel** and provide a channel name.
    4. In the **Webhook URL** field, enter your Cognigy Endpoint Base URL appended with `/handover/chatwoot/`, for example, `https://endpoint-trial.cognigy.com/handover/chatwoot`.
    5. Click **Create API Channel**.
    6. Assign an agent to the inbox by selecting them and clicking **Add agents**.

    <img alt="Chatwoot API channel creation with webhook URL configuration" />

    The new API channel will now appear in your Chatwoot Inbox list.
  </Accordion>
</AccordionGroup>

## Configuration on the Cognigy Side

<AccordionGroup>
  <Accordion title="1. Create a Handover Provider">
    To link your Cognigy Endpoint to Chatwoot:

    1. Go to **Deploy > Handover Providers**.
    2. Click **+ New Handover Provider** and select **Chatwoot** as the handover provider.
    3. Enter the **Base URL** for your Chatwoot instance, which is your Chatwoot installation URL extended with `/api/v1`, for example, `https://app.chatwoot.com/api/v1`. For on-premises setups, use your custom Chatwoot API URL.
    4. Provide the Chatwoot **Account ID**.
    5. Enter the Chatwoot **API Key**.
    6. Click **Save** to store the configuration.
    7. Test the setup by clicking **Open Webchat** and initiating a conversation that triggers a handover.
  </Accordion>

  <Accordion title="2. Add a Handover to Human Agent Node">
    To specify the Chatwoot Inbox for the handover:

    1. In the Flow editor, add a **Handover to Human Agent** Node.
    2. Open the Node editor and go to the **Chatwoot Settings** section.
    3. In the **Chatwoot Inbox Id** field, enter the Inbox ID. You can use [CognigyScript](/ai/platform-features/cognigyscript) to dynamically select an Inbox ID based on the conversation context.

    When the Handover Node is triggered during a conversation, the handover process will begin, and the user's message will appear in the specified Chatwoot Inbox. In Chatwoot, user messages are shown on the left, while agent responses appear on the right.

    <Note>
      Human agents can resolve a conversation once the user's request is addressed. If the user sends a new request, the human agent can reopen the conversation in Chatwoot to continue assisting.
    </Note>
  </Accordion>
</AccordionGroup>

## More Information

* [All Providers](/ai/escalate/handover-reference/overview)
* [Handover](/ai/escalate/handovers)


# Genesys Cloud Guest Chat
Source: https://docs.cognigy.com/ai/escalate/handover-reference/genesys-cloud-guest-chat

Connect Cognigy to Genesys Cloud using Genesys Cloud Guest Chat handover.

<a href="/release-notes/2026.4"><Badge>Updated in 2026.4</Badge></a>

<Frame>
  <img alt="Genesys Cloud Guest Chat handover provider logo" />

  <img alt="Genesys Cloud Guest Chat handover provider logo" />
</Frame>

## Description

<Warning>
  As of June 11, 2025, the Genesys Cloud Guest Chat handover is deprecated. This follows [Genesys ACD Web Chat v2 deprecation](https://help.mypurecloud.com/articles/deprecation-predictive-engagement-support-for-acd-web-chat-version-2/). If you're a new user looking to connect to Genesys, create a new [Genesys Cloud Open Messaging](/ai/escalate/handover-reference/genesys-cloud-open-messaging) handover.
  For existing users, we strongly advise migrating to the [Genesys Cloud Open Messaging](/ai/escalate/handover-reference/genesys-cloud-open-messaging) handover before the deprecation date.
</Warning>

The Genesys Cloud Guest Chat handover provider bridges Cognigy and [Genesys](https://login.mypurecloud.com/),
enabling end users to connect with human agents working in a contact center that uses Genesys Cloud Guest Chat.

The Genesys Cloud Guest Chat handover provider is based on [Genesys Guest Chat APIs](https://developer.genesys.cloud/commdigital/digital/webchat/guestchat). These APIs contain functionality to create a new chat conversation, retrieve previous chat history, send messages, set the typing indicator, and leave the chat. Receiving messages, membership changes, and other real-time events are available via a WebSocket connection.

## Prerequisites

* Access to [Genesys Cloud](https://login.mypurecloud.com/).
* Created Endpoint [compatible with Genesys Cloud Guest Chat](/ai/escalate/handovers#endpoints-compatible-with-handover-providers).

## Configuration on the Handover Provider Side

<AccordionGroup>
  <Accordion title="1. Create a Handover Provider">
    1. Go to **Deploy > Handover Providers**.
    2. Click **+ New Handover Provider** and select **Genesys Guest Chat** from the list.
    3. To fill in the required fields and complete your setup, read the [Genesys Cloud Handover Integration](https://support.cognigy.com/hc/en-us/articles/6229147435292) article in the Cognigy Help Center.
  </Accordion>

  <Accordion title="2. Configure Handover Settings">
    In [Handover to Human Agent](/ai/agents/develop/node-reference/service/handover-to-agent) Node, configure the following settings:

    * **Language** — specify a language for the conversation. For example, `english`, `spanish`, `german`.
    * **Skills** — define skills for the conversation. For example, `escalation`.
    * **Priority** — set the priority for the conversation. For example, `1`. If a priority is set, it triggers a flow in Genesys to prioritize or de-prioritize the conversation within the queue. Note that this functionality requires the appropriate flow to be set up in Genesys.
    * **Display Agent Details** — display the human agent's name and avatar from Genesys in the chat for the end user. The parameter is disabled by default.
    * **Custom Attributes** — add custom attributes in the `Key: Value` format, which allows you to include additional information.
      For example:

    ```json theme={null}
    {
        "section": "section A",
        "region": "Europe"
    }
    ```
  </Accordion>
</AccordionGroup>

## Additional Configuration

<Accordion title="Filter Transcript Messages">
  By default, Cognigy sends the full conversation transcript as a single message once the handover to Genesys occurs. Additionally, you can filter out empty or unsupported messages to keep the transcript relevant and concise:

  * **Filter unsupported messages** — if the system detects a message in an unsupported format, such as `UNSUPPORTED_STRUCTURE_DATA`, the system will exclude the message from the conversation transcript. To activate this feature, use the `FEATURE_FILTER_UNSUPPORTED_MESSAGES_HANDOVER_PROVIDER: "true"` feature flag.
  * **Filter empty messages** — if a user sends a blank message with no text or content, the system will exclude the message from the conversation transcript. To activate this feature, use the `FEATURE_FILTER_EMPTY_MESSAGES_HANDOVER_PROVIDER: "true"` feature flag.
</Accordion>

### Agent Copilot Workspace

Within Genesys integration, you can use the [Agent Copilot workspace](/agent-copilot/overview) as an assistant for your human agents.

By default, two versions of the application are provided:

<AccordionGroup>
  <Accordion title="Standalone">
    Agent Copilot will be available as a standalone application via the following link:

    `https://${AgentCopilotBaseUrl}/?userId=${userId}&sessionId=${sessionId}&URLToken=${URLToken}`

    Where:

    * `AICopilotBaseUrl` — the base URL used to access the Agent Copilot workspace.
    * `sessionId` — the Cognigy session ID. A session represents the interaction between the human agent and the Agent Copilot workspace.
    * `userId` — the Cognigy user ID. It helps Agent Copilot associate the interaction with a specific user.
    * `URLToken` — the Endpoint token from Cognigy. The Endpoint must be linked to the Handover Flow. To find this token, go to the Endpoint that you created previously. In the **Endpoint URL** field, copy the token after `https://endpoint-trial.cognigy.ai/`. For example, in the URL `https://endpoint-trial.cognigy.ai/f38791ae20d4961acf0e97d9f377c4fe3df92894e1eff1c7a774a8ed089a4590`, the token is `f38791ae20d4961acf0e97d9f377c4fe3df92894e1eff1c7a774a8ed089a4590`. Only authorized users can access the Agent Copilot system through this URL.

    Example URL:

    ```txt wrap theme={null}
    https://ai-copilot-trial.cognigy.ai/?sessionId=session123&userId=user123&URLToken=a4d5c86c98f27730311591f28d194510e05ffed30ca148e3344970defd418e7d
    ```
  </Accordion>

  <Accordion title="Embedded">
    By default, you can use the embedded version of the Agent Copilot workspace, but it will overlap part of the screen with the conversation. To place the workspace to the right of the chat with the conversation, follow these steps:

    <AccordionGroup>
      <Accordion title="1. Create a Script">
        To create a script, follow these steps:

        1. In Cognigy, go to Endpoint **Settings > Copilot**.
        2. Copy the Agent Copilot Embedding URL by clicking on it.
        3. In Genesys Cloud, open the **Admin** tab.
        4. Under **Contact Center**, click **Scripts**, then **Create**. Type a name for the script.
        5. Select the **Blank Script** template and click **+ Create**.
        6. On the script page, go to the **Add Components** tab.
        7. Under the **Components** section, select **Web Page**. In the **Web Page Source** field, enter the Agent Copilot Embedding URL that you copied before.
        8. In the **Layout** section, change the size of width and height by clicking the **Stretch** button.
        9. In the upper-right corner, go to the **Variables** tab.
        10. On the **Variables** tab, go to the **Basic Types** section and click **String**.
        11. Create a new variable:<br />
            1. In the **Name** field, enter `userId`.
            2. In the **Description** field, enter `Cognigy variable`.
            3. Enable the **Input** toggle and click **Apply**.
        12. In the upper-left corner of the script page, click **Script > Properties**.
        13. In the **Script Properties** section, activate the **Inbound** and **Message** features.
        14. In the upper-left corner of the script page, click **Script > Save**, then **Publish**.
      </Accordion>

      <Accordion title="2. Add the Script to the Inbound Message Flow">
        To add the Script to the Inbound Message Flow, follow these steps:

        1. In the Genesys Cloud interface, go to **Orchestration > Architect**.
        2. Click or hover over the <img alt="menu" /> icon on the **Flows** tab and select **Inbound Message**.
        3. From the flow list, select a flow that you created before.
        4. To configure a flow, click **Edit**.
        5. In the flow editor, ensure that the **Get Participant Data** action is added. If you haven't done so already, refer to steps 9-11 in the Create an Inbound Message Flow section of the [Genesys Cloud Open Messaging](/ai/escalate/handover-reference/genesys-cloud-open-messaging#3-create-an-inbound-message-flow) article.
        6. Below the **Get Participant Data** action, place a [Set Screen Pop](https://help.mypurecloud.com/articles/set-screen-pop-action/) action. The **Set Screen Pop** action must be positioned above the **Transfer to ACD** action.
        7. From the **Message Script** list of the **Set Screen Pop** action, select the script created before.
        8. In the **Inputs** action, add `userId` in the **userId** field.
        9. In the upper-left corner, click **Save**, then **Publish**.
      </Accordion>
    </AccordionGroup>
  </Accordion>
</AccordionGroup>

## More Information

* [Genesys Cloud Open Messaging](/ai/escalate/handover-reference/genesys-cloud-open-messaging)


# Genesys Cloud Open Messaging
Source: https://docs.cognigy.com/ai/escalate/handover-reference/genesys-cloud-open-messaging



<a href="/release-notes/2026.16"><Badge>Updated in 2026.16.0</Badge></a>

<Frame>
  <img alt="Genesys Cloud Open Messaging handover provider logo" />

  <img alt="Genesys Cloud Open Messaging handover provider logo" />
</Frame>

The Genesys Cloud Open Messaging handover provider bridges Cognigy and [Genesys](https://login.mypurecloud.com/), enabling end users to connect with human agents working in a contact center that uses Genesys Cloud CX.

The Genesys Cloud Open Messaging handover provider is based on [Open Messaging APIs](https://developer.genesys.cloud/commdigital/digital/openmessaging/openmessaging-apis). [Open messaging](https://developer.genesys.cloud/commdigital/digital/openmessaging/) facilitates integrations between Genesys Cloud and a third-party messaging service through a webhook.

## Prerequisites

* Access to [Genesys Cloud](https://login.mypurecloud.com/).
* Create an Endpoint [compatible with Genesys Cloud Open Messaging](/ai/escalate/handovers#endpoints-compatible-with-handover-providers).
* Activate Genesys Cloud Open Messaging on the Cognigy side:
  * If you have a SaaS installation, contact Cognigy technical support.
  * If you have an on-premises installation, specify the following feature flags: `FEATURE_ENABLE_GENESYS_CLOUD_OM="true"`, `FEATURE_ENABLE_GENESYS_CLOUD_OM_WHITELIST="organization1,organization2"`, `FEATURE_USE_GENESYS_BOT_CONNECTOR_ENDPOINT_WHITELIST="organization1,organization2"`. To enable the Genesys Cloud Open Messaging handover provider for all organizations in your installation, use `*` instead of a list of organizations.
* To detect handover completion in Genesys:
  * Starting from Cognigy 4.100, migrate to [Amazon EventBridge](/ai/escalate/handover-reference/genesys-cloud-open-messaging-with-amazon-eventbridge) using the Amazon EventBridge Genesys connector to send requests to Cognigy.
  * For Cognigy 4.99 and earlier versions, you have two options:
    * By default, Cognigy 4.99 and earlier use a WebSocket connection via the Notifications API to detect handover completion in Genesys. However, this approach has [limitations](https://developer.genesys.cloud/notificationsalerts/notifications/#usage-limitations) and may not always be reliable.
    * As an alternative, you can configure [HTTP webhooks](/ai/escalate/handover-reference/genesys-cloud-open-messaging-handover-end-detection) using Triggers and Web Services Data Actions to send requests to Cognigy.

## Configuration on the Handover Provider Side

Before starting the integration with Cognigy, build the Genesys Cloud Open Messaging configuration on the Genesys Cloud CX side.

<AccordionGroup>
  <Accordion title="1. Create a Platform">
    To create a platform config for Genesys Cloud Open Messaging, follow these steps:

    1. In the Genesys Cloud interface, click **Menu** in the upper-left corner and go to **Digital and Telephony > Message > Platform Configurations**.
    2. In the upper-right corner, click **+ Create Profile**.
    3. In the **Create a configuration profile** window, enter a unique name for the platform configuration and click **Create**.
    4. In the left-side menu, select **Platform Integrations**.
    5. On the **Messaging Platforms** page, click **+ Create New Integration** and select **Open Messaging**.
    6. On the **Open Messaging** page, fill in the following fields:
       * **Name** — enter a name without spaces for your integration. Copy and save this name. You need to specify this name in the **Deployment name** field on the Cognigy side.
       * **Outbound Notification Webhook URL** — enter `https://endpoint-<your-environment>/handover/genesysCloudOM`. For example, `https://endpoint-app.cognigy.ai/handover/genesysCloudOM`, where environment is `app.cognigy.ai`.
       * **Outbound Notification Webhook Signature Secret Token** — enter the secret into the X-Hub-Signature-256 header generation for webhook requests sent to the outbound notification webhook URL. For the secret, you can choose any arbitrary but sufficiently random string that you want. The external service should use the secret and signature to validate the message originating from Genesys Cloud. This validation is optional but recommended. For more information about validation, see [Validate webhook notifications in the Genesys Cloud Developer Center](https://developer.genesys.cloud/commdigital/digital/openmessaging/validate). Copy and save this token for future use in Cognigy. Note that if you don't copy and save this token, you will need to recreate it after saving the platform configuration.
    7. Click **Save**.
    8. From the **Platform Config** list, select the config that you created in Platform Configurations.
    9. From the **Supported Content Profile** list, select `default`. Save changes.

    Your Open Messaging platform will appear in the platform list.
  </Accordion>

  <Accordion title="2. Create a Queue">
    To create a Queue, follow these steps:

    1. In the left-side menu, select **User Management > Queues**.
    2. On the **Queues** page, click **+ Create Queue**. The **Create Queue** panel opens on the right side.
    3. In the **Create Queue** panel, fill in the following fields:
       * **Name** — enter a unique name of the queue. Save and copy this name for later use.
       * **Division** — select `Home`.
    4. Click **Save**. Your queue will appear in the queue list.
    5. Open the queue settings by selecting the queue from the list.
    6. In your browser's address bar, find and copy the queue ID from the URL. The ID is located between `/queues/` and `/general`. For example, in the URL, `https://apps.mypurecloud.de/directory/#/admin/organization/queues/d59d0280-6664-4896-ad42-1a2715b7178e/general`, copy the ID `d59d0280-6664-4896-ad42-1a2715b7178e`.
    7. Save the queue ID for later use in Cognigy.
  </Accordion>

  <Accordion title="3. Create an Inbound Message Flow">
    To create an Inbound Message flow, follow these steps:

    1. In the left-side menu, select **Orchestration > Architect**.

    2. Hover over the <img alt="menu" /> icon on the **Flows** tab and select **Inbound Message**.

    3. Click **+ Add** in the upper-left corner. The **Create 'Inbound Message Flow'** dialog box opens.

    4. In the **Name** field, enter a unique name for the inbound message flow.

    5. From the **Divisions** list, select the division to assign the flow to.

    6. Click **Create Flow**. The flow's configuration page opens.

    7. In the **Search Toolbox** field, enter `Send Response` and drag the action below the **Start** action in the messaging flow editor.

    8. In the **Message Body** field of the **Send Response** action, enter `Connected` and select **Literal** from the list next to the field.

    9. *(Optional)* Below the **Send Response** action, add the [Get Participant Data](https://help.mypurecloud.com/articles/get-participant-data-action/) action.

    10. *(Optional)* In the **Get Participant Data** editor, click **+** and add the following attributes:

        | Attribute | Name       | Value        |
        | --------- | ---------- | ------------ |
        | 1         | `Queue ID` | `queueId`    |
        | 2         | `Language` | `myLanguage` |
        | 3         | `Skills`   | `mySkills`   |
        | 4         | `User ID`  | `userId`     |

    11. *(Optional)* For debugging purposes, you can send this data to the human agent within a conversation. To do that, below the **Get Participant Data** action, add a second **Send Response** action. In the message body, enter the attributes in a [message sequence with the String Builder](https://help.mypurecloud.com/articles/set-up-a-message-sequence-with-the-string-builder/).

    12. Below the **Send Response** action, place the [Transfer to ACD](https://help.mypurecloud.com/articles/transfer-acd-action/) action to transfer an interaction to a queueing system.

    13. In the **Queue** field of the **Transfer to ACD** action, select the queue to which you want to transfer the interaction.

    14. In the upper-left corner, click **Save**, then **Publish**.

    After creating your inbound message flow, you will see this flow in the architect list.

    To learn more about designing the flow, refer to the [Configure Inbound Message Flow](https://help.mypurecloud.com/articles/inbound-message-flows/) documentation.
  </Accordion>

  <Accordion title="4. Set up Message Routing">
    To set up Message Routing, follow these steps:

    1. Go to the Genesys Cloud interface, click **Menu** in the upper-left corner and select **Orchestration > Routing > Message Routing**.
    2. In the upper-right corner, click **+ Attach New Addresses to a Flow**. The **Attach New Addresses** page opens.
    3. From the **Select Flow** list, select the Inbound Message Flow you created.
    4. From the **Select Addresses** section, choose the Open Messaging platform you created and click **Attach Address**. Save changes.

    Your Message Routing configuration will appear in the message routing list.
  </Accordion>

  <Accordion title="5. Configure Credentials">
    To configure credentials, follow these steps:

    1. In the left-side menu, select **IT and Integrations > OAuth**, then click **+ Add Client**.
    2. On the **Add New Client** page, configure the following:
       * **App Name** — enter a unique name for the client.
       * **Grant Types** — select **Client Credentials**. Click **Next**.
    3. In the **Roles** list, activate the corresponding role for the client. The role must include at least the following permissions:
       * `messaging-platform:readonly` (View messaging platform integrations)
       * `conversations` (Create, edit, and delete conversation data)
       * `analytics:readonly` (Query aggregate conversation data and view conversation details)
    4. Click **Next**.
    5. In the **Token Duration in seconds** field, enter the token expiration time. Click **Next**.
    6. Click **Generate New Client Secret**, then **Confirm**.
    7. Copy the Client ID and Client Secret, save them for future use, and click **Finish**. Confirm that you copied the Client ID and Client Secret in the dialog box.
    8. In the left-side menu, select **Authorized Applications**.
    9. In the upper-right corner, click **+ Authorize a Client**.
    10. In the **Authorize Client** window, enter the Client ID that you copied previously and click **Authorize Client**.
    11. (Optional) In the **Users that can use this application** section, select the roles of the users who can use this application.
    12. In the **Scope** section, select the minimum scope of the application, as listed in step 3. Click **Authorize**.

    Once your client is authorized, you can start configuring the Genesys Cloud Open Messaging handover provider on the Cognigy side.
  </Accordion>
</AccordionGroup>

## Configuration on the Cognigy Side

<AccordionGroup>
  <Accordion title="1. Create a Handover Provider">
    1. Go to **Deploy > Handover Providers**.
    2. Click **+ New Handover Provider** and select **Genesys Open Messaging** from the list.
    3. Scroll down to **Handover Settings** and select **Genesys Cloud Open Messaging** from the list.
    4. Fill in the following fields:
       * **Host** — enter the login URL for Genesys Cloud customers, which varies by region, such as `mypurecloud.de` for Germany. You can use the base domain like `mypurecloud.com` or `cac1.pure.cloud`, omitting the `apps.` or `login.` part. For more on Genesys Cloud regions, refer to [AWS regions for deployment](https://help.mypurecloud.com/articles/aws-regions-for-genesys-cloud-deployment/).
       * **Deployment Name** — specify the Open Messaging platform name that you created [on the Genesys Cloud CX side](#configuration-on-the-handover-provider-side).
       * **Queue ID** — enter the Queue ID that you copied when you created a queue [on the Genesys Cloud CX side](#configuration-on-the-handover-provider-side).
       * **Webhook Secret** — specify the Outbound Notification Webhook Signature Secret Token that you created [on the Genesys Cloud CX side](#configuration-on-the-handover-provider-side).
       * **Client ID** — enter the Client ID that you generated [on the Genesys Cloud CX side](#configuration-on-the-handover-provider-side).
       * **Client Secret** — enter the Client Secret that you generated [on the Genesys Cloud CX side](#configuration-on-the-handover-provider-side).
    5. *(Optional)* Activate the **Send Profile information** setting if you want to display human agent information, such as the first and last name, to the user. Save changes.
  </Accordion>

  <Accordion title="2. Configure Handover Settings">
    In the [Handover to Human Agent](/ai/agents/develop/node-reference/service/handover-to-agent) Node, configure the following settings:

    * **Language** — specify a language for the conversation. For example, `english`, `spanish`, `german`.
    * **Skills** — define skills for the conversation. For example, `escalation`.
    * **Priority** — set the priority for the conversation. For example, `1`. If a priority is set, it triggers a flow in Genesys to prioritize or de-prioritize the conversation within the queue. Note that this functionality requires the appropriate flow to be set up in Genesys.
    * **Enable User Connects Message** — notify human agents when an end user reconnects to the chat. The parameter is enabled by default. When the parameter is enabled, the message `User joined the conversation` appears in the chat as soon as the end user returns to the chat tab by clicking the **←** (back arrow) at the top bar in the browser, after having opened a new URL on the same tab as the chat.
    * **Enable User Disconnects Message** — notify human agents when an end user disconnects from the chat. The parameter is enabled by default. When the parameter is enabled, the message `User left the conversation` is sent as soon as the end user closes the tab with the chat or switches to a new URL address within the current tab.
    * **Display Agent Details** — display the human agent's name and avatar from Genesys in the chat for the end user. The parameter is disabled by default.
    * **Get Estimated Wait Time** — retrieve the estimated wait time for the end user. End users then receive wait time updates while they wait to be connected to a human agent. The parameter is inactive by default. It requires an active [Amazon EventBridge Source integration](/ai/escalate/handover-reference/genesys-cloud-open-messaging-with-amazon-eventbridge). Genesys Cloud Open Messaging doesn't provide a queue position. Estimated wait time is the only queue parameter available. For more information, read [Real-Time Queue Notifications](/ai/escalate/handover-features/real-time-queue-notifications#configure-settings-in-genesys-cloud-open-messaging).
    * **Custom Attributes** — add custom attributes, which allows you to include additional information. When sending custom attributes from Cognigy to Genesys Cloud, you can enter them in two ways:

    <CodeGroup>
      ```json JSON String theme={null}
      "{\"customerType\":\"premium\",\"tags\":[\"urgent\",\"vip\"]}"
      ```

      ```json JSON Object theme={null}
      {
        "customerType": "premium",
        "tags": ["urgent","vip"],
        "preferences": { "language": "en", "notifications": true }
      }
      ```
    </CodeGroup>

    Before sending to Genesys, Cognigy flattens nested objects, joins or indexes arrays, converts booleans to 1/0, and skips unsupported types.
    To test the connection, click **Open Demo Web Chat** in your Endpoint.
  </Accordion>
</AccordionGroup>

## Additional Configuration

<AccordionGroup>
  <Accordion title="Send Genesys Bot Messages to End Users">
    <Note>
      Before using this feature, add the `GENESYS_CLOUD_OM_HANDLE_BOT_MESSAGE: "true"` feature flag.
    </Note>

    By default, the Genesys Inbound Message flow routes messages to human agents only.
    You can configure your settings so that not only human agents but also end users receive these messages.
    Forwarding messages to the end user can be helpful in the following use cases:

    * When the conversation status or wait time information is relevant to the end user.
    * To provide the end user with updates and transparency throughout the interaction.
    * To allow the end user to make informed decisions while waiting for a human agent, such as requesting a callback.

    The Genesys Inbound flow is responsible for message configuration.
    However, if you want to use additional logic,
    such as allowing end users to see their queue position,
    set up the [In-Queue Message flow](https://help.mypurecloud.com/articles/work-with-in-queue-flows/) in Genesys in addition to the Genesys Inbound flow.

    Cognigy is responsible for message routing logic. To configure this logic, follow the instructions:

    1. In your handover Flow, set a **Lookup** Node below the **Handover to Human Agent** Node. Set the **Lookup** Node as your Entrypoint.
    2. For the **Type** field within the **Lookup** Node, select **Handover Status**.
    3. For the child **Case** Node, specify `genericHandoverUpdate` in the **Value** field.
    4. Add your **Say** Node under the **Case** Node to display the messages to the end user. Select **Text** from the **Output Type** list, and in the **Text** field enter the following **CognigyScript**: `{{ input.data.request.text }}`. The script will then query Genesys for the relevant data, such as a queue position.
    5. In the Handover Settings of the **Say** Node, select **User Only** from the **Handover Output Destination** list.
    6. To display all incoming Genesys Status or Bot messages, add a **Go To** Node below the **Say** Node.
    7. Open the **Go To** Node. From the **Select Node** list, choose **Lookup**. Scroll down to the **Advanced** section. From the **Execution Mode** list, select **Go to Node and wait for Input**.

    The main Flow on Cognigy should look like this:

    <Frame>
      <img alt="Cognigy Flow with Lookup and Go To Nodes for Genesys handover" />
    </Frame>

    <Accordion title="Filter Transcript Messages">
      By default, Cognigy sends the full conversation transcript as a single message once the handover to Genesys occurs. Additionally, you can filter out empty or unsupported messages to keep the transcript relevant and concise:

      * **Filter unsupported messages** — if the system detects a message in an unsupported format, such as `UNSUPPORTED_STRUCTURE_DATA`, the system will exclude the message from the conversation transcript. To activate this feature, use the `FEATURE_FILTER_UNSUPPORTED_MESSAGES_HANDOVER_PROVIDER: "true"` feature flag.
      * **Filter empty messages** — if a user sends a blank message with no text or content, the system will exclude the message from the conversation transcript. To activate this feature, use the `FEATURE_FILTER_EMPTY_MESSAGES_HANDOVER_PROVIDER: "true"` feature flag.
    </Accordion>
  </Accordion>

  <Accordion title="Configure Attachments">
    When using a [Webchat v3](/webchat/v3/configuration) or [WhatsApp](/ai/agents/deploy/endpoint-reference/whatsapp) Endpoint with the Genesys Cloud Open Messaging handover provider, end users can send and receive file attachments during a handover session.

    End users can upload, share, and receive the following file types:

    | Category  | Supported formats                |
    | --------- | -------------------------------- |
    | Documents | `.pdf`, `.docx`, `.pptx`, `.txt` |
    | Images    | `.jpeg`, `.jpg`, `.png`          |
    | Video     | `.mp4`, `.mov`                   |
    | Audio     | `.mp3`, `.ogg`                   |
    | Links     | Web links/URLs                   |

    To activate attachment uploads, configure [File Storage](/ai/agents/deploy/endpoints/file-storage) in the Webchat v3 or WhatsApp Endpoint settings.
  </Accordion>
</AccordionGroup>

### Agent Copilot Workspace

Within the Genesys integration, you can use the [Agent Copilot workspace](/agent-copilot/overview) as an assistant for your human agents.

By default, two versions of the application are provided:

<AccordionGroup>
  <Accordion title="Standalone">
    Agent Copilot will be available as a standalone application via the following link:

    `https://${AgentCopilotBaseUrl}/?userId=${userId}&sessionId=${sessionId}&URLToken=${URLToken}`

    Where:

    * `AICopilotBaseUrl` — the base URL used to access the Agent Copilot workspace.
    * `sessionId` — the Cognigy session ID. A session represents the interaction between the human agent and the Agent Copilot workspace.
    * `userId` — the Cognigy user ID. It helps Agent Copilot associate the interaction with a specific user.
    * `URLToken` — the Endpoint token from Cognigy. The Endpoint must be linked to the Handover Flow. To find this token, go to the Endpoint that you created previously. In the **Endpoint URL** field, copy the token after `https://endpoint-trial.cognigy.ai/`. For example, in the URL `https://endpoint-trial.cognigy.ai/f38791ae20d4961acf0e97d9f377c4fe3df92894e1eff1c7a774a8ed089a4590`, the token is `f38791ae20d4961acf0e97d9f377c4fe3df92894e1eff1c7a774a8ed089a4590`. Only authorized users can access the Agent Copilot system through this URL.

    Example URL:

    ```txt wrap theme={null}
    https://ai-copilot-trial.cognigy.ai/?sessionId=session123&userId=user123&URLToken=a4d5c86c98f27730311591f28d194510e05ffed30ca148e3344970defd418e7d
    ```
  </Accordion>

  <Accordion title="Embedded">
    By default, you can use the embedded version of the Agent Copilot workspace, but it will overlap part of the screen with the conversation. To place the workspace to the right of the chat with the conversation, follow these steps:

    <AccordionGroup>
      <Accordion title="1. Create a Script">
        To create a script, follow these steps:

        1. In Cognigy, go to Endpoint **Settings > Copilot**.
        2. Copy the Agent Copilot Embedding URL by clicking on it.
        3. In Genesys Cloud, open the **Admin** tab.
        4. Under **Contact Center**, click **Scripts**, then **Create**. Type a name for the script.
        5. Select the **Blank Script** template and click **+ Create**.
        6. On the script page, go to the **Add Components** tab.
        7. Under the **Components** section, select **Web Page**. In the **Web Page Source** field, enter the Agent Copilot Embedding URL that you copied before.
        8. In the **Layout** section, change the size of width and height by clicking the **Stretch** button.
        9. In the upper-right corner, go to the **Variables** tab.
        10. On the **Variables** tab, go to the **Basic Types** section and click **String**.
        11. Create a new variable:<br />
            1. In the **Name** field, enter `userId`.
            2. In the **Description** field, enter `Cognigy variable`.
            3. Enable the **Input** toggle and click **Apply**.
        12. In the upper-left corner of the script page, click **Script > Properties**.
        13. In the **Script Properties** section, activate the **Inbound** and **Message** features.
        14. In the upper-left corner of the script page, click **Script > Save**, then **Publish**.
      </Accordion>

      <Accordion title="2. Add the Script to the Inbound Message Flow">
        To add the Script to the Inbound Message Flow, follow these steps:

        1. In the Genesys Cloud interface, go to **Orchestration > Architect**.
        2. Click or hover over the <img alt="menu" /> icon on the **Flows** tab and select **Inbound Message**.
        3. From the flow list, select a flow that you created before.
        4. In the **Toolbox** section, enter `Get Participant Data` in the search field and drag the action to the arrow coming out of the **Start** Node in the messaging flow editor.
        5. In the flow editor, ensure that both the **Send Response** and the **Get Participant Data** actions have been added. If you haven't done so already, refer to steps 7-11 in the [Create an Inbound Message Flow section](#3-create-an-inbound-message-flow) of the handover reference.
        6. Below the **Get Participant Data** action, place a [Set Screen Pop](https://help.mypurecloud.com/articles/set-screen-pop-action/) action. The **Set Screen Pop** action must be positioned above the **Transfer to ACD** action.
        7. From the **Message Script** list of the **Set Screen Pop** action, select the script you previously created.
        8. In the **Inputs** action, enter `userId` in the **userId** field.
        9. In the upper-left corner, click **Save**, then **Publish**.
      </Accordion>
    </AccordionGroup>
  </Accordion>
</AccordionGroup>

## More Information

* [Genesys Cloud Guest Chat](/ai/escalate/handover-reference/genesys-cloud-guest-chat)
* [All Handover Providers](/ai/escalate/handover-reference/overview)
* [Handovers](/ai/escalate/handovers)


# Genesys Cloud Open Messaging: Handover End Detection with HTTP webhooks
Source: https://docs.cognigy.com/ai/escalate/handover-reference/genesys-cloud-open-messaging-handover-end-detection



<Warning>
  The handover end detection with HTTP webhooks applies to Cognigy 4.99 and earlier versions.
  If you use Cognigy 4.100 or later versions, migrate to [Amazon EventBridge](/ai/escalate/handover-reference/genesys-cloud-open-messaging-with-amazon-eventbridge). Remove the previous [configuration steps](#configuration-steps) after migration to Amazon EventBridge.

  In Cognigy 4.100, the handover end detection configuration with HTTP webhooks will continue to work only if it was set up before updating to Cognigy 4.100.
</Warning>

By default, to detect the end of a handover with Genesys Cloud Open Messaging, Cognigy establishes a WebSocket connection that subscribes to a channel created through the [Notifications API](https://developer.genesys.cloud/notificationsalerts/notifications/notifications-apis). However, this approach has [limitations](https://developer.genesys.cloud/notificationsalerts/notifications/#usage-limitations) and may not always be reliable.

As an alternative, you can use HTTP webhooks, which require creating triggers in a workflow and using the Web Services Data Actions extension on the Genesys platform to send requests to Cognigy. This approach ensures that events from ongoing conversations are reliably captured, especially during high volumes or complex flows with multiple queue transfers in Genesys.

In this guide, you will configure HTTP webhooks in Genesys Cloud Open Messaging for the following Genesys analytics events:

* [AfterCallWorkEvent](https://developer.genesys.cloud/analyticsdatamanagement/analytics/detail/analytics-detail-events#aftercallworkevent). The event is triggered when a human agent finishes after-call work by applying a wrap-up.
* [UserEndEvent](https://developer.genesys.cloud/analyticsdatamanagement/analytics/detail/analytics-detail-events#userendevent). The event is triggered when a user or human agent session ends, indicating that the interaction has been completed. Cognigy uses this event to detect when a session ends, which helps avoid prematurely wrapping up the chat in the middle of a transfer or after receiving an `AfterCallWorkEvent` event.

## Prerequisites

* [Configured Genesys Cloud Open Messaging](/ai/escalate/handover-reference/genesys-cloud-open-messaging).
* Applies to Cognigy 4.99 and earlier versions. If you use Cognigy 4.100 or later, migrate to [Amazon EventBridge](/ai/escalate/handover-reference/genesys-cloud-open-messaging-with-amazon-eventbridge).

## Configuration Steps

<AccordionGroup>
  <Accordion title="1. Install the Web Services Data Actions Extension">
    1. In the Genesys Cloud interface, click **Menu** in the upper-left corner and go to **IT and Integrations > Integrations**.
    2. Click **Add Integration**, find **Web Services Data Actions**, then click **Install**.
    3. Enter a name for your connection, for example, `Web Services Data Actions Integration`.
    4. Click **Save & Activate**.
  </Accordion>

  <Accordion title="2. Create an Action">
    1. In the left-side menu, select **Data Actions**.
    2. Click **+ Add Action** and select the integration you created.
    3. Enter a name for the action, for example, `Send Event to Cognigy`.
  </Accordion>

  <Accordion title="3. Configure the Action">
    1. On the **Setup** tab, go to the **Contracts** section.
    2. In **Input Contract > JSON**, select **JSON** and paste the following JSON:

    ```json theme={null}
    {
      "title": "properties",
      "type": "object",
      "properties": {
        "userId": {
          "type": "string"
        },
        "topic": {
          "type": "string"
        },
        "disconnectType": {
          "type": "string"
        }
      },
      "additionalProperties": true
    }
    ```

    4. In the **Configuration** section, follow these steps:
       1. Select **POST** from the **HTTP Method** list.
       2. In the **Request URL Template**, enter the Cognigy API URL in the following format: `https://endpoint-<cognigy-instance-domain>/handover/genesysCloudOM`. For example, `https://endpoint-trial.cognigy.ai/handover/genesysCloudOM`.
    5. Go to the **Test** tab, enter a value in the provided fields, for example `123`. Click **Run Test**.
    6. If the test is successful, click **Save & Publish**. Otherwise, review and correct the previous steps.
  </Accordion>

  <Accordion title="4. Create Workflows">
    1. Go to **Orchestration > Architect**, hover over the <img alt="menu" /> icon on the **Flows** tab and select **Workflow**.
    2. Create a new flow for the `AfterCallWorkEvent` event and name it `Workflow conversation.id.acw`.
    3. In the left-side menu, select **Resources > Data** and create the following String-type variables in the **Add Variable** section:
       * `addressTo`
       * `disconnectType`
    4. In the left-side menu, select **Initial State**, then click the **Data** section on the **Toolbox** tab. Drag **Call Data Action** to the desired location in the flow editor.
    5. Name your action and select **Web Services Data Action** from the **Category** list.
    6. From the **Data Action** list, select the action you created previously.
    7. Fill in the following fields:
       * `userId` — enter `Flow.addressTo` and select **Literal** from the list next to the field.
       * `topic` — enter `v2.detail.events.conversation.id.acw` and select **Literal** from the list next to the field.
       * `disconnectType` — enter `Flow.disconnectType` and select **Literal** from the list next to the field.
         Save and publish your flow.
    8. Create a new flow for the `UserEndEvent` event and repeat steps 1–7 with the following changes:
       1. Name the workflow `Workflow conversation.id.user.end`.
       2. For the `topic` field, select **Literal** and enter `v2.detail.events.conversation.id.user.end`.
  </Accordion>

  <Accordion title="5. Create Triggers">
    1. Go to **Orchestration > Triggers**.
    2. Create a trigger for the `AfterCallWorkEvent` event and name it `Trigger conversation.id.acw`.
    3. Configure the following fields:
       * **Topic Name** — select `v2.detail.events.conversation.{id}.acw`.
       * **Workflow Target** — select `Workflow conversation.id.acw`.
    4. Click **+ Add Condition** and define the criteria:
       * Criterion 1:
         * **JSON Path** — enter `mediaType`.
         * **Operator** — select **Equals (==)**.
         * **Value** — enter `MESSAGE`.
       * Criterion 2:
         * **JSON Path** — enter `messageType`.
         * **Operator** — select **Equals (==)**.
         * **Value** — enter `OPEN`.
    5. Click **Save** and toggle the **Inactive** option next to the trigger name to activate it.
    6. Create a new trigger for the `UserEndEvent` event and repeat steps 1–5 with these changes:
       * **Topic Name** — select `v2.detail.events.conversation.{id}.user.end`.
       * **Workflow Target** — select `Workflow conversation.id.user.end`.

    The integration is complete, ensuring more reliable detection of conversation endings.
  </Accordion>
</AccordionGroup>

## More Information

* [Genesys Cloud Open Messaging Documentation: Create a flow](https://help.mypurecloud.com/articles/create-call-flow/)
* [Genesys Cloud Open Messaging Documentation: Create a custom action for integrations](https://help.mypurecloud.com/articles/create-custom-action-integrations/)
* [Genesys Cloud Open Messaging Documentation: Add a data actions integration](https://help.mypurecloud.com/articles/add-a-data-actions-integration/)


# Genesys Cloud Open Messaging: Handover End Detection with Amazon EventBridge
Source: https://docs.cognigy.com/ai/escalate/handover-reference/genesys-cloud-open-messaging-with-amazon-eventbridge



<a href="/release-notes/2025.22"><Badge>Updated in 2025.22</Badge></a>

The Amazon EventBridge integration offers a more reliable and stable solution for event handling compared to WebSocket.
Unlike WebSocket, which can sometimes result in lost events due to connection instability, EventBridge ensures smooth and consistent event delivery, improving the reliability of the handover process between Cognigy and Genesys Cloud Open Messaging.

In this guide, you will configure Amazon EventBridge connection for the following Genesys analytics events:

<Tabs>
  <Tab title="Cognigy 2025.22 and later">
    * [CustomerEndEvent](https://developer.genesys.cloud/analyticsdatamanagement/analytics/detail/analytics-detail-events#customerendevent). The event is triggered when the customer session ends, for example, when customers leave the chat. This event allows Cognigy to detect when the customer finishes their interaction in Genesys Cloud Open Messaging. This approach ensures that Cognigy doesn't restart the conversation too early, especially if the human agent is still completing after-call tasks.
    * [UserStartEvent](https://developer.genesys.cloud/analyticsdatamanagement/analytics/detail/analytics-detail-events#userstartevent). This event is optional and can be added to display the human agent's image and name in the conversation. This event allows Cognigy to read the human agent's information each time a conversation starts or is transferred, so the human agent's details are always up to date.
  </Tab>

  <Tab title="Cognigy 2025.12-21">
    * [CustomerEndEvent](https://developer.genesys.cloud/analyticsdatamanagement/analytics/detail/analytics-detail-events#customerendevent). The event is triggered when the customer session ends, for example, when customers leave the chat. This event allows Cognigy to detect when the customer finishes their interaction in Genesys Cloud Open Messaging. This approach ensures that Cognigy doesn't restart the conversation too early, especially if the human agent is still completing after-call tasks.
  </Tab>

  <Tab title="Cognigy 2025.11 and earlier">
    * [AfterCallWorkEvent](https://developer.genesys.cloud/analyticsdatamanagement/analytics/detail/analytics-detail-events#aftercallworkevent). The event is triggered when a human agent finishes after-call work by applying a wrap-up.
    * [UserEndEvent](https://developer.genesys.cloud/analyticsdatamanagement/analytics/detail/analytics-detail-events#userendevent). The event is triggered when a user or human agent session ends, indicating that the interaction has been completed. Cognigy uses this event to detect when a session ends, which helps avoid prematurely wrapping up the chat in the middle of a transfer or after receiving an `AfterCallWorkEvent` event.

    The configuration depends on the Cognigy installation you have:

    * [SaaS](#cognigy-saas) – you need access to Genesys and must inform Cognigy technical support about the integration.
    * [On-premises](#cognigy-on-premises) – you need access to both Genesys and [Amazon EventBridge](https://aws.amazon.com/eventbridge/).
  </Tab>
</Tabs>

## Prerequisites

* [Configured Genesys Cloud Open Messaging](/ai/escalate/handover-reference/genesys-cloud-open-messaging).
* You use Cognigy 4.100 and later versions.
* If you want to migrate from [Handover End Detection with HTTP webhooks](/ai/escalate/handover-reference/genesys-cloud-open-messaging-handover-end-detection) to Amazon EventBridge, make sure you remove the [configuration steps](/ai/escalate/handover-reference/genesys-cloud-open-messaging-handover-end-detection#configuration-steps) after migration to Amazon EventBridge.
* For on-premises installations, activate the `FEATURE_USE_GENESYS_OM_HTTP_END_HANDOVER_REQUEST="true"` feature flag.
* Set the timeout to prevent queue transfers from triggering the end of Genesys Cloud Open Messaging handovers with the `GENESYS_OM_HANDOVER_ACW_AFTER_TRANSFER_TIMEOUT_IN_MS` environment variable. The default value is 10 seconds. This timeout is useful when multiple after-call work events are sent through the event bridge and cause handovers to end during queue transfers.
* If you have multiple organizations for one Genesys Cloud Open Messaging account, activate the `SKIP_GENESYS_EVENTBRIDGE_ORGANIZATION_ID_CHECK` feature flag to prevent the validation of the Cognigy organization ID during a handover.

## Restrictions

* The Amazon EventBridge integration for Genesys Cloud Open Messaging can't be set up in the [trial environment](https://trial.cognigy.ai/).

## Cognigy SaaS

<AccordionGroup>
  <Accordion title="1. Contact Cognigy Technical Support">
    1. Contact [Cognigy technical support](/help/get-help).
    2. Provide the `organisationId` for the Cognigy organization where you want to enable the Amazon EventBridge integration. Cognigy technical support will provide you with the AWS Account ID and AWS Account Region Code needed to activate the Amazon EventBridge Source. If you need to create two integrations (one for each organization), provide both `organisationId` values in the same ticket.
  </Accordion>

  <Accordion title="2. Configure Amazon EventBridge in Genesys">
    1. In the Genesys interface, navigate to **Admin > Integrations**, select **Amazon EventBridge Source**, and click **Install**.
    2. On the **Configuration** tab, enter the following information:

       * **AWS Account ID** – enter the 12-digit AWS Account ID provided by Cognigy technical support.
       * **AWS Account Region** – enter the AWS region provided by Cognigy technical support.
       * **Event Source Suffix** – enter the `organisationId` of your Cognigy organization.
       * **Topic Filtering** – add the following filters:

    <Tabs>
      <Tab title="Cognigy 2025.12-21">
        * `v2.detail.events.conversation.{id}.customer.end`
        * *(Optional)* `v2.detail.events.conversation.{id}.user.start`
      </Tab>

      <Tab title="Cognigy 2025.12-21">
        * `v2.detail.events.conversation.{id}.customer.end`
      </Tab>

      <Tab title="Cognigy 2025.11 and earlier">
        * `v2.detail.events.conversation.{id}.user.end`
        * `v2.detail.events.conversation.{id}.acw`
      </Tab>
    </Tabs>

    3. Click **Save**, go to the **Details** tab, and activate the integration.
  </Accordion>

  <Accordion title="3. Confirm the Integration Setup">
    1. Reply to the same support ticket to inform Cognigy technical support that the setup is complete.
    2. Cognigy technical support will verify the integration and contact you to confirm whether it was successful. This process takes up to 5 business days.
  </Accordion>
</AccordionGroup>

## Cognigy On-Premises

<AccordionGroup>
  <Accordion title="1. Configure Amazon EventBridge in Genesys">
    1. In the Genesys interface, navigate to **Admin > Integrations**, select **Amazon EventBridge Source**, and click **Install**.
    2. On the **Configuration** tab, fill in the following fields:
       * **AWS Account ID** – enter the 12-digit [AWS Account ID](https://docs.aws.amazon.com/accounts/latest/reference/manage-acct-identifiers.html) where the event source will be available for the event bus.
       * **AWS Account Region** – enter the [AWS region](https://docs.aws.amazon.com/eventbridge/latest/userguide/eb-saas.html) where the event source will be made available for the event bus. This parameter defines the geographical location of the AWS resources that will handle the event traffic.
       * **Event Source Suffix** – enter the `organisationId` of your Cognigy organization.
       * **Topic Filtering** – add the following filters:

    <Tabs>
      <Tab title="Cognigy 2025.12 and later">
        * `v2.detail.events.conversation.{id}.customer.end`
      </Tab>

      <Tab title="Cognigy 2025.11 and earlier">
        * `v2.detail.events.conversation.{id}.user.end`
        * `v2.detail.events.conversation.{id}.acw`

        3. Click **Save**, go to the **Details** tab, and activate the integration.
      </Tab>
    </Tabs>
  </Accordion>

  <Accordion title="2. Set up an EventBridge Flow">
    1. Go to [Amazon EventBridge](https://console.aws.amazon.com/events).
    2. In the left navigation pane, go to **Integration > API destinations**, then click **Create API destination**.
    3. On the **Create API destination** page, fill in the following fields:
       * **Name** – enter `service-handover-api-destination`.
       * **Endpoint** – enter `https://endpoint-<your-environment>/handover/genesysCloudOM`. For example, `https://endpoint-app.cognigy.ai/handover/genesysCloudOM`, where `your-environment` is `app.cognigy.ai`.
       * **HTTP Method** – select **POST**.
       * **Rate limit** – enter `300`.
    4. In the **Connection configuration** section, select **Create new connection**, then fill in the following fields:
       * **Name** – enter `service-handover-connection`.
       * **Description** – enter a meaningful description. This parameter is optional.
       * **Authorization Type** – select **API Key**.
       * **API Key Name** – enter `Authorization`.
       * **API Key Value** – this value will be stored in your Kubernetes Secret. Use a securely generated random alphanumeric string (23 characters), base64-encoded for enhanced security.
    5. Save changes by clicking **Create**.
  </Accordion>

  <Accordion title="3. Create a Custom Event Bus">
    1. In the left navigation pane of the Amazon EventBridge console, go to **Integration > Partner event sources**.
    2. On the **Partner event sources** page, you will see a partner event source prefixed with `aws.partner/genesys.com/cloud` and suffixed with your Cognigy `organisationId`. If it's missing, the Genesys setup was incorrect. Delete the EventBridge integration in Genesys and create it again.
    3. In the top-right corner, click **Associate with event bus** and confirm the integration.
  </Accordion>

  <Accordion title="4. Create a Rule">
    1. In the left navigation pane of the Amazon EventBridge console, go to **Buses > Event buses**.
    2. Select your custom event bus under **Custom event buses**.
    3. On the **Rules** tab, click **Create rule**.
    4. At the **Define rule detail** step, enter a rule name and its description. Make sure the correct event bus and event pattern are preselected.
    5. At the **Build event pattern** step, in the **Event source** section, select **Other**, then enter the following JSON in the **Event pattern** field:
       ```json theme={null}
       {
       "detail": {
           "eventBody": {
           "mediaType": ["MESSAGE"],
           "messageType": ["OPEN"]
           }
         }
       }
       ```
    6. At the **Select target(s)** step, proceed as follows:<br />
       6.1 Select **EventBridge API destination** in the **Target types** section.<br />
       6.2 In the **API destination** section, select **Use an existing API destination**.<br />
       6.3 In the **Execution role** section, create a role that uniquely identifies this integration, for example, `Amazon_EventBridge_Invoke_API_Destination_Genesys_Integration`.
    7. Skip the Configure tags step. At the **Review and create** step, complete the configuration.
  </Accordion>

  <Accordion title="5. Store your API Key as a Secret">
    You must expose the EventBridge API key to the Cognigy handover service.
    Run the following command by replacing `<your-key>` with the actual API key:

    ```bash theme={null}
    kubectl create secret generic genesys-handover-secret \
    --from-literal=GENESYS_EVENTBRIDGE_HANDOVER_API_KEY=<your-key>
    ```

    Your Cognigy deployment must read this secret and inject the value into the service via environment variables:

    ```bash theme={null}
    env:
      - name: GENESYS_EVENTBRIDGE_HANDOVER_API_KEY
        valueFrom:
          secretKeyRef:
            name: genesys-handover-secret
            key: GENESYS_EVENTBRIDGE_HANDOVER_API_KEY
    ```

    Now, your integration is complete.
  </Accordion>
</AccordionGroup>

## More Information

* [Genesys Cloud Open Messaging Documentation: About the Amazon EventBridge integration](https://help.mypurecloud.com/articles/about-the-amazon-eventbridge-integration/)
* [Genesys Cloud Open Messaging Documentation: Configure the Amazon EventBridge integration](https://help.mypurecloud.com/articles/configure-the-amazon-eventbridge-integration/)


# Handover End Detection
Source: https://docs.cognigy.com/ai/escalate/handover-reference/handover-end-detection

Configure how Genesys Cloud Open Messaging detects when a handover session ends.

Genesys Cloud Open Messaging supports two methods for detecting when a handover session ends.

<CardGroup>
  <Card title="HTTP Webhooks" href="/ai/escalate/handover-reference/genesys-cloud-open-messaging-handover-end-detection">
    Detect handover end events in Genesys Cloud Open Messaging using HTTP webhooks.
  </Card>

  <Card title="Amazon EventBridge" href="/ai/escalate/handover-reference/genesys-cloud-open-messaging-with-amazon-eventbridge">
    Detect handover end events in Genesys Cloud Open Messaging using Amazon EventBridge.
  </Card>
</CardGroup>


# Live Agent
Source: https://docs.cognigy.com/ai/escalate/handover-reference/live-agent



<a href="/release-notes/4.97"><Badge>Updated in 4.97</Badge></a>

The Live Agent handover provider bridges Cognigy with the native contact center [Cognigy Live Agent](/live-agent/index),
enabling end users to connect with human agents working within the Cognigy Live Agent contact center.

## Key Benefits

* **Deep Cognigy Integration**. Offers superior compatibility with Cognigy compared to other handover services, creating a cohesive ecosystem for collaboration between AI Agents and human agents.
* **Rapid Deployment**. Allows you to configure a Live Agent Inbox within a Cognigy Project quickly, reducing setup time and technical overhead.
* **AI-Powered Support for Human Agents**. Features an AI Agent that helps human agents in real time, improving response quality and reducing agent workload.
* **User-Friendly Administration**. Offers an intuitive interface for creating and managing accounts, as well as adding and overseeing human agents with minimal complexity.

## Get Started with Cognigy Live Agent

To get started with Cognigy Live Agent, go to the [Getting Started](/live-agent/getting-started/overview) guide.

## More Information

* [All Handover Providers](/ai/escalate/handover-reference/overview)
* [Handovers](/ai/escalate/handovers)


# NiCE CXone (Integrated)
Source: https://docs.cognigy.com/ai/escalate/handover-reference/nice-cxone/nice-cxone-integrated

Integration with NiCE CXone for handing over conversations to human agents in environments that are pre-integrated with NiCE CXone.

<a href="/release-notes/2026.16"><Badge>Updated in 2026.16</Badge></a>

<Frame>
  <img alt="NiCE CXone handover provider logo" />

  <img alt="NiCE CXone handover provider logo" />
</Frame>

## Description

The NiCE CXone (Integrated) handover provider bridges Cognigy and [NiCE CXone](https://www.nice.com/), enabling end users to connect with human agents in a NICE CXone contact center.

NiCE CXone (Integrated) is for Cognigy installations that are deployed as part of the NiCE CXone platform. Both systems share the same infrastructure, so the handover provider handles authentication and channel configuration automatically.

## Prerequisites

* Access to [NiCE CXone](https://cxone.niceincontact.com/).
* Cognigy installation is integrated with [NiCE CXone environment](/ai/administer/access/management-ui#create-an-organization). If your Cognigy installation isn't integrated with NiCE CXone at the platform level, use the [NiCE CXone (Non-Integrated)](/ai/escalate/handover-reference/nice-cxone/nice-cxone-non-integrated) handover provider instead.

## Restrictions

* The [Transfer Node](/ai/agents/develop/node-reference/voice/generic/transfer) isn't supported.
* If your NiCE CXone Studio script is configured to send a message to the user after a handover is initiated, the **Repeat Handover Accepted Message** setting in the [Handover to Human Agent](/ai/agents/develop/node-reference/service/handover-to-agent) Node won't work. The message sent by the Studio script is treated as a human agent response, which prevents the **Handover Accepted Message** from repeating.
* Configuring the estimated wait time logic isn't supported on the Cognigy side for the NiCE CXone integration. Instead, add an [Estimated Wait Time](https://help.nicecxone.com/content/studio/actionscx/ewtestimate/ewtestimate.htm?cshid=EWTESTIMATE) action to your NiCE CXone Studio script.

- After importing a NICE CXone (Integrated) handover provider from a Package or Snapshot:
  * In NiCE CXone, update the Studio script if you customized it.
  * [Reconfigure on the NiCE CXone side](#configuration-on-the-handover-provider-side) because this configuration isn't included in the import.

## Configuration on the Cognigy Side

<AccordionGroup>
  <Accordion title="1. Create a Handover Provider">
    1. Go to **Deploy > Handover Providers**.
    2. Click **+ New Handover Provider** and select **NiCE CXone**.
    3. Enter a unique name for the provider. This name will be used as a name for the channel in NiCE CXone.
    4. Save the configuration.

    After creating a handover provider, you'll see one of the following statuses:

    * **Pending** or **Syncing** — no action needed. Wait for the process to complete.
    * **Failed** — review the [prerequisites](#prerequisites) and try again. If the issue persists, contact [Cognigy technical support](/help/get-help).
    * **Ready** — you can continue the setup on the NiCE CXone side.

    Once the status is **Ready**, a Studio script is automatically created in NiCE CXone with the ID in the format `Cognigy\Handover from Cognigy <cognigy-handover-providerId>`, for example, `Cognigy\Handover from Cognigy 6a462ef9109c0c66eb873918`. Use this ID to identify the script that maps to the handover provider you created.
  </Accordion>

  <Accordion title="2. Add a Handover to Human Agent Node">
    1. In the Flow editor, add a [Handover to Human Agent](/ai/agents/develop/node-reference/service/handover-to-agent) Node.
    2. From the **Handover Provider** list, select the NiCE CXone provider that you created.
    3. Configure the generic handover behavior, such as accepted messages or cancellation handling, as needed.
    4. *(Optional)* Configure additional settings:
       * **Send Transcript as first message** - allow NiCE CXone to receive the conversation transcript as the first message when the handover begins. This setting is turned on by default.
       * **Display Agent Details** - display the human agent's name and avatar from NiCE CXone in the chat for the end user. This setting is turned off by default.
       * **Custom Attributes** — add custom attributes, which allows you to include additional information. When sending custom attributes from Cognigy to NiCE CXone, you can enter them in two ways:

    <CodeGroup>
      ```json JSON String theme={null}
      "{\"customerType\":\"premium\",\"tags\":[\"urgent\",\"vip\"]}"
      ```

      ```json JSON Object theme={null}
      {
        "customerType": "premium",
        "tags": ["urgent","vip"],
        "preferences": { "language": "en", "notifications": true }
      }
      ```
    </CodeGroup>

    <Note>
      No NiCE CXone-specific fields are required in the Node. Routing to the correct queue or human agent team is configured on the NiCE CXone side.
    </Note>
  </Accordion>
</AccordionGroup>

## Configuration on the Handover Provider Side

<AccordionGroup>
  <Accordion title="1. Configure Handover Permissions">
    1. In NiCE CXone, click the **Applications** app selector in the upper-left corner.
    2. In the left-side **CXone Mpower Hubs** menu, go to **Omnichannel Routing > ACD > Roles**.
    3. Select the role assigned to your account, for example, **Administrator**.
    4. In **Assigned Channels**, select the channel with the same name as the NiCE CXone handover provider you created in Cognigy.
    5. Click **Update Role Permissions** and assign either all permissions or the minimum scope:
       * `Send Replies`
       * `Assign / Reassign Conversations`
       * `Reassign Conversations`
         Save your changes.
  </Accordion>

  <Accordion title="2. Create a Skill">
    To ensure that conversations are automatically assigned to human agents, create a skill and assign it to the channel and human agent.

    1. In NiCE CXone, click the **Applications** app selector in the upper-left corner.
    2. In the left-side **CXone Mpower Hubs** menu, go to **Omnichannel Routing > ACD > ACD Skills**.
    3. Click **Create New > Single Skill**.
    4. Configure the following fields:

    * **Media Type** — select **Digital**.
    * **Skill Name** — enter a name for your skill.
    * **Inbound/Outbound** — select **Inbound**.
    * **Campaign** — select any existing campaign.

    5. Click **Create**. After creating the skill, assign it to the channel:
       1. In the left-side **CXone Mpower Hubs** menu, go to **Omnichannel Routing > ACD > Digital > Points of Contact Digital**.
       2. Find the channel with the same name as your Cognigy handover provider. Click **Edit**, go to the **Skills** tab, and select the created skill from the **Select default skill** list.
    6. Verify the skill is assigned to a human agent:
       1. In the left-side **CXone Mpower Hubs** menu, go to **Omnichannel Routing > ACD > ACD Users**.
       2. Select the user from the list, and check the **Skills** tab to ensure the required skill is assigned.
  </Accordion>

  <Accordion title="3. Review the Studio Script">
    1. Click the **Applications** app selector in the upper-left corner. In the left-side **CXone Mpower Hubs** menu, go to **Omnichannel Routing > Studio**.
    2. Open the **Cognigy** folder. Inside, locate the script named **Cognigy Handover Script `<cognigy-handover-providerId>`**, where `<cognigy-handover-providerId>` is the ID of the handover provider in Cognigy.
       Don't rename the folder because Cognigy references the script by its original name when deleting the provider.

    You can adjust the script later using the [NiCE CXone Studio Scripts](https://help.nicecxone.com/content/studio/fundamentals/workwithscriptscx.htm?Highlight=scripts) documentation. In the script, make sure not to change the `ESCALATE` and `END ESCALATION` actions because Cognigy uses them to start and end handovers.

    The `ONRELEASE` action notifies Cognigy that the human agent has closed the conversation. Cognigy then triggers the **On Resolve Options** setting configured in the [Handover to Human Agent](/ai/agents/develop/node-reference/service/handover-to-agent) Node.
  </Accordion>

  <Accordion title="4. Test the Handover">
    Use a Cognigy [Webchat v3 Endpoint](/webchat/v3/configuration) for quick testing.

    1. In Cognigy, go to the Webchat v3 Endpoint settings. Open **Demo Webchat** and trigger the handover.
    2. In NiCE CXone, click the **Applications** app selector in the upper-left corner.
    3. In the left-side **CXone Mpower Hubs** menu, go to **Omnichannel Routing > Agent**.
    4. In the **Agent** window, select your preferred option for how you want to connect. For example, select **Set Phone Number**, enter a phone number, and click **Launch**.
    5. With the [correct skill assigned](#create-a-skill), the conversation will appear in the agent's queue. Otherwise, you can select **Search** in the left-side menu and find the recent conversation that was created.
       1. Verify that the conversation appears in NiCE CXone. Confirm that it can be picked up by a human agent.
       2. Send a reply from NiCE CXone and verify that it appears in the **Demo Webchat**.
  </Accordion>
</AccordionGroup>

## Additional Configuration

<Accordion title="Configure Attachments">
  When using a [Webchat v3](/webchat/v3/configuration) or [WhatsApp](/ai/agents/deploy/endpoint-reference/whatsapp) Endpoint with the NiCE CXone handover provider, end users can send and receive file attachments during a handover session.

  End users can upload, share, and receive the following file types:

  | Category  | Supported formats                |
  | --------- | -------------------------------- |
  | Documents | `.pdf`, `.docx`, `.pptx`, `.txt` |
  | Images    | `.jpeg`, `.jpg`, `.png`          |
  | Video     | `.mp4`, `.mov`                   |
  | Audio     | `.mp3`, `.ogg`                   |
  | Links     | Web links/URLs                   |

  To activate attachment uploads, configure [File Storage](/ai/agents/deploy/endpoints/file-storage) in the Webchat v3 or WhatsApp Endpoint settings.
</Accordion>

## More Information

* [NiCE CXone (Non-Integrated) Handover Provider](/ai/escalate/handover-reference/nice-cxone/nice-cxone-non-integrated)
* [NiCE CXone Handover Provider Overview](/ai/escalate/handover-reference/nice-cxone/overview)
* [All Handover Providers](/ai/escalate/handover-reference/overview)
* [Handovers](/ai/escalate/handovers)
* [NiCE CXone Endpoint](/ai/agents/deploy/endpoint-reference/nice)


# NiCE CXone (Non-Integrated)
Source: https://docs.cognigy.com/ai/escalate/handover-reference/nice-cxone/nice-cxone-non-integrated

Integration with NiCE CXone for handing over conversations to human agents in environments that are not pre-integrated with NiCE CXone.

<a href="/release-notes/2026.16"><Badge>Updated in 2026.16</Badge></a>

<Frame>
  <img alt="NiCE CXone handover provider logo" />

  <img alt="NiCE CXone handover provider logo" />
</Frame>

## Description

The NiCE CXone (Non-Integrated) handover provider bridges Cognigy and [NiCE CXone](https://www.nice.com/), enabling end users to connect with human agents in a NICE CXone contact center.

NiCE CXone (Non-Integrated) is for Cognigy installations that are deployed independently from NiCE CXone. The two systems run on separate infrastructure, so you must configure a BYOC (Bring Your Own Channel) channel and provide NiCE CXone credentials manually.

## Prerequisites

* Access to [NiCE CXone](https://cxone.niceincontact.com/).
* Cognigy installation isn't integrated with [NiCE CXone environment](/ai/administer/access/management-ui#create-an-organization). If your Cognigy installation is already integrated with NiCE CXone at the platform level, use the standard [NiCE CXone (Integrated) handover provider](/ai/escalate/handover-reference/nice-cxone/nice-cxone-integrated).
* An Administrator role in NiCE CXone to access the required configuration settings and create API credentials.
* A Client ID and Client Secret obtained by submitting the [CXone Application Registration](https://forms.microsoft.com/Pages/ResponsePage.aspx?id=vdojcYcOqU2cubfsggEarfHlkRVlgSlMjqsp52ASGttUM1U2VTQ4N0pXNk5XSzROV0VNNkhBU1gzMC4u) form. NiCE sends the credentials to your registered technical contacts within 2–3 business days. You need these credentials later in Cognigy.

## Restrictions

* The [Transfer Node](/ai/agents/develop/node-reference/voice/generic/transfer) isn't supported.
* If your NiCE CXone Studio script is configured to send a message to the user after a handover is initiated, the **Repeat Handover Accepted Message** setting in the [Handover to Human Agent](/ai/agents/develop/node-reference/service/handover-to-agent) Node won't work. The message sent by the Studio script is treated as a human agent response, which prevents the **Handover Accepted Message** from repeating.
* Configuring the estimated wait time logic isn't supported on the Cognigy side for the NiCE CXone integration. Instead, add an [Estimated Wait Time](https://help.nicecxone.com/content/studio/actionscx/ewtestimate/ewtestimate.htm?cshid=EWTESTIMATE) action to your NiCE CXone Studio script.

- After importing a NICE CXone (Non-Integrated) handover provider from a Package or Snapshot, repeat [Create and Configure a Handover Provider](#1-create-and-configure-a-handover-provider) starting from step 4, then follow the steps in [Configuration on the NICE CXone side](#configuration-on-the-handover-provider-side). This configuration isn't included in the import.

## Configuration on the Cognigy Side

<AccordionGroup>
  <Accordion title="1. Create and Configure a Handover Provider">
    1. In Cognigy, go to **Deploy > Handover Providers**.
    2. Click **+ New Handover Provider** and select **NiCE CXone**.
    3. Enter a unique name for the provider. This name will be used as the channel name in NiCE CXone.
    4. Click **Save**. The **Handover Provider Editor** page opens. In the browser address bar, locate the URL and copy the handover provider ID after `/handover-provider/`. For example, in the following URL: `https://trial.cognigy.ai/project/65ae8e3adc2d77e6e260aa37/69b009f72e3deaa53a602d66/handover-provider/69f9bba8844ecf15d07954f9`, the handover provider ID is `69f9bba8844ecf15d07954f9`. Copy and save this value. You need it when configuring the [Channel Integration ID](#channel-integration-id) field.
    5. Configure the following fields:

    <AccordionGroup>
      <Accordion title="Tenant ID">
        The unique identifier of your NiCE CXone instance.

        1. In NiCE CXone, click the **Applications** app selector in the upper-left corner.
        2. In the left-side **CXone Mpower Hubs** menu, go to **General > Admin**.
        3. Open the upper-right menu, click **My Profile And More** and select **My Profile**.
        4. On the **General** tab, copy the value from the **Tenant ID** field and enter it in Cognigy.
      </Accordion>

      <Accordion title="Business Unit ID">
        The identifier of your NiCE CXone business unit.

        1. In NiCE CXone, click the **Applications** app selector in the upper-left corner.
        2. In the left-side **CXone Mpower Hubs** menu, go to **Omnichannel Routing > ACD > ACD Configuration > Business Units**.
        3. On the **Business Unit details** page, copy the value from the **Unit Number** field and enter it in Cognigy.
      </Accordion>

      <Accordion title="NiCE CXone Authentication Outbound">
        Credentials for authenticating outbound API calls from Cognigy to NiCE CXone. Configure the following parameters:

        * **Name** — a unique name for your connection.
        * **Access Key ID** — the Access Key ID for your NiCE CXone user:
          1. In NiCE CXone, click the **Applications** app selector in the upper-left corner.
          2. In the left-side **CXone Mpower Hubs** menu, go to **General > Admin**.
          3. Open the employee profile used for API access.
          4. On the **Security** tab, click **Add Access Key**.
          5. Click **Show Secret Key** and immediately copy the Access Key ID and Access Key Secret to a secure location. The Access Key Secret is shown only once.
        * **Access Key Secret** — the Access Key Secret copied in the same step as the Access Key ID.
        * **Client ID** — the Client ID received from NiCE CXone as part of [Prerequisites](#prerequisites).
        * **Client Secret** — the Client Secret received from NiCE CXone as part of [Prerequisites](#prerequisites).
      </Accordion>

      <Accordion title="Channel Integration ID">
        The identifier of the channel integration in NiCE CXone.

        1. In NiCE CXone, click the **Applications** app selector in the upper-left corner.
        2. In the left-side **CXone Mpower Hubs** menu, go to **Omnichannel Routing > ACD**.
        3. Go to **Digital > Bring Your Own Channel** and click **+ Add**.
        4. Configure the following fields:
           <Note>
             Replace `<cognigy-instance-domain>` in the URL with the domain of your Cognigy instance. For example, if your instance URL is `https://trial.cognigy.ai`, replace `<cognigy-instance-domain>` with `trial`.
           </Note>
           * **Name** — a name for the integration, for example, `Cognigy Integration`.
           * **Integration box URL** — `https://endpoint-<cognigy-instance-domain>.cognigy.ai/handover/niceCxOne`.
           * **Authorization URL** — `https://endpoint-<cognigy-instance-domain>.cognigy.ai/handover/niceCxOne/{handover-provider-id}/auth/token`, where `{handover-provider-id}` is the handover provider ID copied earlier.
           * **Add action URL** — `https://endpoint-<cognigy-instance-domain>.cognigy.ai/handover/niceCxOne/add`.
           * **Reconnect action URL** — `https://endpoint-<cognigy-instance-domain>.cognigy.ai/handover/niceCxOne/reconnect`.
           * **Remove action URL** — `https://endpoint-<cognigy-instance-domain>.cognigy.ai/handover/niceCxOne/remove`.
           * **Client ID** — a Client ID of your choice, for example, `nice-cxone-dev-tech-writers`. Save this value. You need it for the **Client ID** in [Channel Integration Authentication](#channel-integration-authentication).
           * **Client Secret** — a 32-character string. You can generate this value using any secure random string generator. Save this value. You need it for the **Client Secret** in [Channel Integration Authentication](#channel-integration-authentication).
        5. Click **Save**. Copy the generated **Integration ID** and enter it in Cognigy.
      </Accordion>

      <Accordion title="Channel Integration Authentication">
        Credentials for authenticating the channel integration. Configure the following parameters:

        * **Name** — a unique name for your connection.
        * **Client ID** — the Client ID you entered when creating the [channel integration in NiCE CXone](#channel-integration-id).
        * **Client Secret** — the Client Secret you entered when creating the [channel integration in NiCE CXone](#channel-integration-id).
      </Accordion>
    </AccordionGroup>

    6. Click **Save**. A channel is created on the NiCE CXone side.
    7. Verify the created channel in NiCE CXone:
       1. In NiCE CXone, go to **Omnichannel Routing > ACD**, then **Digital > Points of Contact Digital**.
       2. Find the channel with the same name as your NiCE CXone handover provider in Cognigy. If you don't find the channel, the connection wasn't established. Check the handover provider configuration to ensure that all credentials are valid.
  </Accordion>

  <Accordion title="2. Add a Handover to Human Agent Node">
    1. In the Flow editor, add a [Handover to Human Agent](/ai/agents/develop/node-reference/service/handover-to-agent) Node.
    2. From the **Handover Provider** list, select the NiCE CXone provider that you created.
    3. Configure the handover behavior, such as accepted messages or cancellation handling, as needed.
    4. *(Optional)* Configure additional settings:

       * **Send Transcript as first message** - allow NiCE CXone to receive the conversation transcript as the first message when the handover begins. This setting is turned on by default.
       * **Display Agent Details** - display the human agent's name and avatar from NiCE CXone in the chat for the end user. This setting is turned off by default.
       * **Custom Attributes** — add custom attributes, which allows you to include additional information. When sending custom attributes from Cognigy to NiCE CXone, you can enter them in two ways:

       <CodeGroup>
         ```json JSON String theme={null}
         "{\"customerType\":\"premium\",\"tags\":[\"urgent\",\"vip\"]}"
         ```

         ```json JSON Object theme={null}
         {
           "customerType": "premium",
           "tags": ["urgent","vip"],
           "preferences": { "language": "en", "notifications": true }
         }
         ```
       </CodeGroup>

    <Note>
      No NiCE CXone-specific fields are required in the Node. Routing to the correct queue or human agent team is configured on the NiCE CXone side.
    </Note>
  </Accordion>
</AccordionGroup>

## Configuration on the NiCE CXone Side

<AccordionGroup>
  <Accordion title="1. Configure Handover Permissions">
    1. In NiCE CXone, click the **Applications** app selector in the upper-left corner.
    2. In the left-side **CXone Mpower Hubs** menu, go to **Omnichannel Routing > ACD > Roles**.
    3. Select the role assigned to your account, for example, **Administrator**.
    4. In **Assigned Channels**, select the channel with the same name as the NiCE CXone handover provider you created in Cognigy.
    5. Click **Update Role Permissions** and assign either all permissions or the minimum scope:
       * `Send Replies`
       * `Assign / Reassign Conversations`
       * `Reassign Conversations`
         Save your changes.
  </Accordion>

  <Accordion title="2. Create a Skill">
    To ensure that conversations are automatically assigned to human agents, create a skill and assign it to the channel and human agent.

    1. In NiCE CXone, click the **Applications** app selector in the upper-left corner.
    2. In the left-side **CXone Mpower Hubs** menu, go to **Omnichannel Routing > ACD > ACD Skills**.
    3. Click **Create New > Single Skill**.
    4. Configure the following fields:

    * **Media Type** — select **Digital**.
    * **Skill Name** — enter a name for your skill.
    * **Inbound/Outbound** — select **Inbound**.
    * **Campaign** — select any existing campaign.

    5. Click **Create**. After creating the skill, assign it to the channel:
       1. In the left-side **CXone Mpower Hubs** menu, go to **Omnichannel Routing > ACD > Digital > Points of Contact Digital**.
       2. Find the channel with the same name as your Cognigy handover provider. Click **Edit**, go to the **Skills** tab, and select the created skill from the **Select default skill** list.
    6. Verify the skill is assigned to a human agent:
       1. In the left-side **CXone Mpower Hubs** menu, go to **Omnichannel Routing > ACD > ACD Users**.
       2. Select the user from the list, and check the **Skills** tab to ensure the required skill is assigned.
  </Accordion>

  <Accordion title="3. Review the Studio Script">
    1. Click the **Applications** app selector in the upper-left corner. In the left-side **CXone Mpower Hubs** menu, go to **Omnichannel Routing > Studio**.
    2. Open the **Cognigy** folder. Inside, locate the script named **Cognigy Handover Script `<cognigy-handover-providerId>`**, where `<cognigy-handover-providerId>` is the ID of the handover provider in Cognigy.
       Don't rename the folder because Cognigy references the script by its original name when deleting the provider.

    You can adjust the script later using the [NiCE CXone Studio Scripts](https://help.nicecxone.com/content/studio/fundamentals/workwithscriptscx.htm?Highlight=scripts) documentation. In the script, make sure not to change the `ESCALATE` and `END ESCALATION` actions because Cognigy uses them to start and end handovers.

    The `ONRELEASE` action notifies Cognigy that the human agent has closed the conversation. Cognigy then triggers the **On Resolve Options** setting configured in the [Handover to Human Agent](/ai/agents/develop/node-reference/service/handover-to-agent) Node.
  </Accordion>

  <Accordion title="4. Test the Handover">
    Use a Cognigy [Webchat v3 Endpoint](/webchat/v3/configuration) for quick testing.

    1. In Cognigy, go to the Webchat v3 Endpoint settings. Open **Demo Webchat** and trigger the handover.
    2. In NiCE CXone, click the **Applications** app selector in the upper-left corner.
    3. In the left-side **CXone Mpower Hubs** menu, go to **Omnichannel Routing > Agent**.
    4. In the **Agent** window, select your preferred option for how you want to connect. For example, select **Set Phone Number**, enter a phone number, and click **Launch**.
    5. With the [correct skill assigned](#create-a-skill), the conversation will appear in the agent's queue. Otherwise, you can select **Search** in the left-side menu and find the recent conversation that was created.
       1. Verify that the conversation appears in NiCE CXone. Confirm that it can be picked up by a human agent.
       2. Send a reply from NiCE CXone and verify that it appears in the **Demo Webchat**.
  </Accordion>
</AccordionGroup>

## Additional Configuration

<Accordion title="Configure Attachments">
  When using a [Webchat v3](/webchat/v3/configuration) or [WhatsApp](/ai/agents/deploy/endpoint-reference/whatsapp) Endpoint with the NiCE CXone handover provider, end users can send and receive file attachments during a handover session.

  End users can upload, share, and receive the following file types:

  | Category  | Supported formats                |
  | --------- | -------------------------------- |
  | Documents | `.pdf`, `.docx`, `.pptx`, `.txt` |
  | Images    | `.jpeg`, `.jpg`, `.png`          |
  | Video     | `.mp4`, `.mov`                   |
  | Audio     | `.mp3`, `.ogg`                   |
  | Links     | Web links/URLs                   |

  To activate attachment uploads, configure [File Storage](/ai/agents/deploy/endpoints/file-storage) in the Webchat v3 or WhatsApp Endpoint settings.
</Accordion>

## Troubleshooting

| Issue                                           | Cause                                      | Recommended Action                                                                                                                                    |
| ----------------------------------------------- | ------------------------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------- |
| The **Save** button remains active after saving | NiCE CXone authentication or timeout issue | Verify the NiCE CXone authentication outbound credentials or try clicking the **Save** button again later because NiCE may experience timeout issues. |

## More Information

* [NiCE CXone (Integrated) Handover Provider](/ai/escalate/handover-reference/nice-cxone/nice-cxone-integrated)
* [NiCE CXone Handover Provider Overview](/ai/escalate/handover-reference/nice-cxone/overview)
* [All Handover Providers](/ai/escalate/handover-reference/overview)
* [Handovers](/ai/escalate/handovers)
* [NiCE CXone Endpoint](/ai/agents/deploy/endpoint-reference/nice)


# NiCE CXone
Source: https://docs.cognigy.com/ai/escalate/handover-reference/nice-cxone/overview

Integration with NiCE CXone for handover of conversations to human agents in a contact center.

The NiCE CXone handover providers connect Cognigy with [NiCE CXone](https://www.nice.com/), enabling end users to reach human agents in a NiCE CXone contact center.

Cognigy supports two NiCE CXone handover providers depending on your installation's integration with NiCE CXone at the platform level. Only one option is available at a time: either **NiCE CXone (Integrated)** or **NiCE CXone (Non-Integrated)**. Follow the setup instructions for the relevant provider:

<CardGroup>
  <Card title="NiCE CXone (Integrated)" href="/ai/escalate/handover-reference/nice-cxone/nice-cxone-integrated">
    For Cognigy installations with platform-level integration with NiCE CXone.

    * No additional credentials required.
    * A NiCE CXone channel is created automatically through the platform-level integration.
  </Card>

  <Card title="NiCE CXone (Non-Integrated)" href="/ai/escalate/handover-reference/nice-cxone/nice-cxone-non-integrated">
    For Cognigy installations without platform-level integration with NiCE CXone.

    * Requires Tenant ID, Business Unit ID, and API credentials from NiCE CXone.
    * A BYOC (Bring Your Own Channel) channel is created in NiCE CXone.
  </Card>
</CardGroup>


# Overview
Source: https://docs.cognigy.com/ai/escalate/handover-reference/overview

Connect AI Agents to contact centers using Cognigy handover providers.

*Handover Providers* connect AI Agents to contact centers, enabling seamless handoffs from automated conversations to human agents.

Select the connector for the contact center your human agents use.

<CardGroup>
  <Card title="NiCE CXone" href="/ai/escalate/handover-reference/nice-cxone/overview">
    Integration with NiCE CXone for environments integrated at the platform level or without platform-level integration.
  </Card>

  <Card title="Cognigy Live Agent" href="/ai/escalate/handover-reference/live-agent">
    Integration with the Cognigy-native contact center for handover of chat conversations.
  </Card>

  <Card title="Genesys Cloud Open Messaging" href="/ai/escalate/handover-reference/genesys-cloud-open-messaging">
    Integration with Genesys Cloud Open Messaging.
  </Card>

  <Card title="Salesforce MIAW" href="/ai/escalate/handover-reference/salesforce-miaw">
    Integration with Salesforce MIAW.
  </Card>

  <Card title="RingCentral Engage" href="/ai/escalate/handover-reference/ring-central-engage">
    Integration with RingCentral Engage.
  </Card>

  <Card title="8x8" href="/ai/escalate/handover-reference/8x8">
    Integration with the 8x8 platform.
  </Card>

  <Card title="storm" href="/ai/escalate/handover-reference/storm">
    Integration with storm, the contact center platform from Content Guru.
  </Card>

  <Card title="Chatwoot" href="/ai/escalate/handover-reference/chatwoot">
    Integration with the Chatwoot platform.
  </Card>

  <Card title="Genesys Cloud Guest Chat" href="/ai/escalate/handover-reference/genesys-cloud-guest-chat">
    Integration with Genesys Cloud for guest chat functionality.
  </Card>

  <Card title="Salesforce Service Cloud" href="/ai/escalate/handover-reference/salesforce">
    Integration with Salesforce Service Cloud.
  </Card>
</CardGroup>


# RingCentral Engage Integration
Source: https://docs.cognigy.com/ai/escalate/handover-reference/ring-central-engage



<a href="/release-notes/4.97"><Badge>Updated in 4.97.0</Badge></a>

<Frame>
  <img alt="RingCentral Engage handover provider logo" />
</Frame>

## Description

The RingCentral Engage handover provider bridges Cognigy and [RingCentral Engage](https://engage.ringcentral.com/),
enabling end users to connect with human agents working in a contact center that uses RingCentral Engage.

## Prerequisites

* Access to [RingCentral Engage](https://login.ringcentral.com/)
* Created Endpoint [compatible with RingCentral Engage](/ai/escalate/handovers#endpoints-compatible-with-handover-providers).
* Configuration steps must be performed by a user with the `Admin` role in RingCentral Engage.
* A community must already exist within the RingCentral Engage platform.

## Configuration on the Handover Provider Side

<AccordionGroup>
  <Accordion title="1. Create API Token">
    1. Go to the **ADMIN** section in the top bar of the RingCentral Engage platform.
    2. Select **API access tokens** from the left sidebar.
    3. Click **+** in the top-right corner to create a new API Access Token.
    4. Select an agent to handle user chats.
    5. Copy the token text displayed next to the **API Access Token** name. Save it for later use.
  </Accordion>

  <Accordion title="2. Create the Categories">
    1. In the **ADMIN** section, go to **Categories** in the left sidebar.
    2. Click the **+** button to create a new category, for example, `Cognigy`, leave other fields as default, and save changes.
    3. Create two subcategories, `Bot` and `Agent`, with `Cognigy` as the parent category.
    4. Double-click `Bot` or `Agent` to open their configuration pages. Extract the `CATEGORY_ID` from the URL, for example, `xxxxx.engagement.dimelo.com/admin/categories/CATEGORY_ID/edit`.
       These are needed for the **Bot Category Id** and **Agent Category Id** fields in Cognigy Handover Settings.
  </Accordion>

  <Accordion title="3. Create the Source">
    1. Navigate to **Sources** in the **ADMIN** section sidebar.
    2. Create a new source of type `Engage Digital Source SDK`.
    3. Set the **Base URI** field to `<endpoint-url>/handover/rce`. For example, `https://endpoint-dev.cognigy.ai/handover/rce`.
    4. In **Message Settings**, set the default category to the **Bot** category created earlier.
    5. Copy the **API access token** field value from the source into the **Realtime Access Token** field in Cognigy Handover Settings.
    6. Copy the **Realtime Endpoint URL** from the **Implementation Details** section into the corresponding field in Cognigy Handover Settings.
  </Accordion>

  <Accordion title="4. Configure Agent Permissions">
    1. In the **Agents** overview, click the key icon for each agent who will use the source.
    2. Grant `read` and `reply` permissions for the source.
  </Accordion>

  <Accordion title="5. Create the Webhook">
    1. Go to **Webhooks** in the **ADMIN** section sidebar.
    2. Click the **+** button to create a new webhook.
    3. In the **SETTINGS** section:
       1. Select the previously created API Access Token.
       2. Set **Active** to **on**.
       3. Set **URL** to `<endpoint-url>/handover/rce`, for example, `https://endpoint-demo.cognigy.ai/handover/rce`.
       4. Enter a randomly generated value in the **Secret** field. Copy this value for the **Webhook Secret** field in Cognigy.
    4. In the **FILTERS** section:
       1. Set **Sources filtering strategy** to **No sources except**.
       2. Select the previously created source.
    5. In the **SUBSCRIPTIONS** section, add the **intervention.closed** event.
    6. Update the **Verification Token** to match your Cognigy environment's token. By default, the value is `cognigy123`. Save changes.
  </Accordion>

  <Accordion title="6. Get the Base API URL">
    Build the Base API URL by appending `api` to your RingCentral Engage environment name, for example `cognigy.engagement.dimelo.com` becomes `cognigy.api.engagement.dimelo.com`.
    Save it for later use as the **Base API URL** in Cognigy.
  </Accordion>

  <Accordion title="7. Create Bot and Agent Identities">
    1. Go to the **Identities** page in RingCentral Engage.
    2. Locate the Bot identity, check **Controlled identity**, and save. This ensures Bot messages appear on the right side.
    3. Locate the Agent identity, make it controlled, and assign it to agents via the **Agents** view by selecting the community and Agent user and saving it.
  </Accordion>

  <Accordion title="8. Set Up Folders for Chatting">
    1. Go to **Desks > Folders** in RingCentral Engage.
    2. Create a folder filtering chats with the **Agent** category for easy agent access.
  </Accordion>
</AccordionGroup>

## Configuration on the Cognigy Side

<AccordionGroup>
  <Accordion title="1. Create a Handover Provider">
    <Note>
      This feature is in beta. We encourage you to try it out and provide us with feedback.
    </Note>

    1. Go to **Build > Handover Providers**.
    2. Click **+ New handover Provider** and select the RingCentral Engage as the handover provider.
    3. *(Optional)* Toggle the **Forward Only Handover Conversations** setting to choose whether to forward all conversations to the provider or only those that trigger a handover. If this setting is enabled, Cognigy will forward only conversations where the handover event is triggered.
    4. *(Optional)* Toggle the **Get Queue Updates** setting to receive queue updates as events in the conversation.
    5. Configure the handover settings:

    | Setting                   | Description                                                   |
    | ------------------------- | ------------------------------------------------------------- |
    | **API Access Token**      | Paste the token from the RingCentral Engage API Access Token. |
    | **Bot Category ID**       | Enter the `CATEGORY_ID` for the `Bot` category.               |
    | **Agent Category ID**     | Enter the `CATEGORY_ID` for the `Agent` category.             |
    | **Realtime Access Token** | Paste the token from the RingCentral Engage Source.           |
    | **Realtime Endpoint URL** | Paste the URL from the RingCentral Engage Source.             |
    | **Webhook Secret**        | Paste the secret from the RingCentral Engage Webhook.         |
    | **Base API URL**          | Enter the Base API URL that you built previously.             |
  </Accordion>

  <Accordion title="2. Add a Handover to Human Agent Node">
    1. Create a Flow in Cognigy with a Handover to Human Agent Node.
    2. Link the Endpoint to this Flow.
    3. When triggered, this Node escalates the conversation to RingCentral Engage, assigning it the `Agent` category and sending the Cognigy User Profile to RingCentral Engage.
       See [Send User Profile](#additional-configuration).

    After configuring the Handover Settings, test the setup: send a message to the Endpoint and make sure that a conversation appears in RingCentral Engage.
  </Accordion>
</AccordionGroup>

## Additional Configuration

<AccordionGroup>
  <Accordion title="Apple Business Chat">
    Ensure human agents with API token ownership have `reply` permissions on all non-custom SDK sources.
  </Accordion>

  <Accordion title="Send User Profile to RingCentral Engage">
    During a Handover event, the Contact Profile is sent to RingCentral Engage to provide helpful information to human agents.

    <Warning>
      The Cognigy Contact Profile will be sent to RingCentral Engage and matched up with the object keys existing on RingCentral Engage. The information that wasn't matched will be added to the `context_data` object.

      * [RingCentral Engage accepted User Profile Object structure](https://developers.ringcentral.com/engage/digital/guide/sdks/source-sdk/objects#users) sent during the handover.
      * [Cognigy Profile schema](/ai/agents/analyze/contact-profiles#profile-schema). Cognigy Profile Schema can be edited and additional fields with specific object keys added.
    </Warning>
  </Accordion>

  <Accordion title="Get Queue Updates">
    The **Get Queue Update** setting enables the injection of queue updates into the conversation flow as events. This capability allows you to react to queue updates within your conversation.

    In addition to enabling the **Get Queue Update** option, you need to configure the Lookup Node so that the handover type is selected as **Queue**. For more details, refer to the [Handover Status](/ai/agents/develop/node-reference/logic/lookup) section.

    The queue updates from RingCentral Engage are visible in the `input.data.handover` object.
    The object contains the following properties:

    * `estimatedWaitTime` — the approximate wait time for the user, measured in milliseconds.
    * `position` — the conversation's position in the queue.
  </Accordion>
</AccordionGroup>

## More Information

* [All Handover Providers](/ai/escalate/handover-reference/overview)
* [Handovers](/ai/escalate/handovers)


# Salesforce Service Cloud
Source: https://docs.cognigy.com/ai/escalate/handover-reference/salesforce



<a href="/release-notes/4.100"><Badge>Updated in 4.100</Badge></a>

<Frame>
  <img alt="Salesforce Service Cloud handover provider logo" />
</Frame>

## Description

<Warning>
  Salesforce announced the removal date for [LiveAgent, Salesforce Chat, Embedded Chat, and Service Chat](https://help.salesforce.com/s/articleView?id=001790618\&type=1). The removal date is February 14, 2026. This handover provider connector will remain active until the removal date. Before this date, migrate to a [new connector](/ai/escalate/handover-reference/salesforce-miaw) for the Salesforce provider that supports [Messaging for In-App and Web](https://help.salesforce.com/s/articleView?id=service.miaw_intro_landing.htm\&type=5).
  If you configured the embedded Agent Copilot workspace, make sure that you update its [configuration](/ai/escalate/handover-reference/salesforce-miaw#additional-configuration) during the migration.

  The Salesforce handover provider bridges Cognigy and [Salesforce Service Cloud](https://salesforce.com/),
  enabling end users to connect with human agents working in a contact center that uses Salesforce Service Cloud.
</Warning>

## Prerequisites

* Access to [Salesforce Service Cloud](https://login.salesforce.com/).
* Created Endpoint [compatible with Salesforce](/ai/escalate/handovers#endpoints-compatible-with-handover-providers).

## Configuration on the Handover Provider Side

<Accordion title="Configure Salesforce Service Cloud">
  1. [Log in](https://login.salesforce.com/) to Salesforce.
  2. If no chat deployment is available in Salesforce, use the following guides:
     * [Create a Basic Chat Implementation](https://help.salesforce.com/s/articleView?id=sf.live_agent_set_up_basic_implementation.htm)
     * [Create Chat Deployments](https://help.salesforce.com/s/articleView?id=sf.live_agent_create_deployments.htm).
  3. On the Salesforce home page, navigate to **Feature Settings > Service > Chat > Chat Settings**. On the **Chat Settings** page, go to the **API Information > Chat API Endpoint**. Copy the Chat API Endpoint value without the `/chat/rest` part, for example `https://x.xxx-xx-xxx.salesforceliveagent.com`. Save it for future use.
  4. Navigate to **Feature Settings > Service > Embedded Service > Embedded Service Deployments**. In the row with the deployment you want to use, click <img alt="menu" />, then select **View**.
  5. In the **Embedded Service Code Snippets** section, click **Get Code**. Copy and save the values from the following parameters without the single quotes `''`:

     * `deploymentId`
     * `buttonId`
     * `organizationId`

     <br />

     ```js theme={null}
     embedded_svc.init(
         'https://...', 
         'https://...', 
         gslbBaseURL, 
         '<your-organization-id>', // Copy the Organization ID
         '...', 
         {
             baseLiveAgentContentURL: '...',
             deploymentId: '<your-deployment-id>', // Copy the Deployment ID
             buttonId: '<your-button-id>', // Copy the Button ID
         }
     );
     ```
</Accordion>

## Configuration on the Cognigy Side

<AccordionGroup>
  <Accordion title="1. Create a Handover Provider">
    1. Go to **Build > Handover Providers** and click **+ New Handover Provider**.
    2. In the **New Handover Provider** window, select **Salesforce** from the list and name the connector.
    3. *(Optional)* Activate the **Forward Events to the Flow** setting. This setting forwards any event to the Flow, which can be handled in the Lookup Node. Select the **Handover Status** type in the Lookup Node. In the Case Node, select **Events** from the **Handover Status** list. Agent Replies and Conversation Closed events are not forwarded.
    4. Fill in the following fields:

    | **Field**           | **Description**                                                                        |
    | ------------------- | -------------------------------------------------------------------------------------- |
    | Live Agent Base URL | The URL, copied from the **Chat API Endpoint** field.                                  |
    | API Version         | The API version is `52`.                                                               |
    | Organization ID     | The unique identifier for your Salesforce organization, copied from `organizationId`.  |
    | Deployment ID       | The ID of your deployment in Salesforce, copied from `deploymentId`.                   |
    | Button ID           | The ID of the button for the related deployment in Salesforce, copied from `buttonId`. |
  </Accordion>

  <Accordion title="2. Add a Handover to Human Agent Node">
    In the **Salesforce Settings** section of the [Handover to Human Agent](/ai/agents/develop/node-reference/service/handover-to-agent) Node, configure the following settings:

    * **Send Transcript as first message** — the setting allows Salesforce to receive the conversation transcript as the first message. This setting is turned off by default.
    * **Button ID** — the ID of the Button you created in the Salesforce installation. This setting will override the default Button ID that you specified in the Endpoint configuration. You can use different Button IDs for individual Handover to Human Agent Nodes.
    * **Pre-chat Entities and Details** — the additional context sent in a custom JSON structure, providing the information before engaging with the human agent.
  </Accordion>

  <Accordion title="2.1 Configure Pre-Chat Entities and Details">
    By default, Cognigy doesn't send a conversation transcript from the chat with the AI Agent to the chat with the human agent. If you activate the Send Transcript as a first message setting, Cognigy sends the conversation transcript to the chat with the human agent as the first message. The conversation transcript includes the entire chat between the user and the AI Agent.

    The conversation transcript is displayed when the human agent receives the first message from the user and can be found in the Salesforce Service Console, under the **Transcripts** tab, in the **Transcript > Details > Transcript > Body** section.

    Moreover, Cognigy allows users to send Pre-chat Details and Pre-chat Entities arrays as a custom JSON structure. This feature allows users to include additional context or relevant information before initiating the chat with the human agent. This information can then be used for further processing or personalizing the conversation based on the provided details.

    **Pre-Chat Entities**

    Cognigy allows the creation of fields
    and linking User details obtained through Pre-Chat Details to Salesforce Entities like Contact or Case.

    To do this, you must send the Pre-Chat Entities array to the live chat session on Salesforce.

    The JSON example shows how to connect the conversation to a Case and use Pre-Chat Details:

    ```json theme={null}
    [
       {
            "entityName":"Contact",         
            "saveToTranscript": "Contact",
            "linkToEntityName":"Case",
            "linkToEntityField":"ContactId",
            "showOnCreate":true,          
            "entityFieldsMaps":[
                {
                  "fieldName":"LastName",
                  "label":"LastName",
                  "doFind":true,
                  "isExactMatch":true,
                  "doCreate":true
               },
               {
                  "fieldName":"FirstName",
                  "label":"FirstName",
                  "doFind":true,
                  "isExactMatch":true,
                  "doCreate":true
               },
               {
                  "fieldName":"Email",
                  "label":"Email",
                  "doFind":true,
                  "isExactMatch":true,
                  "doCreate":true
               }          
            ]
         }   
      ]
    ```

    **Pre-Chat Details**

    Cognigy allows sending conversation information to Salesforce Live Agent using Pre-chat Details.

    By using the **entityMaps** field, this information can be mapped to specific Pre-Chat Entity fields.

    CognigyScript can fetch values from Cognigy input, profile, and context objects before sending the data.

    ```json theme={null}
    [
          {
            "label":"LastName",
            "value":"Mustermann",
            "entityMaps":[
               {
                  "entityName":"contact",
                  "fieldName":"LastName"
               }
            ],
            "transcriptFields":[
               "LastName__c"
            ],
            "displayToAgent":true
         },
         {
            "label":"FirstName",
            "value":"Max",
            "entityMaps":[
               {
                  "entityName":"contact",
                  "fieldName":"FirstName"
               }
            ],
            "transcriptFields":[
               "FirstName__c"
            ],
            "displayToAgent":true
         },
         {
            "label":"Email",
            "value":"max.mustermann@mail.de",
            "entityMaps":[
               {
                  "entityName":"contact",
                  "fieldName":"Email"
               }
            ],
            "transcriptFields":[
               "Email__c"
            ],
            "displayToAgent":true
         }
    ]
    ```

    For more information about the Pre-Chat Entities and Pre-Chat Details fields and their usage, read the [Salesforce API documentation](https://developer.salesforce.com/docs/atlas.en-us.live_agent_rest.meta/live_agent_rest/live_agent_rest_data_types.htm#CustomDetail).

    To check the handover, go to the installation that your server administrator has deployed.
  </Accordion>
</AccordionGroup>

## Additional Configuration

<Accordion title="Agent Copilot Workspace">
  To integrate Agent Copilot directly into Salesforce Service Cloud, refer to the [Salesforce Integrations](https://github.com/Cognigy/salesforce-integrations) documentation on GitHub.
</Accordion>

## More Information

* [All Handover Providers](/ai/escalate/handover-reference/overview)
* [Handovers](/ai/escalate/handovers)


# Salesforce MIAW
Source: https://docs.cognigy.com/ai/escalate/handover-reference/salesforce-miaw



<a href="/release-notes/2026.13"><Badge>Updated in 2026.13</Badge></a>

<Frame>
  <img alt="Salesforce MIAW handover provider logo" />
</Frame>

## Description

The Salesforce MIAW handover provider bridges Cognigy and [Salesforce Messaging for In-App and Web](https://help.salesforce.com/s/articleView?id=service.miaw_intro_landing.htm\&type=5),
enabling end users to connect with human agents working in a contact center that uses Salesforce.

## Prerequisites

* Access to [Salesforce](https://login.salesforce.com/).
* Created Endpoint [compatible with Salesforce](/ai/escalate/handovers#endpoints-compatible-with-handover-providers).
* Give [Salesforce users access](https://help.salesforce.com/s/articleView?id=service.miaw_non_agent_permissions_1.htm\&type=5).

## Restrictions

* The Salesforce MIAW handover provider supports attachments only when used with [Webchat](/webchat/overview).

## Limitations

* Salesforce MIAW has file size limitations for attachments. For more information, refer to the Salesforce documentation on [File Size and Sharing Limits](https://help.salesforce.com/s/articleView?id=experience.collab_files_size_limits.htm\&type=5).

## Configuration on the Handover Provider Side

<AccordionGroup>
  <Accordion title="1. Configure Omni-Channel Settings">
    1. Log in to [Salesforce](https://login.salesforce.com/).
    2. On the **Salesforce home** page, select **Setup** in the upper-right corner, then select **Setup** from the menu.
    3. In the **Quick Find** box, enter `Omni-Channel Settings` and click **Omni-Channel Settings**. Alternatively, navigate to **Feature Settings > Service > Omni-Channel > Omni-Channel Settings** in the left-side menu.
    4. On the **Omni-Channel Settings** page, ensure that the **Enable Omni-Channel** setting is activated.
    5. In the **Quick Find** box, enter `Service Channels` and click **Service Channels**. Alternatively, navigate to **Feature Settings > Service > Omni-Channel > Service Channels** in the left-side menu.
    6. Create a channel by clicking **New**. Fill in the following fields:
       * **Service Channel Name** – enter the name of your channel, for example, `Messaging`.
       * **Developer Name** – the developer name will be generated automatically as soon as you enter the service channel name. Change the developer name if needed.
       * **Salesforce Object** – select **Messaging Session**. In case the **Messaging Session** is not available for you, select **Case**.
    7. Save changes.
  </Accordion>

  <Accordion title="2. Create a Queue">
    1. In the **Quick Find** box, enter `Queues` and click **Queues**. Alternatively, navigate to **Users > Queues** in the left-side menu.
    2. Create a new queue by clicking **New**. Fill in the following fields:
       * **Label** – enter the name of your channel, for example, `Messaging`.
       * **Queue Name** – the queue name will be generated automatically as soon as you enter the channel name. Change the queue name if needed.
    3. In the **Supported Objects** section, add the following options:
       * **Messaging User**
       * **Messaging Session**
       * **Chat Session**
       * **Case** (if you selected **Case** as a Salesforce object in the **Service Channels** settings)
    4. In the **Queue Members** section, add the human agents you want to include in the queue. Save changes.
  </Accordion>

  <Accordion title="3. Create a Flow">
    1. In the **Quick Find** box, enter `Flows`, then click **Flows**.
    2. To create a new flow, click **New Flow** in the upper-right corner and select **Omni-Channel Flow**.
    3. Open the **Toolbox** tab by clicking the **Toggle Toolbox** button in the upper-left corner. Then, click **New Resource** on the **Manager** tab and configure the following fields:
       * From the **Resource Type** list, select **Variable**.
       * In the **API Name** field, enter `recordId`.
       * From the **Data Type** list, select **Text**.
       * Activate **Available for input**, then click **Done**.
    4. In the Flow editor, click **+** to add an element. Select **Route Work** from the list.
    5. In the **Label** field, enter a name for the element. The value in the **API Name** field will be generated automatically.
    6. In the **Set Input Values** section, configure the following fields:
       * **Record ID Variable** — select `recordId`.
       * **Service Channel** — select the service channel that you created previously.
       * **Route To** — select **Queue**.
    7. In the **Queue** section, select the ID related to your queue from the **Queue ID** list.
    8. Click **Save** and fill in the following fields:
       * **Flow Label** — enter the name of your flow.
       * **Flow API Name** — the flow API name will be generated automatically. Change flow API name if needed.
    9. Click **Activate**, then **Publish**. At this stage, you will route your flow via the queue. You can also configure routing for a specific agent, bot, or skill. For more details, refer to the Salesforce [documentation](https://help.salesforce.com/s/articleView?id=service.omnichannel_routing_targets.htm\&type=5).
  </Accordion>

  <Accordion title="4. Create a Messaging Channel">
    1. Go back to the **Setup** page, enter `Messaging Settings` in the **Quick Find** box, and click **Messaging Settings**. Alternatively, navigate to **Feature Settings > Service > Messaging > Messaging Settings** in the left-side menu.
    2. On the **Messaging Settings** page, click **New Channel**, then **Next**. Select **Messaging for In-App and Web** and fill in the following fields:
       * **Channel Name** — enter the name of your messaging channel.
       * **Developer Name** — the developer name will be created automatically.
       * **Deployment Type** — select the interface where users will interact with Salesforce MIAW:
         * **Web** — you need to enter the URL of the website on which users can interact with Salesforce MIAW in the **Domain** field.
         * **Mobile**
    3. In the **Channel Routing** section, configure the following:
       * **Routing Type** — select **Omni-Flow**.
       * **Flow Definition** — enter the Flow name that you created.
       * **Fallback Queue** — enter the queue name that you created. Save changes.
    4. Ensure that your channel is active. On the **Messaging Settings** page, check if the channel is marked as **Active** in the **Active** column. If not, select your channel and click **Activate** in the upper-right corner.
  </Accordion>

  <Accordion title="5. Create an Embedded Service">
    1. In the **Quick Find** box, enter `Embedded Service Deployments` and click **Embedded Service Deployments**. Alternatively, navigate to **Feature Settings > Service > Messaging > Embedded Service Deployments** in the left-side menu.
    2. On the **Embedded Service Deployments** page, click **New Deployment**.
    3. Select the **Messaging for In-App and Web** type, then **Custom Client**. Fill in the following fields:
       * **Embedded Service Deployment Name** — enter the name of your deployment.
       * **API Name** — the API name will be created automatically. Change API name if needed.
    4. From the **Messaging Channel** list, select the channel you created previously. Save changes.
    5. Go to **Code Snippet**. Copy the `OrganizationId`, `DeveloperName`, and `Url` values for later use in Cognigy.
    6. Click **Done**. Check whether your deployment is active and click **Publish**. The deployment will take about 10 minutes. Afterward, you can go to Cognigy to configure the handover settings.
  </Accordion>
</AccordionGroup>

## Configuration on the Cognigy Side

<AccordionGroup>
  <Accordion title="1. Create a Handover Provider">
    1. Go to **Build > Handover Providers** and click **+ New Handover Provider**.
    2. In the **New Handover Provider** window, select **Salesforce MIAW** from the list and name the connector.
    3. *(Optional)* Activate the **Forward Events to the Flow** setting. This setting forwards any event to the Flow, which can be handled in the Lookup Node. Select the **Handover Status** type in the Lookup Node. In the Case Node, select **Events** from the **Handover Status** list. Agent Replies and Conversation Closed events are not forwarded.
    4. Fill in the following fields:

    | **Field**                       | **Description**                                                                         |
    | ------------------------------- | --------------------------------------------------------------------------------------- |
    | Base URL                        | The URL of your configured service in Salesforce, copied from `Url`.                    |
    | Organization ID                 | The unique identifier of your organization in Salesforce, copied from `OrganizationId`. |
    | Embedded Service Developer Name | The ID of your deployment in Salesforce, copied from `DeveloperName`.                   |
    | Capabilities Version            | The API version is `1`.                                                                 |
  </Accordion>

  <Accordion title="2. Add a Handover to Human Agent Node">
    In the **Salesforce Settings** section of the [Handover to Human Agent](/ai/agents/develop/node-reference/service/handover-to-agent) Node, configure the following settings:

    * **Send Transcript as first message** — the setting allows Salesforce to receive the conversation transcript as the first message. If the transcript exceeds 3000 characters, it is split into multiple messages. This setting is turned off by default.
    * **Conversation Routing Attributes** — the setting supports key-value pairs, where the keys are variables you manually specify in the Omni-Channel flow in Salesforce. To specify variables, follow the [Map Pre-Chat Values in Omni-Channel Flow](https://help.salesforce.com/s/articleView?id=service.miaw_map_messaging_2.htm) guide in the Salesforce documentation. Then, enter the key-value pairs in JSON format. For example:
      ```json theme={null}
      {
        "customerName": "John Doe",
        "issueType": "Billing"
      }
      ```
      where `customerName` and `issueType` are variables you specified on the Salesforce side, and `John Doe` and `Billing` are the values you want to pass to the flow in Salesforce.
    * **Identity Token for Authenticated Users** — the setting allows you to specify an access token generated via the [Salesforce API](https://developer.salesforce.com/docs/service/messaging-api/references/miaw-api-reference?meta=generateAccessTokenForAuthenticatedUser). This token enables users who have already authenticated within your system to interact securely with Salesforce MIAW. Before using this token, [set up user verification](https://help.salesforce.com/s/articleView?id=sf.user_verification_setup.htm) on the Salesforce side. By default, the setting is turned off, meaning that unauthenticated users can interact with Salesforce MIAW.
    * **Enable User Connects Message** - the setting allows you to notify human agents when an end user reconnects to the chat. The parameter is enabled by default. When the parameter is enabled, the message `User joined the conversation` appears in the chat as soon as the end user opens a new URL on the same tab as the chat and then returns to the chat tab by clicking the **←** (back arrow) at the top bar in the browser.
    * **Enable User Disconnects Message** - the setting allows you to notify human agents when an end user disconnects from the chat. The parameter is enabled by default. When the parameter is enabled, the message `User left the conversation` is sent as soon as the user closes the tab with the chat or switches to a new URL address within the current tab.
    * **End Chat When Agent Leaves** — the setting allows the handover to end automatically when a human agent is removed from a Salesforce MIAW conversation. This setting is turned off by default. Instead of waiting for the conversation close event, which [Salesforce may not always send](https://help.salesforce.com/s/articleView?id=000397177\&type=1), this setting relies on the participant change event to trigger the end of the handover.
    * **Agent Change Notification** — this setting sends a system message to the end user indicating when the conversation is reassigned to another human agent. This parameter is deactivated by default. When the conversation is reassigned, the system message is sent under the human avatar name set in the [Webchat v3 Endpoint settings](/webchat/v3/configuration#webchat-layout).
    * **Get Queue Position** — activate the setting to retrieve the end user's current queue position. This setting is turned off by default.
    * **Get Estimated Wait Time** — activate the setting to retrieve the estimated wait time for the end user. This setting is turned off by default.
    * **Alternative Update** — activate this setting to send an alternative message to the end user when certain conditions are met. This setting is turned off by default.
    * **Maximum Queue Position** — set the threshold for the maximum queue position. If this threshold is reached, the alternative text is sent instead of the actual queue position. This parameter appears if **Alternative Update** is active.
    * **Alternative Text** — specify the message to be sent when the maximum queue position is reached. This parameter appears if **Alternative Update** is active.

    Test your Flow to ensure it works as expected. You can use Demo Webchat to send messages to the Salesforce contact center.
    On the Salesforce side, go to the Service Console. In the Service Console, open [Messaging Sessions](https://help.salesforce.com/s/articleView?id=service.livemessage_create_console_app.htm\&type=5), and make sure that your human agent is online.
  </Accordion>
</AccordionGroup>

## Additional Configuration

### Agent Copilot Workspace

Within the Salesforce MIAW integration, you can use the [Agent Copilot workspace](/agent-copilot/overview) as an assistant to your human agents.

By default, two versions of the application are provided:

<AccordionGroup>
  <Accordion title="Standalone">
    Agent Copilot will be available as a standalone application via the following link:

    `https://${AgentCopilotBaseUrl}/?userId=${userId}&sessionId=${sessionId}&URLToken=${URLToken}`

    Where:

    * `AICopilotBaseUrl` — the base URL used to access the Agent Copilot workspace.
    * `sessionId` — the Cognigy session ID. A session represents the interaction between the human agent and the Agent Copilot workspace.
    * `userId` — the Cognigy user ID. It helps Agent Copilot associate the interaction with a specific user.
    * `URLToken` — the Endpoint token from Cognigy. The Endpoint must be linked to the Handover Flow. To find this token, go to the Endpoint that you created previously. In the **Endpoint URL** field, copy the token after `https://endpoint-trial.cognigy.ai/`. For example, in the URL `https://endpoint-trial.cognigy.ai/f38791ae20d4961acf0e97d9f377c4fe3df92894e1eff1c7a774a8ed089a4590`, the token is `f38791ae20d4961acf0e97d9f377c4fe3df92894e1eff1c7a774a8ed089a4590`. Only authorized users can access the Agent Copilot system through this URL.

    Example URL:

    ```txt wrap theme={null}
    https://ai-copilot-trial.cognigy.ai/?sessionId=session123&userId=user123&URLToken=a4d5c86c98f27730311591f28d194510e05ffed30ca148e3344970defd418e7d
    ```
  </Accordion>

  <Accordion title="Embedded">
    To use the embedded version of the Agent Copilot workspace, you need to update the settings on the Salesforce side and configure the Agent Copilot UI components.

    <AccordionGroup>
      <Accordion title="1. Update Messaging Settings">
        1. In the [Salesforce](https://login.salesforce.com/) interface, go to **Setup > Messaging Settings** and select the messaging channel that you created previously.
        2. In the messaging channel, go to **Custom Parameters** and click **New**.
        3. In the **New Custom Parameter** window, configure the following parameters:
           * **Parameter Name** — enter **Copilot URL**.
           * **Parameter API Name** — the parameter API name will be generated automatically as soon as you enter the parameter name. Change the parameter API name if needed.
           * **Channel Variable Name** — the channel variable name will be generated automatically as soon as you enter the parameter name. Change the channel variable name if needed.
           * **Data Type** — select **String**.
           * **Maximum Length** — enter `255`. Save changes.
        4. In the **Parameter Mappings** section, click **New**.
        5. In the **New Parameter Mapping** window, configure the following parameters:
           * **Parameter** — select **Copilot URL** that was added under **Custom Parameters**.
           * **Flow Variable Name** — enter `copilotUrl`. <br /> Save changes.
      </Accordion>

      <Accordion title="2. Update Embedded Service Deployments">
        1. In the **Quick Find** box, enter `Embedded Service Deployments` and click **Embedded Service Deployments**. Alternatively, navigate to **Feature Settings > Service > Messaging > Embedded Service Deployments** in the left-side menu.
        2. Select the deployment that you created previously, then select **Pre-Chat**.
        3. Enable the **Activate the pre-chat feature** setting.
        4. Go to the **Visible Pre-Chat Fields** section, click **Add > Custom**.
        5. In the **New Custom** window, fill in the following fields:
           * **Field Type** — select **text**.
           * **Channel Variable Name** — select `Copilot_URL` that you added in **Messaging Settings** previously. <br /> Save changes.
        6. Click **Save**. On the **Embedded Service Deployment Settings** page, click **Publish**.
      </Accordion>

      <Accordion title="3. Update the Flow">
        1. In the **Quick Find** box, enter `Flows`, then click **Flows**. Alternatively, navigate to **Feature Settings > Service > Messaging > Flows** in the left-side menu.
        2. Go to the Flow that you created previously.
        3. Open the **Toolbox** tab by clicking the **Toggle Toolbox** button in the upper-left corner. Then, click **New Resource** on the **Manager** tab and configure the following fields:
           1. From the **Resource Type** list, select **Variable**.
           2. In the **API Name** field, enter `copilotUrl`.
           3. From the **Data Type** list, select **Text**.
           4. Activate **Available for Input**, then click **Done**.
        4. In the Flow editor, above the **Route Work** action, click **+** to add an element. Select **Update Records** from the list.
        5. In the **Label** field, enter a name for the element. The value in the **API Name** field will be generated automatically.
        6. In the **How to Find Records to Update and Set Their Values** section, select **Specify conditions to identify records**.
        7. In the **Update Records of This Object Type** section, select **Messaging Session** from the **Object** list.
        8. In the **Filter Messaging Session Records** section, configure the following fields:
           * **Condition Requirement to Update Records** — select **All Conditions Are Met (AND)**.
           * **Field** — enter `Id`.
           * **Operator** — select **Equals**.
           * **Value** — enter `recordId`.
        9. In the **Set Field Values for the Messaging Session Records** section, configure the following fields:
           * **Field** — enter `Copilot__c`.
           * **Value** — enter `copilotUrl`.
        10. Click **Save a New Version**, then **Activate**.
      </Accordion>

      <Accordion title="4. Configure UI Components">
        To configure the Agent Copilot UI components, refer to the [Salesforce Integrations](https://github.com/Cognigy/salesforce-integrations) documentation on GitHub. Once the components are configured, you can test the embedded Agent Copilot workspace using [Demo Webchat](/webchat/demo).

        If you used an embedded Agent Copilot workspace with the [deprecated Salesforce integration](/ai/escalate/handover-reference/salesforce) and migrated to Salesforce MIAW, you need to update the Agent Copilot UI components. Refer to the [Set up the Component](https://github.com/Cognigy/salesforce-integrations?tab=readme-ov-file#set-up-the-componentrefer) documentation on GitHub.
      </Accordion>
    </AccordionGroup>
  </Accordion>
</AccordionGroup>

#### Automated Messaging Components

The Salesforce MIAW integration supports automated messaging components for sending auto-responses during messaging sessions, such as welcome messages, inactivity reminders, and survey links. To learn how to use these components, refer to the [Salesforce documentation on Automated Messaging Components](https://help.salesforce.com/s/articleView?id=service.messaging_components_auto_response.htm\&type=5).

### Prevent Conversations from Automatically Closing

Cognigy prevents Salesforce MIAW chats from closing automatically when human agents leave the conversation and continues polling for new events.

#### Cognigy 2025.16 and earlier

Salesforce MIAW chat closes automatically when a human agent leaves the conversation. You can control this behavior by setting the `DISABLE_ABORT_HANDOVER_ON_SALESFORCE_MIAW_PARTICIPANT_CHANGE` variable in the `values.yaml` file. By default, this variable is set to `false`. Setting it to `true` keeps the conversation open after human agents leave and continues polling for new events.

### Access Token Expiration

When a handover starts, Salesforce MIAW provides an access token with a configurable expiration time to Cognigy in the backend. To make sure that the session is synchronized, Cognigy polls Salesforce MIAW approximately every 2 minutes to check for new messages or actions, and verifies if the access token is still valid. Additionally, Cognigy has a configurable grace period, counted from the end of the access token expiration time to control the session termination. The default grace period is 30 minutes. When the grace period is reached, Cognigy terminates the session on both Cognigy and Salesforce MIAW sides. This approach prevents unnecessary API calls to Salesforce MIAW sessions with expired access tokens.

<AccordionGroup>
  <Accordion title="Set the Access Token Expiration Time">
    1. On the **Salesforce home** page, click **Setup** in the upper-right corner, then select **Setup** from the menu.
    2. In the **Quick Find** box, enter `Messaging Settings` and follow the **Feature Settings > Service > Messaging** path.
    3. On the **Messaging Settings** page, click the **Show one more action** arrow next to the channel you want to configure, and select **Edit**.
    4. On the **Messaging for In-App and Web** page, configure as follows:

    * Without user verification — enter the expiration time in **Authorization Token Expiration Time for Unverified Users**. The default value is 360 minutes. Save changes.
    * With user verification — select **Add User Verification** and enter the expiration time in **Authorization Token Expiration Time for Verified Users**. The default value is 60 minutes. Save changes.
  </Accordion>

  <Accordion title="Set the Grace Period">
    * For on-premises installations, set the grace period in seconds using the
      `SALESFORCE_MIAW_ACCESS_TOKEN_GRACE_PERIOD` variable in the `values.yaml`
      file.
    * For SaaS installations, contact [Cognigy Support](/help/get-help).
  </Accordion>

  <Accordion title="Example">
    If you have an access token expiration time of 6 hours (360 minutes) and a
    grace period of 30 minutes, Cognigy terminates the session on both
    platforms 5 hours and 30 minutes after the handover started.
  </Accordion>
</AccordionGroup>

### Mark Conversations as Inactive

Human agents managing multiple active conversations can quickly reach capacity. Without a way to pause unresolved sessions, new incoming conversations may be delayed, affecting response times.
Your human agents can temporarily pause unresolved conversations by marking them as inactive. This approach frees up capacity for new sessions while still allowing customers to reconnect later.

To mark a conversation as inactive, open it in Salesforce MIAW, click the dropdown next to **End Chat**, and select **Customer Inactive**. For more details, refer to [Ending or Inactivating Messaging Sessions Automatically](https://help.salesforce.com/s/articleView?id=service.messaging_auto_end.htm\&type=5) and [Inactivate a Messaging Session](https://help.salesforce.com/s/articleView?id=service.livemessage_agent_inactivate_session.htm\&type=5) in the Salesforce documentation.

## Troubleshooting

| Possible Error                                                                                  | How to Resolve                                                                                                                                                                                                                                                                                                                                                                           |
| ----------------------------------------------------------------------------------------------- | :--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Could not find field `Copilot__c`. Make sure to add this field to your Salesforce installation. | An embedded Agent Copilot doesn't work. Make sure that you've updated the settings in Salesforce. Check your flow and make sure that `Copilot__c` is included in the **Update Records** action.                                                                                                                                                                                          |
| You are trying to pull data from a Live Chat transcript while you are on a Messaging session.   | This error occurs when you migrate from the [deprecated Salesforce integration](/ai/escalate/handover-reference/salesforce) to Salesforce MIAW without updating the Agent Copilot UI components. To update the components, refer to the [Set up the Component](https://github.com/Cognigy/salesforce-integrations?tab=readme-ov-file#set-up-the-componentrefer) documentation on GitHub. |
| Login to Omni-Channel failed.                                                                   | This error occurs when a resource limit is exceeded. In the Salesforce Console, delete old messaging sessions, chat transcripts, and chat sessions. This action frees up space.                                                                                                                                                                                                          |

## More Information

* [All Handover Providers](/ai/escalate/handover-reference/overview)
* [Handovers](/ai/escalate/handovers)


# storm
Source: https://docs.cognigy.com/ai/escalate/handover-reference/storm



<a href="/release-notes/2025.24"><Badge>Updated in 2025.24</Badge></a>

<Frame>
  <img alt="storm handover provider logo" />
</Frame>

## Description

The storm handover provider bridges Cognigy and [storm](https://www.contentguru.com/en-gb/products/storm-cx/omni-channel-customer-engagement/omni-channel-agent-desktop/), the contact center platform from Content Guru, enabling end users to connect with human agents working in a contact center that uses storm.

## Prerequisites

* Access to [storm DTA](https://www.timeforstorm.eu/dta).

## Configuration on the Cognigy Side

<AccordionGroup>
  <Accordion title="1. Create a Handover Provider">
    To link your Cognigy Endpoint to storm:

    1. Go to **Deploy > Handover Providers**.

    2. Click **+ New Handover Provider** and select **storm®** as the handover provider.

    3. Configure the following fields:
       * **Base API URL** — the URL of your configured service in storm.
       * **Storm Authentication** — select a storm connection from the list. Alternatively, click <img alt="add storm connection" /> next to the field to add a storm connection. Fill in the following fields:

         | **Field**       | **Description**                                     |
         | --------------- | --------------------------------------------------- |
         | Connection name | Enter a unique name for your connection.            |
         | API Secret      | Enter the API secret for your Content Guru service. |
         | API Key         | Enter the API key for your Content Guru service.    |

    4. Click **Save** to store the configuration.
  </Accordion>

  <Accordion title="2. Add a Handover to Human Agent Node">
    To add a Handover to Human Agent Node, follow these steps:

    1. In the Flow editor, add a **Handover to Human Agent** Node.
    2. Open the Node editor and select the storm handover provider from the **Handover Provider** list. Save changes.
  </Accordion>
</AccordionGroup>

## Configuration on the Handover Provider Side

Contact your Content Guru representative to finalize the integration between Cognigy and storm.

## More Information

* [All Handover Providers](/ai/escalate/handover-reference/overview)
* [Handovers](/ai/escalate/handovers)


# Handover Providers
Source: https://docs.cognigy.com/ai/escalate/handovers

Connect AI Agents to contact centers with Handover Providers for seamless escalation to human agents.

<a href="/release-notes/2026.16"><Badge>Updated in 2026.16.0</Badge></a>

*Handover Providers* connect AI Agents to contact centers, enabling seamless handoffs from automated conversations to human agents.
The handover happens through the [Handover to Human Agent Node](/ai/agents/develop/node-reference/service/handover-to-agent) within a Flow,
allowing escalation when human assistance is required — whether due to complexity, personalization, or AI Agent's limitations.
With multiple Handover to Human Agent Nodes in your Flow, you can route conversations to different contact centers. For example, VIP customers can be directed to NiCE CXone, while basic inquiries go to Cognigy Live Agent.

To enable the handover process, in addition to the Handover to Human Agent Node, you'll need:

* [A compatible contact center and its connector](/ai/escalate/handover-reference/overview)
* [An Endpoint that supports handovers](#endpoints-compatible-with-handover-providers)

Cognigy offers a range of tools to manage this escalation effectively:

* Nodes:
  * [Handover to Human Agent](/ai/agents/develop/node-reference/service/handover-to-agent) — initiates the transfer to a human agent.
  * [Question](/ai/agents/develop/node-reference/basic/question) — initiates the transfer to a human agent either when a [specific Intent is detected](/ai/agents/develop/node-reference/basic/question#escalation-intents) or when the [AI Agent misunderstands the user's input](/ai/agents/develop/node-reference/basic/question#escalation-wrong-answers).
  * [Check Agent Availability](/ai/agents/develop/node-reference/service/check-agent-availability) — verifies whether human agents are free to take over.
  * [Lookup](/ai/agents/develop/node-reference/logic/lookup) — monitors handover status, for example, completed, canceled, error, queue.
  * [Close Handover](/ai/agents/develop/node-reference/service/close-handover) — ends the handover process when appropriate.
  * [Set Handover Inactivity](/ai/agents/develop/node-reference/service/set-handover-inactivity) — defines timeouts for inactive handovers.
* Features:
  * [User Inactivity Detection](/ai/escalate/handover-features/user-inactivity-detection) — triggers actions based on user silence.
  * [Real-Time Queue Notifications](/ai/escalate/handover-features/real-time-queue-notifications) — provides live updates on queue status (applicable only to Cognigy Live Agent).

## Endpoints Compatible with Handover Providers

The Endpoints that support handover are listed below:

* [8x8](/ai/agents/deploy/endpoint-reference/8x8)
* [Azure Bot Services](/ai/agents/deploy/endpoint-reference/azure-bot-services)
* [Facebook (Meta) Messenger](/ai/agents/deploy/endpoint-reference/facebook-messenger)
* [LINE Messenger](/ai/agents/deploy/endpoint-reference/line-messenger)
* [RingCentral Engage](/ai/agents/deploy/endpoint-reference/ringcentral-engage)
* [Slack](/ai/agents/deploy/endpoint-reference/slack)
* [Socket](/ai/agents/deploy/endpoint-reference/socketio)
* [Sunshine Conversations](/ai/agents/deploy/endpoint-reference/sunshine-conversations)
* [Voice Gateway](/ai/agents/deploy/endpoint-reference/voice-gateway)
* [Webchat](/webchat/overview)
* [Webhook](/ai/agents/deploy/endpoint-reference/webhook)
* [WhatsApp](/ai/agents/deploy/endpoint-reference/whatsapp)

## Working with Handover Providers

<Tabs>
  <Tab title="GUI">
    You can create, edit, and delete handover providers in **Deploy > Handover Providers**.
  </Tab>

  <Tab title="API">
    You can view, edit, and delete handover providers using the [Cognigy API](https://api-trial.cognigy.ai/openapi#post-/v2.0/handover/providers).
  </Tab>
</Tabs>


# Migrate from Endpoint Settings to Handover Provider Settings
Source: https://docs.cognigy.com/ai/escalate/migration



<a href="/release-notes/2026.16"><Badge>Updated in 2026.16.0</Badge></a>

In 2026.16.0, the **Handover Providers** section in the Endpoint settings was removed.

If you are using version 2026.15 or earlier, migrate your handover provider settings before upgrading to 2026.16.0. The new interface provides more flexibility and improved organization for managing contact center integrations.

Follow this guide to migrate your handover provider settings to the new interface.

The migration process is seamless and doesn't involve any downtime.

### Benefits of the New Handover Provider Interface

* **Multiple Handover Providers**. The new interface lets you use multiple handover providers in a single Flow, providing greater flexibility to route between different contact centers.
* **Isolated Logic**. The handover provider logic is now isolated from the Endpoint logic, simplifying AI Agent deployment.

### Migrate Handover Settings to the New Interface

1. Go to **Deploy > Endpoint** and select the desired Endpoint.
2. In the **Handover Settings** section, click **Migrate to Handover Providers**.
3. Enter the name for the new handover provider to migrate the settings to the **Handover Providers** page.
4. Confirm the migration. The handover settings will be copied to the new interface, and the old settings will be preserved.

Some providers may require adding the settings manually, such as adding secrets or API keys.
For example, for Genesys, you may need to select the existing connection or add a new one in the **Genesys Cloud Authentication** field.

### Update Flow Settings to Reflect Handover Provider Changes

1. Go to **Build > Flows** and select your Flow. Update the settings in the following Nodes:

   * **Handover to Human Agent Node**
   * **Check Agent Availability Node**

   In each of these Nodes, go to the settings for the handover provider you migrated and copy these settings for later use.

2. In each of these Nodes, go to the **Handover Provider** list and switch from the **Legacy** option to the name of the handover provider you created. Note that this option works if the handover settings are still present in the Endpoint. If they're deleted, the **Legacy** option won't function.

3. In the handover provider settings section, manually fill in the fields with the respective parameters using the values copied earlier.

4. Save your changes to apply them.

### Final Testing and Cleanup

1. Test your Flow to ensure it works as expected.
2. Once you have tested your Flow, you can delete the **Handover to Human Agent Settings** from the Endpoint. Note that deleting old Handover settings from the Endpoint will result in permanent loss, and you can't perform a rollback.

## More Information

* [Handovers](/ai/escalate/handovers)


# AI Default Channel Formats
Source: https://docs.cognigy.com/ai/for-developers/code/ai-default-channel-formats

Reference for JSON payload formats used with the Cognigy default channel in Code Nodes. Covers text, Quick Replies, Adaptive Cards, and channel output.

You can send JSON-formatted output directly against your [Endpoints](/ai/agents/deploy/endpoints/overview) for it to be output to the respective channels.

The [AI Default Channel](/ai/agents/develop/node-reference/basic/say#default-cognigy-channel) is special in a way that content from the Default Channel is automatically converted into the correct payloads for all other channels.

Using the [Code](/ai/for-developers/code/overview) Node, you can also send JSON-formatted output to the Endpoints for the AI Default Channel, and it will handle the conversion into the appropriate content. The required payload formats are described below.

## Payload Formats

<AccordionGroup>
  <Accordion title="Text">
    Sends text to the Endpoint.

    <Tabs>
      <Tab title="JSON Format">
        ```js theme={null}
        api.output("Your text", null);
        ```
      </Tab>

      <Tab title="Say Node">
        <Frame>
          <img alt="Text output configured in a Say Node" />
        </Frame>
      </Tab>

      <Tab title="Webchat">
        <Frame>
          <img alt="Text output rendered in Webchat" />
        </Frame>
      </Tab>
    </Tabs>
  </Accordion>

  <Accordion title="Text with Quick Replies">
    Sends text and a number of quick reply buttons.

    <Tabs>
      <Tab title="JSON Format">
        ```js theme={null}
        api.output(null, {
          "type": "quickReplies",
          "_cognigy": {
            "_default": {
              "_quickReplies": {
                "type": "quick_replies",
                "text": "Quick Reply Text",
                "quickReplies": [
                  {
                    "id": 111,
                    "contentType": "postback",
                    "payload": "Postback 1",
                    "title": "Button 1",
                    "imageUrl": "https://www.cognigy.com/hubfs/Cognigy.svg"
                  },
                  {
                    "id": 112,
                    "contentType": "phone_number",
                    "payload": "+49123456",
                    "title": "Button 2"
                  },
                  {
                    "id": 112,
                    "contentType": "trigger_intent",
                    "payload": "",
                    "title": "Button 3",
                    "intentName": "Intent Name"
                  }
                ]
              }
            }
          }
        });
        ```
      </Tab>

      <Tab title="Say Node">
        <Frame>
          <img alt="Quick replies configured in a Say Node" />
        </Frame>

        <Frame>
          <img alt="Quick reply button detailed settings" />
        </Frame>
      </Tab>

      <Tab title="Webchat">
        <Frame>
          <img alt="Quick reply buttons rendered in Webchat" />
        </Frame>
      </Tab>
    </Tabs>
  </Accordion>

  <Accordion title="Gallery">
    Outputs a carousel of cards.

    <Tabs>
      <Tab title="JSON Format">
        ```js theme={null}
        api.output(null, {
          "type": "gallery",
          "_cognigy": {
            "_default": {
              "_gallery": {
                "type": "carousel",
                "items": [
                  {
                    "id": 111,
                    "title": "Cognigy",
                    "subtitle": "Cognigy is a conversational AI platform.",
                    "imageUrl": "https://www.cognigy.com/hubfs/AI%20Copy%205.svg",
                    "buttons": [
                      {
                        "id": 112,
                        "payload": "Postback 1",
                        "type": "postback",
                        "title": "Button 1"
                      },
                      {
                        "id": 113,
                        "payload": "",
                        "type": "web_url",
                        "title": "URL Button",
                        "url": "https://www.cognigy.com/products/cognigy-ai"
                      }
                    ]
                  },
                  {
                    "id": 121,
                    "title": "Cognigy.VG",
                    "subtitle": "Cognigy Voice Gateway connects your AI to your contact center",
                    "imageUrl": "https://www.cognigy.com/hubfs/cognigy-vg-logo.svg",
                    "buttons": [
                      {
                        "id": 122,
                        "payload": "",
                        "type": "web_url",
                        "title": "Click me",
                        "url": "https://www.cognigy.com/products/voice-gateway"
                      }
                    ]
                  }
                ],
                "fallbackText": "Fallback text for voice channels"
              }
            }
          }
        });
        ```
      </Tab>

      <Tab title="Say Node">
        <div>
          <Frame>
            <img alt="Gallery carousel configured in a Say Node" />
          </Frame>

          <Frame>
            <img alt="Gallery card button settings in the Node editor" />
          </Frame>
        </div>
      </Tab>

      <Tab title="Webchat">
        <Frame>
          <img alt="Gallery carousel rendered in Webchat" />
        </Frame>
      </Tab>
    </Tabs>
  </Accordion>

  <Accordion title="Text with Buttons">
    Outputs text with a list of buttons.

    <Tabs>
      <Tab title="JSON Format">
        ```js theme={null}
        api.output(null, {
          "type": "buttons",
          "_cognigy": {
            "_default": {
              "_buttons": {
                "type": "buttons",
                "text": "Text with Buttons",
                "buttons": [
                  {
                    "id": 111,
                    "payload": "Postback Value",
                    "type": "postback",
                    "title": "Postback Value Button"
                  },
                  {
                    "id": 112,
                    "payload": "",
                    "type": "web_url",
                    "title": "URL Button",
                    "url": "https:\\\\google.com"
                  },
                  {
                    "id": 113,
                    "payload": "+491234567",
                    "type": "phone_number",
                    "title": "Phone Number Button"
                  },
                  {
                    "id": 114,
                    "payload": "",
                    "type": "trigger_intent",
                    "title": "Intent Button",
                    "intentName": "Intent Name"
                  }
                ]
              }
            }
          }
        });
        ```
      </Tab>

      <Tab title="Say Node">
        <div>
          <Frame>
            <img alt="Text with buttons configured in a Say Node" />
          </Frame>

          <Frame>
            <img alt="Text with buttons detailed settings in the Node editor" />
          </Frame>
        </div>
      </Tab>

      <Tab title="Webchat">
        <Frame>
          <img alt="Text with buttons rendered in Webchat" />
        </Frame>
      </Tab>
    </Tabs>
  </Accordion>

  <Accordion title="List">
    Outputs a list.

    <Tabs>
      <Tab title="JSON Format">
        ```js theme={null}
        api.output(null, {
          "type": "list",
          "_cognigy": {
            "_default": {
              "_list": {
                "type": "list",
                "items": [
                  {
                    "title": "List Item 1",
                    "subtitle": "Subtitle",
                    "imageUrl": "https://www.cognigy.com/hubfs/Cognigy.svg",
                    "defaultActionUrl": "",
                    "imageAltText": "Alternate Text",
                    "buttons": [
                      {
                        "type": "postback",
                        "payload": "Postback Value",
                        "title": "Postback Button"
                      }
                    ]
                  },
                  {
                    "title": "List Item 2",
                    "subtitle": "",
                    "imageUrl": "",
                    "defaultActionUrl": "",
                    "buttons": [
                      {
                        "type": "web_url",
                        "payload": "",
                        "title": "URL Button",
                        "url": "https:\\\\www.google.com"
                      }
                    ]
                  },
                  {
                    "title": "List Item 3",
                    "subtitle": "",
                    "imageUrl": "",
                    "defaultActionUrl": "",
                    "buttons": [
                      {
                        "type": "phone_number",
                        "payload": "+491234567",
                        "title": "Phone Number Button"
                      }
                    ]
                  },
                  {
                    "title": "List Item 4",
                    "subtitle": "",
                    "imageUrl": "",
                    "defaultActionUrl": "",
                    "buttons": [
                      {
                        "type": "trigger_intent",
                        "payload": "",
                        "title": "Intent Button",
                        "intentName": "Intent Name"
                      }
                    ]
                  }
                ],
                "button": {
                  "type": "postback",
                  "condition": "",
                  "payload": "Postback value",
                  "title": "Separate Button from List"
                }
              }
            }
          }
        });
        ```
      </Tab>

      <Tab title="Say Node">
        <div>
          <Frame>
            <img alt="List output configured in a Say Node" />
          </Frame>

          <Frame>
            <img alt="List item settings in the Say Node editor" />
          </Frame>
        </div>
      </Tab>

      <Tab title="Webchat">
        <Frame>
          <img alt="List items rendered in Webchat" />
        </Frame>
      </Tab>
    </Tabs>
  </Accordion>

  <Accordion title="Audio">
    Outputs audio from a URL.

    <Tabs>
      <Tab title="JSON Format">
        ```js theme={null}
        api.output(null, {
          "type": "audio",
          "_cognigy": {
            "_default": {
              "_audio": {
                "type": "audio",
                "audioUrl": "https://audio.test/test.wav"
              }
            }
          }
        });
        ```
      </Tab>

      <Tab title="Say Node">
        <Frame>
          <img alt="Audio output configured in a Say Node" />
        </Frame>
      </Tab>

      <Tab title="Webchat">
        <Frame>
          <img alt="Audio player rendered in Webchat" />
        </Frame>
      </Tab>
    </Tabs>
  </Accordion>

  <Accordion title="Image">
    Outputs image from a URL.

    <Tabs>
      <Tab title="JSON Format">
        ```js theme={null}
        api.output(null, {
          "type": "image",
          "_cognigy": {
            "_default": {
              "_image": {
                "type": "image",
                "imageUrl": "https://testURL.test/image.jpeg"
              }
            }
          }
        });
        ```
      </Tab>

      <Tab title="Say Node">
        <Frame>
          <img alt="Image output configured in a Say Node" />
        </Frame>
      </Tab>

      <Tab title="Webchat">
        <Frame>
          <img alt="Image rendered in Webchat" />
        </Frame>
      </Tab>
    </Tabs>
  </Accordion>

  <Accordion title="Video">
    Outputs video from a URL.

    <Tabs>
      <Tab title="JSON Format">
        ```js theme={null}
        api.output(null, {
          "type": "video",
          "_cognigy": {
            "_default": {
              "_video": {
                "type": "video",
                "videoUrl": "https://www.youtube.com/watch?v=iGc3fekZQOw"
              }
            }
          }
        });
        ```
      </Tab>

      <Tab title="Say Node">
        <Frame>
          <img alt="Video output configured in a Say Node" />
        </Frame>
      </Tab>

      <Tab title="Webchat">
        <Frame>
          <img alt="Video player rendered in Webchat" />
        </Frame>
      </Tab>
    </Tabs>
  </Accordion>

  <Accordion title="Adaptive Card">
    Outputs an adaptive card.

    <Tabs>
      <Tab title="JSON Format">
        ```js theme={null}
        api.output(null, {
          "type": "adaptiveCard",
          "_cognigy": {
            "_default": {
              "_adaptiveCard": {
                "type": "adaptiveCard",
                "adaptiveCard": {
                  "$schema": "http://adaptivecards.io/schemas/adaptive-card.json",
                  "type": "AdaptiveCard",
                  "body": [
                    {
                      "type": "TextBlock",
                      "size": "Medium",
                      "weight": "Bolder",
                      "text": "Publish Adaptive Card Schema"
                    },
                    {
                      "type": "ColumnSet",
                      "columns": [
                        {
                          "type": "Column",
                          "items": [
                            {
                              "type": "Image",
                              "style": "Person",
                              "url": "https://pbs.twimg.com/profile_images/3647943215/d7f12830b3c17a5a9e4afcc370e3a37e_400x400.jpeg",
                              "size": "Small"
                            }
                          ],
                          "width": "auto"
                        },
                        {
                          "type": "Column",
                          "items": [
                            {
                              "type": "TextBlock",
                              "weight": "Bolder",
                              "text": "Matt Hidinger",
                              "wrap": true
                            },
                            {
                              "type": "TextBlock",
                              "spacing": "None",
                              "text": "Created Tue, Feb 14, 2017",
                              "isSubtle": true,
                              "wrap": true
                            }
                          ],
                          "width": "stretch"
                        }
                      ]
                    },
                    {
                      "type": "TextBlock",
                      "text": "Publish Adaptive Card Schema easily.",
                      "wrap": true
                    },
                    {
                      "type": "Action.OpenUrl",
                      "title": "View",
                      "url": "https://adaptivecards.io"
                    }
                  ],
                  "actions": [],
                  "version": "1.5"
                }
              }
            }
          }
        });
        ```
      </Tab>

      <Tab title="Webchat">
        <Frame>
          <img alt="Adaptive Card rendered in Webchat" />
        </Frame>
      </Tab>
    </Tabs>
  </Accordion>
</AccordionGroup>


# Analytics Data
Source: https://docs.cognigy.com/ai/for-developers/code/analytics-data

Learn how to read and modify analytics data in Cognigy Code Nodes, including Intent, custom fields, and Flow-level tracking for improved analytics.

<a href="/release-notes/4.90"><Badge>Updated in 4.90</Badge></a>

In a [Code Node](/ai/for-developers/code/overview), you can modify certain parts of the analytics data to get more control of what exactly is written to analytics.

You can access the analytics data in a Code Node by using the key `analyticsdata`. For example:

```js theme={null}
analyticsdata.intent = "test";
```

## Analytics Data

<Warning>
  As of Cognigy 4.91.0, the Understood Messages concept is deprecated. Use the [Goals and Tasks](https://docs.cognigy.com/ai/analyze/goals-and-tasks/overview/) concepts to evaluate whether your AI Agent understood a particular phrase. In Cognigy 2025.19, the charts related to this concept were removed from Insights reports.

  As of Cognigy 4.91.0, the `understood` field in the Cognigy OData endpoint is deprecated. The `understood` field has been removed from the Cognigy OData endpoint. Use the [Goals collection](https://docs.cognigy.com/ai/analyze/odata#goals) in the Cognigy OData endpoint.

  Refer to the list of all upcoming removals on the [Deprecations and Removals](https://docs.cognigy.com/release-notes/deprecations-and-removals/) page.
</Warning>

<Accordion title="Analytics Data Fields">
  | Field                  | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                | Type              |
  | ---------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ----------------- |
  | intent                 | The intent that will be written to analytics. You can for instance modify this field to store the intent that was found in an executed Flow in analytics.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  | String            |
  | intentFlow             | The ID of the Flow where the intent was found (main Flow or attached Flow).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                | String            |
  | intentScore            | The score of the Intent that was found.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    | Number            |
  | inputText              | The input text.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            | String            |
  | inputData              | The input data object.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     | `{"key":"value"}` |
  | state                  | The State the Flow is currently in.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        | String            |
  | slots                  | The slots that were found in the user's input.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             | Object            |
  | completedGoals         | A list of goals the user has completed.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    | List of Strings   |
  | understood             | An input is considered `understood` if any of the following conditions are met:<ul><li>An Intent or Slot is identified.</li><li>The input is marked as understood through the Nodes: Code, Overwrite Analytics, Search Extract Output, or AI Agent.</li><li>The sentence type is `pAnswer`, `nAnswer`, or `Greeting` (if **Confirmation Word logic** is activated).</li></ul><br />An input is considered `not understood` if any of the following conditions are met:<ul><li>There is an active handover without triggering an Agent Copilot Whisper Flow.</li><li>The message is explicitly marked as `Don't count` or `Null`.</li></ul> | Boolean           |
  | handoverEscalations    | A counter of how many times the session was escalated from the AI Agent to a human agent.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  | Number            |
  | inHandoverRequest      | A flag indicating whether a handover has been initiated but not yet accepted: `true` means waiting in the queue, `false` means there is no pending handover.                                                                                                                                                                                                                                                                                                                                                                                                                                                                               | Boolean           |
  | inHandoverConversation | A flag indicating whether a human agent is actively handling the conversation: `true` means active handling, `false` means there is no human agent involvement.                                                                                                                                                                                                                                                                                                                                                                                                                                                                            | Boolean           |
  | custom1                | A custom field that can be set to any string value during Flow execution. Defaults to null.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                | String            |
  | custom2                | A custom field that can be set to any string value during Flow execution. Defaults to null.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                | String            |
  | custom3                | A custom field that can be set to any string value during Flow execution. Defaults to null.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                | String            |
  | custom4                | A custom field that can be set to any string value during Flow execution. Defaults to null.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                | String            |
  | custom5                | A custom field that can be set to any string value during Flow execution. Defaults to null.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                | String            |
  | custom6                | A custom field that can be set to any string value during Flow execution. Defaults to null.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                | String            |
  | custom7                | A custom field that can be set to any string value during Flow execution. Defaults to null.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                | String            |
  | custom8                | A custom field that can be set to any string value during Flow execution. Defaults to null.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                | String            |
  | custom9                | A custom field that can be set to any string value during Flow execution. Defaults to null.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                | String            |
  | custom10               | A custom field that can be set to any string value during Flow execution. Defaults to null.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                | String            |
</Accordion>

## Custom Fields

Custom fields let you add extra data to your Analytics reports.
This feature helps you analyze your data more deeply and get insights beyond the standard metrics. By customizing your data, you can make better decisions that fit your specific business needs and goals.

### Limitations

* The maximum length of a custom field is 1024 characters.

<Accordion title="Store Complex Custom Data">
  For more complex data, you can use a string field object.

  In this example, `custom1` stores a simple string value, while `custom2` uses a JSON representation of the `customData` object to store more complex information.

  ```js theme={null}
  const customData = {
    key: "value"
  };

  analyticsdata.custom1 = "userFocusGroup: focusGroup1";
  analyticsdata.custom2 = JSON.stringify(customData);
  ```
</Accordion>

## Analytics Data Lifespan

The analyticsdata object is generated for each Flow Execution.
This means that any data you store in the analyticsdata object
(for example, a custom field) will only be stored in analytics for this Flow Execution.
On the next Flow Execution, there will be a new analyticsdata object with a new Intent.


# API Functions
Source: https://docs.cognigy.com/ai/for-developers/code/api-functions



<a href="/release-notes/2025.24"><Badge>Updated in 2025.24</Badge></a>

In a Code Node, you can use the functions of the `api` classes. These functions are also available in Extension Nodes. For example, you can use the `api` object to execute `api.say()`.

<Note>
  Previously, you could use both `actions` and `api`. Now, only `api` is supported. Use `api` functions for all implementations.
</Note>

## CognigyScript

Use these functions to parse and evaluate CognigyScript expressions in text or conditions at run-time.

<AccordionGroup>
  <Accordion title="api.parseCognigyScript">
    Parses CognigyScript text or conditions.

    **Parameters**

    | Parameter | Type    | Description                                    |
    | --------- | ------- | ---------------------------------------------- |
    | text      | string  | The text containing CognigyScript expressions. |
    | condition | boolean | The CognigyScript condition to evaluate.       |

    **Returns**

    `Promise<any>`

    **Example**

    ```js theme={null}
    // Parse CognigyScript as text
    const parsedText = api.parseCognigyScript(
    "Hello {{cc.text}}, welcome!",
    false
    );
    api.say(parsedText);

    // Parse CognigyScript as a condition
    const parsedCondition = api.parseCognigyScript(
    "{{input.currentTime.year > 2024}}",
    true
    );

    if (parsedCondition === "true") {
    api.say("Condition is true");
    } else {
    api.say("Condition is false");
    }
    ```
  </Accordion>

  <Accordion title="api.parseCognigyScriptCondition">
    Evaluates a CognigyScript expression and returns a `true` or `false` value.

    CognigyScript allows you to embed expressions like `{{context.value > 5}}` within strings. This function extracts and evaluates such an expression, used in logic checks for Flows, conditions, or custom functions.

    This function is useful for dynamically checking conditions at run-time based on session data and information from the Input, Context, or Profile objects.

    **Parameters**

    | Parameter | Type   | Description                              |
    | --------- | ------ | ---------------------------------------- |
    | condition | string | The CognigyScript condition to evaluate. |

    **Returns**

    `Promise<string>`

    **Example**

    ```js theme={null}
    const parseCognigyScriptCondition = api.parseCognigyScriptCondition(
    "{{input.currentTime.year > 2024}}"
    );

    api.say(`${parseCognigyScriptCondition}`);
    ```
  </Accordion>

  <Accordion title="api.parseCognigyScriptText">
    Parses and renders a text string containing CognigyScript expressions. Each CognigyScript expression, for example, `{{cc.userName}}`, is evaluated using the current execution context.

    **Parameters**

    | Parameter | Type   | Description                                    |
    | --------- | ------ | ---------------------------------------------- |
    | text      | string | The text containing CognigyScript expressions. |

    **Returns**

    `Promise<string>`

    **Example**

    ```js theme={null}
    const output = api.parseCognigyScriptText("Hello {{cc.text}}, welcome to {{cc.city}}!");
    api.say(`The current state is: ${output}`);
    ```
  </Accordion>
</AccordionGroup>

## Contact Profiles

Use these functions to activate, deactivate, delete, merge, and update Contact Profiles during a conversation.

<AccordionGroup>
  <Accordion title="api.activateProfile">
    Activates the Contact Profile.
    Call this function to reactivate read and write operations on a previously deactivated Contact Profile.

    **Parameters**

    None

    **Returns**

    `Promise<any>`

    **Example**

    ```js theme={null}
    // Reactivate a previously deactivated Profile
    api.activateProfile();
    ```
  </Accordion>

  <Accordion title="api.addContactMemory">
    Adds an entry to the `memories` array of the [Profile](/ai/agents/develop/ai-agent-memory/profile) object.

    **Parameters**

    | Parameter | Type   | Description                                                                   |
    | --------- | ------ | ----------------------------------------------------------------------------- |
    | memory    | string | The text to add as a new entry in the `memories` array of the Profile object. |

    **Returns**

    `void`

    **Example**

    ```js theme={null}
    api.addContactMemory("ppl");
    ```
  </Accordion>

  <Accordion title="api.deactivateProfile">
    Deactivates the Contact Profile. While the Contact Profile is deactivated, no data can be read from or written to the Contact Profile.

    **Parameters**

    None

    **Returns**

    `Promise<any>`

    **Example**

    ```js theme={null}
    // Deactivate a Contact Profile
    api.deactivateProfile();
    ```
  </Accordion>

  <Accordion title="api.deleteProfile">
    Permanently deletes the Contact Profile from the system.

    **Parameters**

    None

    **Returns**

    `Promise<any>`

    **Example**

    ```js theme={null}
    // Delete the Contact Profile
    api.deleteProfile();
    ```
  </Accordion>

  <Accordion title="api.mergeProfile">
    Merges the current Contact Profile with another one.
    This function is useful for combining duplicate Contact Profiles or moving data between users.

    **Parameters**

    | Parameter | Type   | Description                                  |
    | --------- | ------ | -------------------------------------------- |
    | contactId | string | The ID of the Contact Profile to merge into. |

    **Returns**

    `Promise<any>`

    **Example**

    ```js theme={null}
    api.mergeProfile("c-89374201a7bd5a9e4");
    ```
  </Accordion>

  <Accordion title="api.updateProfile">
    Updates a value in the current Contact Profile.

    **Parameters**

    | Parameter | Type   | Description                                     |
    | --------- | ------ | ----------------------------------------------- |
    | key       | string | The key of the Contact Profile field to update. |
    | value     | any    | The value to set for the specified key.         |

    **Returns**

    `Promise<any>`

    **Example**

    ```js theme={null}
    // Update the 'email' field of the current Contact Profile
    api.updateProfile("email", "user@example.com");

    // Update a custom string field
    api.updateProfile("subscriptionStatus", "active");

    // Update a numeric field
    api.updateProfile("loyaltyPoints", 120);

    // Update a boolean field
    api.updateProfile("marketingConsent", true);
    ```
  </Accordion>
</AccordionGroup>

## Context Object

Use these functions to read, write, and reset the Context object so you can store and manipulate conversation-scoped data across Nodes.

<AccordionGroup>
  <Accordion title="api.addToContext">
    Adds a value to the Context object, either replacing an existing value or appending to it as an array.

    **Parameters**

    | Parameter | Type   | Description                                                                                                                                                   |
    | --------- | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | key       | string | The key to add to the Context object.                                                                                                                         |
    | value     | any    | The value to assign to the key.                                                                                                                               |
    | mode      | string | Determines how the value is added: <ul><li>`"simple"` – replaces any existing value.</li><li>`"array"` – appends to the existing value as an array.</li></ul> |

    **Returns**

    `void`

    **Example**

    ```javascript theme={null}
    // Simple mode: replaces existing value
    api.addToContext("user.city", "Berlin", "simple");
    // Context: { user: { city: "Berlin" } }

    api.addToContext("user.city", "Munich", "simple");
    // Context: { user: { city: "Munich" } }

    // Array mode: appends to an array
    api.addToContext("user.purchases", "book", "array");
    // Context: { user: { purchases: ["book"] } }

    api.addToContext("user.purchases", "laptop", "array");
    api.addToContext("user.purchases", "coffee", "array");
    // Context: { user: { purchases: ["book", "laptop", "coffee"] } }
    ```
  </Accordion>

  <Accordion title="api.deleteContext">
    Deletes the key and its value from the Context object.

    **Parameters**

    | Parameter | Type   | Description                              |
    | --------- | ------ | ---------------------------------------- |
    | key       | string | The key in the Context object to remove. |

    **Returns**

    `void`

    **Example**

    ```js theme={null}
    api.deleteContext("user.city");
    ```
  </Accordion>

  <Accordion title="api.getContext">
    Retrieves data from the specified key in the Context object.

    **Parameters**

    | Parameter | Type   | Description                                           |
    | --------- | ------ | ----------------------------------------------------- |
    | key       | string | The key to retrieve the data from the Context object. |

    **Returns**

    `any`: The requested context data.

    **Example**

    ```js theme={null}
    api.getContext("user.profile.name");
    ```
  </Accordion>

  <Accordion title="api.getConversationTranscript">
    Retrieves the current conversation transcript, including the last 10 user inputs and the associated outputs from the AI Agent.

    **Parameters**

    | Parameter         | Type   | Description                                                                                                                                                                                                                                                     |
    | ----------------- | ------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | mode              | string | This parameter is optional. Specifies the format of the returned transcript: <ul><li>`"string"` – returns a plain text version of the conversation.</li><li>`"json"` – returns a structured array of message objects. This format is used by default.</li></ul> |
    | options           | object | This parameter is optional. An object to configure the retrieved transcript data.                                                                                                                                                                               |
    | options.turnLimit | number | This parameter is optional. Limits the transcript to the most recent N conversational turns. If the parameter is omitted, the full transcript is returned.                                                                                                      |

    **Returns**

    Either the JSON array of conversation entries or the string representation.

    JSON:

    ```json theme={null}
    [
        { "source":"user", "text":"hello" },
        { "source":"bot", "text":"You said: hello" },
        { "source":"user", "text":"you are an ai agent" },
        { "source":"bot", "text":"You said: you are an ai agent" },
        { "source":"user", "text":"show transcript" }
    ]
    ```

    String:

    ```text theme={null}
    - USER: hello
    - BOT: You said: hello
    - USER: you are an ai agent
    - BOT: You said: you are an ai agent
    - USER: show transcript
    ```

    **Example**

    JSON:

    ```js theme={null}
    // Get the transcript
    const transcriptJson = api.getConversationTranscript({ turnLimit: 5 });

    // Convert the JSON array to a string
    const transcriptJsonPrint= JSON.stringify(transcriptJson, null, 2);

    // Send it to the user
    api.say(transcriptJsonPrint);
    ```

    String:

    ```js theme={null}
    // Get the transcript
    const transcriptString = api.getConversationTranscript("string", { turnLimit: 1 });

    // Send it to the user
    api.say(transcriptString);
    ```
  </Accordion>

  <Accordion title="api.removeFromContext">
    Removes information from the Context object either by deleting a key entirely or by removing specific entries from an array stored at that key.

    **Parameters**

    | Parameter | Type   | Description                                                                                                                                                       |
    | --------- | ------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | key       | string | The key in the Context object to remove information from.                                                                                                         |
    | value     | any    | The value to remove when using `"array"` mode. Ignored in `"simple"` mode.                                                                                        |
    | mode      | string | Determines how the removal is performed: <ul><li>`"simple"` – removes the key entirely.</li><li>`"array"` – removes only matching values from an array.</li></ul> |

    **Returns**

    `void`

    **Example**

    ```js theme={null}
    // Simple mode: remove the entire key
    api.removeFromContext("user.city", null, "simple");
    // Context: { user: { city: undefined } }

    // Array mode: remove specific values from an array
    api.addToContext("user.purchases", "book", "array");
    api.addToContext("user.purchases", "laptop", "array");
    api.addToContext("user.purchases", "coffee", "array");

    // Remove "laptop" from the purchases array
    api.removeFromContext("user.purchases", "laptop", "array");
    // Context: { user: { purchases: ["book", "coffee"] } }
    ```
  </Accordion>

  <Accordion title="api.resetContext">
    Resets the Context object for the current conversation. All keys and values added during the conversation are removed.

    **Parameters**

    None

    **Returns**

    `void`

    **Example**

    ```js theme={null}
    // Remove all context data added during the conversation
    api.resetContext();
    ```
  </Accordion>

  <Accordion title="api.setContext">
    Sets the value for a specific key in the Context object, replacing any existing value.

    **Parameters**

    | Parameter | Type   | Description                           |
    | --------- | ------ | ------------------------------------- |
    | key       | string | The key in the Context object to set. |
    | value     | any    | The value to set for the key.         |

    **Returns**

    `void`

    **Example**

    ```js theme={null}
    api.setContext("user.city", "Berlin");
    ```
  </Accordion>
</AccordionGroup>

## Encoding and Decoding

Use these functions to handle string encoding and decoding, such as converting text to Base64 for safe transmission or decoding Base64 strings back to UTF-8.

<AccordionGroup>
  <Accordion title="api.base64Decode">
    Base64 decodes a given string.

    **Parameters**

    | Parameter | Type   | Description               |
    | --------- | ------ | ------------------------- |
    | data      | string | The string to be decoded. |

    **Returns**

    `string`: The decoded string as UTF-8.

    **Example**

    ```js theme={null}
    const decoded = api.base64Decode("SGVsbG8sIFdvcmxkIQ==");
    api.say(`Decoded string: ${decoded}`);
    // Output: "Hello, World!"
    ```
  </Accordion>

  <Accordion title="api.base64Encode">
    Base64 encodes a given string.

    **Parameters**

    | Parameter | Type   | Description               |
    | --------- | ------ | ------------------------- |
    | data      | string | The string to be encoded. |

    **Returns**

    `string`: The base64 encoded string.

    **Example**

    ```js theme={null}
    const encoded = api.base64Encode("Hello, World!");
    api.say(`Encoded string: ${encoded}`);
    // Output: "SGVsbG8sIFdvcmxkIQ=="
    ```
  </Accordion>
</AccordionGroup>

## Execution Control

Use these functions to control Flow execution, including adding conditional Entrypoints, jumping to specific Nodes or Flows, resetting overrides, stopping execution, and simulating new inputs.

<AccordionGroup>
  <Accordion title="api.addConditionalEntrypoint">
    Adds a conditional Entrypoint for the current conversation.

    If `{condition}` is met within the next `{retentionTime}` inputs, Flow execution starts at `{entrypoint}`.

    **Parameters**

    | Parameter     | Type   | Description                                                                                                   |
    | ------------- | ------ | ------------------------------------------------------------------------------------------------------------- |
    | entrypoint    | string | The ID of the Node to execute when the condition evaluates to `true`.                                         |
    | retentionTime | number | The number of times this conditional Entrypoint can trigger before it is automatically removed.               |
    | condition     | string | A string expression evaluated. For example, `input.text.includes('@')` checks if the user input contains `@`. |

    **Returns**

    `void`

    **Example**

    ```js theme={null}
    api.addConditionalEntrypoint({
    entrypoint: "c40c50f3-2f51-422c-8a3e-748e7cc92ed0",
    retentionTime: 2,
    condition: "input.text.includes('@')"
    });
    ```
  </Accordion>

  <Accordion title="api.resetNextNodes">
    Clears any active overrides for the next Node in the Flow.

    This function is useful when you've changed the next Node using the `api.setNextNode` function, and you want to return to the default Flow behavior.

    **Parameters**

    None

    **Returns**

    `void`

    **Example**

    ```js theme={null}
    // Imagine you previously used api.setNextNode("fallback-node") to redirect the Flow.
    // Now, you want to clear that override so the Flow continues as originally defined.

    api.resetNextNodes();
    ```
  </Accordion>

  <Accordion title="api.setNextNode">
    Sets the Node with the given `nodeId` to be executed immediately after the current one.
    This function is used by If and Lookup Nodes to select child Nodes based on their configuration.

    **Parameters**

    | Parameter | Type   | Description                                                                                                                          |
    | --------- | ------ | ------------------------------------------------------------------------------------------------------------------------------------ |
    | nodeId    | string | The ID of the Node to execute next.                                                                                                  |
    | newFlowId | string | The ID of the Flow containing the target Node. This parameter is optional and is required if the target Node is in a different Flow. |

    **Returns**

    `void`

    **Example**

    ```js theme={null}
    // The current Node is executing inside a Flow, and you want to control which Node should be executed next based on some logic.
    // In this example, the execution will jump to a Node with the nodeId "c4c5c267-55d8-44ae-a123-f351d237f496" immediately after the current Node finishes.

    api.setNextNode("c4c5c267-55d8-44ae-a123-f351d237f496");

    // To continue the conversation in a different Flow after the current Node, specify both the target nodeId and the flowId.
    // The Node with nodeId "c4c5c267-55d8-44ae-a123-f351d237f496" must be marked as an Entrypoint in the target Flow with flowId "67f511b11acdc139f91946b8".
    // Execution will jump directly to that Node after the current one finishes.

    api.setNextNode("c4c5c267-55d8-44ae-a123-f351d237f496", "67f511b11acdc139f91946b8");
    ```
  </Accordion>

  <Accordion title="api.stopExecution">
    Stops the execution of the Flow.

    **Parameters**

    None

    **Returns**

    `void`

    **Example**

    ```js theme={null}
    api.stopExecution();
    ```
  </Accordion>

  <Accordion title="api.thinkV2">
    Restarts Flow execution with a simulated input using the provided text and data.

    **Parameters**

    | Parameter | Type   | Description                                                            |
    | --------- | ------ | ---------------------------------------------------------------------- |
    | text      | string | A simulated input text                                                 |
    | data      | object | A simulated input data object with optional fields and additional keys |

    **Returns**

    `void`

    **Example**

    ```js theme={null}
    api.thinkV2( "Hello, how's the weather?", { location: "Berlin", units: "Celsius" });
    ```
  </Accordion>
</AccordionGroup>

## Input and Output

Use these functions to enrich the Input object and send text or data responses to the user from a Code or Extension Node.

<AccordionGroup>
  <Accordion title="api.addToInput">
    Adds a key-value pair to the Input object.

    **Parameters**

    | Parameter | Type   | Description                         |
    | --------- | ------ | ----------------------------------- |
    | key       | string | The key to add to the Input object. |
    | value     | string | The value to assign to the key.     |

    **Returns**

    `void`

    **Example**

    ```js theme={null}
    api.addToInput("username", "John Doe");
    ```
  </Accordion>

  <Accordion title="api.output / api.say">
    Executes a basic output to the contact, similar to the Say Node.

    **Parameters**

    | Parameter | Type   | Description                               |
    | --------- | ------ | ----------------------------------------- |
    | text      | string | The output text that is sent to the user. |
    | data      | any    | The output data that is sent to the user. |

    <Danger>
      `_cognigy` is an internal parameter and not part of the public API. This parameter is unsupported and may change without notice.
    </Danger>

    **Returns**

    `void`

    **Example**

    ```js theme={null}
    // Send text with additional data
    api.say("Hello, this is a message from the Code Node!", { customData: 123 });

    // `api.output` is an alias of `api.say`
    api.output("Hello, this is a message from the Code Node!", { customData: 123 });
    ```
  </Accordion>
</AccordionGroup>

## Logging and Monitoring

Use these functions to write Project logs, send debug messages, inspect LLM token usage, track analytics steps, and configure sensitive logging behavior.

<AccordionGroup>
  <Accordion title="api.completeGoal">
    Completes the task. The information that the task has been completed is recorded and is available in Cognigy Insights and the Cognigy OData endpoint.

    **Parameters**

    | Parameter | Type   | Description           |
    | --------- | ------ | --------------------- |
    | name      | string | The name of the task. |

    **Returns**

    `Promise<void>`

    **Example**

    ```js theme={null}
    api.completeGoal("Order Pizza");
    ```
  </Accordion>

  <Accordion title="api.emitToOpsCenter">
    Forwards an error to the Ops Center Error page. Automatically sets the component to `Flow` and the subcomponent to `FlowNodeExecution`.

    **Parameters**

    | Parameter | Type   | Description                                   |
    | --------- | ------ | --------------------------------------------- |
    | title     | string | The title of the error to log.                |
    | projectId | string | The ID of the Project where the error occurs. |

    **Returns**

    `void`

    **Example**

    ```js theme={null}
    api.emitToOpsCenter({
        projectId: "6874becb018c552bc267cc51",
        title: "Something went wrong",
    });
    ```
  </Accordion>

  <Accordion title="api.getLLMTokenUsageForSession">
    Retrieves the LLM token usage information for the current session.

    **Parameters**

    None

    **Returns**

    ```json theme={null}
    {
        "llmDisplayName": string,
        "providerType": string,
        "modelType": string,
        "usage": {
        "inputTokens": number,
        "outputTokens": number
        }
    }
    ```

    **Example**

    ```js theme={null}
    api.getLLMTokenUsageForSession();
    ```
  </Accordion>

  <Accordion title="api.log">
    Writes a message to the Project logs.

    **Parameters**

    | Parameter | Type   | Description                                                                                                                                                                                                                                                                                                    |
    | --------- | ------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | level     | string | The log level. Cognigy includes the following log levels:<ul><li>`info` — general information about the normal operation of a Flow.</li><li>`error` — information about errors that occur during Flow execution.</li><li>`debug` — detailed information about Flow execution for debugging purposes.</li></ul> |
    | text      | string | The message to write to the logs.                                                                                                                                                                                                                                                                              |

    **Returns**

    `void`

    **Example**

    ```js theme={null}
    api.log("info", "This is an informational message.");
    api.log("error", "An error occurred during Flow execution.");
    api.log("debug", "Debugging details for Flow execution.");
    ```
  </Accordion>

  <Accordion title="api.logDebugMessage / api.logDebugError">
    Sends a debug message. The message is visible only in the Interaction Panel with activated debug mode.

    **Parameters**

    | Parameter | Type            | Description                                                                                                          |
    | --------- | --------------- | -------------------------------------------------------------------------------------------------------------------- |
    | message   | string / object | The debug message to log. When logging an object, use `JSON.stringify` to make it readable in the Interaction Panel. |
    | header    | string          | An optional label or header for the log message, used to categorize it in the debug panel.                           |

    **Returns**

    `void`

    **Example**

    ```js theme={null}
    // Log a simple string
    api.logDebugMessage("Flow started successfully", "Flow Info");

    // Log multiple variables together
    const sessionInfo = { sessionId: "abc123", timestamp: new Date().toISOString() };
    const userData = { name: "Alice", age: 25 };

    // Combine objects into one for logging
    api.logDebugMessage(JSON.stringify({ sessionInfo, userData }, null, 2), "Session & User Info");
    ```
  </Accordion>

  <Accordion title="api.setSensitiveLoggingSettings">
    Configures masking of sensitive data in logs, analytics, and conversation content for the session.

    **Parameters**

    | Parameter            | Type    | Description                                                                                                                                                                                                                                 |
    | -------------------- | ------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | maskLogging          | boolean | Masks sensitive information in system logs on the Logs page. Replaces fields like `userId`, `text`, and `data` with asterisks to protect user privacy.                                                                                      |
    | maskAnalytics        | boolean | Masks sensitive information in Input, Insights, and Analytics (`inputText`, `inputData`, `userLanguageText`, and user IP). Disables the Intent Trainer because input text isn't stored. Can selectively ignore analytics using a Code Node. |
    | maskIPAddress        | boolean | This parameter is optional. Masks the user's IP address in the Input object, Insights, and Analytics. IP addresses are replaced with asterisks and aren't available via OData Analytics.                                                    |
    | disableConversations | boolean | Disables logging of conversation content entirely. No conversation data will be stored.                                                                                                                                                     |
    | disableIntentTrainer | boolean | This parameter is optional. Disables logging for the Intent Trainer, preventing user input from being stored or used for training intents.                                                                                                  |
    | traceId              | string  | This parameter is optional. A unique ID for tracing this API call. Helps track requests in logs and diagnostics without affecting masking or logging behavior.                                                                              |

    **Returns**

    `void`

    **Example**

    ```js theme={null}
    api.setSensitiveLoggingSettings({
        maskLogging: true, // Mask sensitive information in conversation logs
        maskAnalytics: true, // Mask sensitive information in analytics and Insights
        maskIPAddress: true, // Optional: mask user IP addresses
        disableConversations: true, // Disable conversation logging entirely
        disableIntentTrainer: true // Disable logging for the Intent Trainer
        }, "trace-12345"); // Optional trace ID for debugging
    ```
  </Accordion>

  <Accordion title="api.trackAnalyticsStep">
    Tracks a custom analytics step for the session. Useful for monitoring user interactions or workflow progress.

    **Parameters**

    | Parameter | Type   | Description                               |
    | --------- | ------ | ----------------------------------------- |
    | stepLabel | string | The label of the analytics step to track. |

    **Returns**

    `void`

    **Example**

    ```javascript theme={null}
    // Track that the user reached the "Checkout" step
    api.trackAnalyticsStep("Checkout");
    ```
  </Accordion>
</AccordionGroup>

## States (deprecated)

<Warning>
  As of Cognigy 2026.7.0, States are deprecated. The removal of States is planned for Cognigy 2026.12.0. Before the removal, delete your States and retrain your NLU model so it no longer uses state-based Intent masking. For alternative approaches to control Intent recognition, use [Intent Conditions](/ai/platform-features/nlu/intents/conditions).
</Warning>

Use these functions to get, set, and reset the conversation State of a Flow.

<AccordionGroup>
  <Accordion title="api.getState">
    Returns the current State of the conversation.

    **Parameters**

    None

    **Returns**

    `string`

    **Example**

    ```js theme={null}
    // Get the current State of the Flow
    const currentState = api.getState();

    // Send the State to the user
    api.say(`The current state is: ${currentState}`);
    ```
  </Accordion>

  <Accordion title="api.resetState">
    Resets the State of the Flow to the initial State.

    **Parameters**

    None

    **Returns**

    `Promise<string>`

    **Example**

    ```js theme={null}
    api.resetState();
    ```
  </Accordion>

  <Accordion title="api.setState">
    Updates the conversation State to a specified State.

    **Parameters**

    | Parameter | Type   | Description                         |
    | --------- | ------ | ----------------------------------- |
    | state     | string | The State that should be activated. |

    **Returns**

    `void`

    **Example**

    ```js theme={null}
    // Set the conversation state to "awaiting_input"
    api.setState("awaiting_input");

    // Verify by getting the current state
    const currentState = api.getState();
    api.say(`The current state is: ${currentState}`);
    ```
  </Accordion>
</AccordionGroup>

## xApps

Use these functions to manage conversations with xApps.

<Accordion title="api.setAppState">
  Updates the state of the xApp Session for the current conversation. Only works if an xApp Session was started beforehand.

  **Parameters**

  | Parameter       | Type   | Description                                                                                                                                                               |
  | --------------- | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
  | appTemplateId   | string | The xApp template identifier. Can be a URL to a hosted xApp template, `generic-html` to load inline HTML with the xApp page SDK, or `null` to close the active xApp page. |
  | appTemplateData | object | The data payload passed to the xApp page. For `generic-html`, this must include an `html` field containing the HTML to render.                                            |

  **Returns**

  `void`

  **Example**

  ```js theme={null}
  // Load an xApp template hosted at a URL
  api.setAppState("https://my-xapp-template-url", {
  my: "xapp-template-data",
  });

  // Load a generic HTML xApp page
  const html = `
  <html>
  <body>
  Hello, World!
  <button type="button" onclick="SDK.submit({});">continue</button>
  <script src="/sdk/app-page-sdk.js"></script>
  </body>
  </html>
  `;

  api.setAppState("generic-html", { html });

  // Close the active xApp page
  api.setAppState(null, {});
  ```
</Accordion>

## Other

Use these functions to manage conversation behavior, such as adding Keyphrases, handling handovers, and localization settings.

<AccordionGroup>
  <Accordion title="api.addLexiconKeyphrase">
    Adds a new Keyphrase to a Lexicon.

    **Parameters**

    | Parameter | Type      | Description                                                                                           |
    | --------- | --------- | ----------------------------------------------------------------------------------------------------- |
    | lexiconId | string    | The ID of the Lexicon to which the Keyphrase will be added.                                           |
    | keyphrase | string    | The primary phrase to add to the Lexicon.                                                             |
    | tags      | string\[] | An array of strings used to categorize or filter the Keyphrase.                                       |
    | synonyms  | string\[] | An array of strings representing alternative phrases that should be recognized as the same Keyphrase. |
    | data      | object    | This parameter is optional. The custom data associated with the Keyphrase.                            |

    **Returns**

    `Promise<string>`

    **Example**

    ```js theme={null}
    // Add a new keyphrase to the existing Lexicon
    api.addLexiconKeyphrase(
      "6942550bf16d8343c6f1adec",
      "password reset",
      ["account", "security"], // tags
      ["reset password", "forgot password"], // synonyms
      {
        priority: "high",
        source: "dynamic",
      }
    );
    ```
  </Accordion>

  <Accordion title="api.evaluateRule">
    Evaluates a condition based on a rule configuration that follows the `left`, `operand`, `right` pattern. This function works similarly to the If Node. The `left` value is compared with the `right` value using the specified operator, for example, `!=`, `>`, `includes`. If the condition evaluates to `true`, the associated logic or transition is executed.

    **Parameters**

    | Parameter | Type   | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
    | --------- | ------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | left      | string | The primary value to be evaluated.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | operand   | string | The comparison operator used to evaluate the relationship between the left and right values. Supported operators: <ul><li>`eq` – checks if values are equal. Example: `5 eq 5 → true`</li><li>`neq` – checks if values aren't equal. Example: `5 neq 3 → true`</li><li>`lt` – left is less than right. Example: `3 lt 5 → true`</li><li>`lte` – left is less than or equal. Example: `5 lte 5 → true`</li><li>`gt` – left is greater than right. Example: `7 gt 5 → true`</li><li>`gte` – left is greater than or equal. Example: `5 gte 5 → true`</li><li>`contains` – left contains right. Example: `"hello" contains "ell" → true`</li><li>`ncontains` – left doesn't contain right. Example: `"hello" ncontains "xyz" → true`</li><li>`exists` – checks if a value exists. Example: `x exists → true if x is defined`</li><li>`nexists` – checks if a value doesn't exist. Example: `x nexists → true if x is undefined`</li><li>`isyes` – checks if a value is `"yes"`. Example: `"yes" isyes → true`</li><li>`isno` – checks if value is `"no"`. Example: `"no" isno → true`</li></ul>. |
    | right     | string | The reference value to compare against the `left` value. This parameter is ignored for operators that check only the state of the left value, such as `exists`, `nexists`, `isyes`, and `isno`.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |

    **Returns**

    `Promise<boolean>`

    **Example**

    ```js theme={null}
    api.evaluateRule({
        left: "hello world",
        operand: "contains",
        right: "hello"
    });
    ```
  </Accordion>

  <Accordion title="api.handover">
    Initiates a handover to a human agent or another system.

    **Parameters**

    | Parameter          | Type   | Description                                                             |
    | ------------------ | ------ | ----------------------------------------------------------------------- |
    | text               | string | The message to send to the user.                                        |
    | cancelIntent       | string | The Intent that will be triggered if the user cancels the handover.     |
    | unavailable        | string | The message shown to the user if no agent is available.                 |
    | unsupportedChannel | string | The message shown if the current channel doesn't support handover.      |
    | quickReply         | string | The text for a quick reply button offered to the user.                  |
    | chatwootInboxId    | string | The ID of the Chatwoot inbox where the conversation should be routed.   |
    | liveAgentInboxId   | string | The ID of the Live Agent inbox where the conversation should be routed. |

    **Returns**

    `Promise<void>`

    **Example**

    ```js theme={null}
    api.handover({
        text: "Connecting you to a human agent...",
        cancelIntent: "CancelHandover",
        unavailable: "No agents are available at the moment.",
        unsupportedChannel: "This channel doesn't support handover.",
        quickReply: "Talk to an agent",
        chatwootInboxId: "chatwoot_123",
        liveAgentInboxId: "liveagent_456"
    });
    ```
  </Accordion>

  <Accordion title="api.setLocaleReferenceId">
    Sets the Reference ID to specify which Locale is used for localization.

    **Parameters**

    | Parameter         | Type   | Description                     |
    | ----------------- | ------ | ------------------------------- |
    | localeReferenceID | string | The Locale Reference ID to set. |

    **Returns**

    `void`

    **Example**

    ```js theme={null}
    api.setLocaleReferenceId("b4dc8ff5-bd05-467a-a281-2514122c4aa2")
    ```
  </Accordion>

  <Accordion title="api.setRating">
    Sets a rating for the conversation to prefill a rating prompt with a specific value or update it based on other actions or user inputs. Using this function, you can collect feedback at specific steps.

    **Parameters**

    | Parameter | Type   | Description                                                             |
    | --------- | ------ | ----------------------------------------------------------------------- |
    | rating    | number | The rating for the conversation: `-1` for negative or `1` for positive. |
    | comment   | string | Additional information about the rating.                                |

    **Returns**

    `void`

    **Example**

    ```js theme={null}
    api.setRating({
        rating: 1,
        comment: "The conversation was helpful."
    });
    ```
  </Accordion>

  <Accordion title="api.setTimezoneOffset">
    Sets the time zone offset from UTC for the session, allowing time-related data to be interpreted correctly. You can view the applied time zone in the Input object under `input.currentTime.timezoneOffset`.

    **Parameters**

    | Parameter | Type   | Description               |
    | --------- | ------ | ------------------------- |
    | offset    | number | The time offset in hours. |

    **Returns**

    `void`

    **Example**

    ```js theme={null}
    // Set the session timezone offset to UTC+4
    api.setTimezoneOffset(4);
    ```
  </Accordion>
</AccordionGroup>

## More Information

* [Node Execution API: Function Reference](/ai/for-developers/function-reference/all-functions)


# Channel Output
Source: https://docs.cognigy.com/ai/for-developers/code/channel-output

Send custom JSON payloads to Channels using Cognigy Code Nodes. Covers api.output and api.say for Webchat, Messenger, and other channels.

With a Code Node, you can output a text message along with a data object by using the following `actions` functions:

```js theme={null}
api.output("<YOUR_TEXT>", {"key": "value" });
// OR
api.say("<YOUR_TEXT>", {"key": "value" });
```

The `data` parameter should take the JSON payload that defines the custom channel message:

```js theme={null}
const jsonPayload = {
 "_cognigy":{
  // Channel Specific Payload 
 }
}

api.say("<YOUR_TEXT>", jsonPayload);
```

See the next sections for the required format.

<AccordionGroup>
  <Accordion title="AI Default Channel">
    The Cognigy default channel has a specific channel format which you can find more about at [AI default channel format](/ai/for-developers/code/ai-default-channel-formats).
  </Accordion>

  <Accordion title="Alexa">
    [Alexa Response JSON Reference](https://developer.amazon.com/de/docs/custom-skills/request-and-response-json-reference.html)

    ### Example

    We'll output a simple card to our contact by using the following snippet:

    ```JSON theme={null}
    {
        "_cognigy": {
            "_alexa": {
                "response": {
                    "shouldEndSession": false,
                    "card":  {
                        "type": "Standard",
                        "title": "Title of the card",
                        "content": "Content of a simple card",
                        "text": "Text content for a standard card",
                        "image": {
                            "smallImageUrl": "https://url-to-small-card-image",
                            "largeImageUrl": "https://url-to-large-card-image"
                        }
                    }
                }
            }
        }
    }
    ```
  </Accordion>

  <Accordion title="Messenger">
    [Messenger Messaging Reference](https://developers.facebook.com/docs/messenger-platform/send-messages)

    For the messenger you can send multiple templates by building a JSON object or by using the **facebook-bot-messenger module**. For further information go to their [GitHub Page](https://github.com/snlangsuan/facebook-bot-messenger#readme).

    ### Example

    The following snippet will send a simple text message and a quick reply to the contact:

    ```JSON theme={null}
    {
        "_cognigy": {
            "_facebook": {      
              "message": {
                "text": "Hello World",
                "quick_replies": [
                  {
                    "content_type": "text",
                    "condition": "",
                    "title": "Hi",
                    "image_url": "",
                    "payload": "Hi"
                  }
                ]
              }
            }
        }
    }
    ```

    Or you can use the **facebook-bot-messenger** module for building custom json. Here we'll send quick replies to the contact:

    ```js theme={null}
    // use facebook-bot-messenger to compile reply
    const builder = new MessengerPlatform.QuickRepliesMessageBuilder('Pick a color:');
    builder.addImageOption('Red', 'DEVELOPER_DEFINED_PAYLOAD_FOR_PICKING_RED', 'http://petersfantastichats.com/img/red.png')
           .addImageOption('Green', 'DEVELOPER_DEFINED_PAYLOAD_FOR_PICKING_GREEN', 'http://petersfantastichats.com/img/green.png');

    // output the reply
    api.output("test", { "_cognigy": { "_facebook": {"message": builder.buildMessage() }}});
    ```
  </Accordion>

  <Accordion title="Webchat">
    The Webchat uses the same format as the **Messenger** but instead of adding the `_facebook` property you have to add `_webchat`.

    ### Example

    The following example sends a text message and a quick reply to the contact:

    ```JSON theme={null}
    {
        "_cognigy": {
            "_webchat": {      
                "message": {
                    "text": "Hello World",
                    "quick_replies": [
                        {
                            "content_type": "text",
                            "condition": "",
                            "title": "Hi",
                            "image_url": "",
                            "payload": "Hi"
                        }
                    ]
                }
            }
        }
    }
    ```
  </Accordion>

  <Accordion title="AI">
    The default AI channel, which is always selected from the default, uses the almost same format as the webchat above. The only difference is that you type `_default` instead of `_webchat`.

    ### Example

    The following example sends a text message and a quick reply to the contact:

    ```JSON theme={null}
    {
        "_cognigy": {
            "_default": {      
                "message": {
                    "text": "Hello World",
                    "quick_replies": [
                        {
                            "content_type": "text",
                            "condition": "",
                            "title": "Hi",
                            "image_url": "",
                            "payload": "Hi"
                        }
                    ]
                }
            }
        }
    }
    ```
  </Accordion>

  <Accordion title="LINE">
    [LINE Messaging API](https://developers.line.me/en/reference/messaging-api/#template-messages)

    ### Example

    The following outputs a simple confirm prompt to the contact:

    ```JSON theme={null}
    {
        "_cognigy": {
            "_line": {
                    "type": "template",
                    "altText": "this is a confirm template",
                    "template": {
                        "type": "confirm",
                        "text": "Are you sure?",
                        "actions": [
                            {
                                "type": "message",
                                "label": "Yes",
                                "text": "yes"
                            },
                            {
                                "type": "message",
                                "label": "No",
                                "text": "no"
                            }
                        ]
                    }
            }
        }
    }
    ```
  </Accordion>

  <Accordion title="Smooch">
    [Smooch messaging API](https://docs.smooch.io/rest/?shell#message-types)

    ### Example

    The following outputs a simple image to the end user:

    ```JSON theme={null}
    {
        "role": "appMaker",
        "type": "image",
        "text": "Hello!",
        "mediaUrl": "http://example.org/image.jpg",
        "actions": [{
            "text": "More info",
            "type": "link",
            "uri": "http://example.org"
        }]
    }
    ```
  </Accordion>

  <Accordion title="RingCentral Engage">
    ### Example

    The following outputs a gallery with images to the end user. This example is quite interesting since one needs to [create an attachment for each image](https://developers.ringcentral.com/engage/digital/api-reference/Attachments/createAttachment) using the RingCentral Engage API. The returned **attachmentIds** are inserted in the **Custom JSON** below:

    ```JSON theme={null}
    {
      "command": "structured-content",
      "body": "",
      "structuredContent": {
        "type": "carousel",
        "items": [
            {
                "attachment_id": "5f73470a0e69dc",
                "title": "This is the title",
                "url": "https://url-to-image.com",
                "subtitle": "This is the first subtitle",
                "items": [
                    {
                        "type": "url",
                        "url": "https://www.cognigy.com/",
                        "title": "Visit Website"
                    },
                    {
                        "type": "reply",
                        "payload": "I want to select this one",
                        "title": "Select"
                    }
                ]
            },
            "...": "..."
        ]
      }
    }
    ```
  </Accordion>
</AccordionGroup>


# Modules
Source: https://docs.cognigy.com/ai/for-developers/code/modules

Reference for npm modules available in Cognigy Code Nodes, including Moment.js, Lodash, and other preinstalled packages for data manipulation.

In a Code Node, you can use a selection of predefined NPM modules. On this page, you'll find all the necessary information to run these modules in a Code Node.

<AccordionGroup>
  <Accordion title="Moment.js">
    You can use the moment.js module for parsing, validating, manipulation and displaying dates.

    ## **Usage**

    In a Code Node, call the following function:

    ```js theme={null}
    moment(inp?:any, format?: any, strict?: boolean);
    ```

    ## **Example**

    the following example gets the current UTC time and date and outputs it back to the contact.

    ```js theme={null}
    const utc = moment.utc();
    api.output(utc);
    ```

    For more information, see the [Moment.js Documentation](https://momentjs.com/docs/).
  </Accordion>

  <Accordion title="Lodash">
    [Lodash Documentation](https://lodash.com/docs/4.17.10)

    Cognigy also supports the utility library Lodash within Code Nodes.

    ## **Usage**

    You can call functions of the Lodash module by using the following syntax:

    ```js theme={null}
    _.keys(<OBJECT>);
    ```

    You can use any Lodash function within a Code Node. See their [documentation](https://lodash.com/docs/4.17.10) for further information.

    ## **Example**

    The following example shows a typical use-case and syntax of the Lodash module.

    ```js theme={null}
    const favouriteFoods = ["pizza", "spaghetti", "burger"];
    const last = _.last(favouriteFoods);
    api.output(last);
    ```

    In this example we're extracting the last element of the `favouriteFoods` array and sending it as an output message.
  </Accordion>

  <Accordion title="Messenger Platform">
    [XML-js Documentation](https://www.npmjs.com/package/xml-js)

    In case you want to parse XML data, you can use this module.

    ## **Usage**

    Within a Code Node, you can call the following method:

    ```js theme={null}
    const xml =
        '<?xml version="1.0" encoding="utf-8"?>' +
        '<note importance="high" logged="true">' +
        '    <title>Happy</title>' +
        '    <todo>Work</todo>' +
        '    <todo>Play</todo>' +
        '</note>';

    const result = xmljs.xml2json(xml, { compact: true, spaces: 4 });

    // result will now contain a JSON representation of the xml above
    ```
  </Accordion>

  <Accordion title="Text Cleaner">
    ## **Usage**

    Within a Code Node, you can invoke the `getTextCleaner` function, which retrieves an instance of the [Text Cleaner](/ai/platform-features/nlu/text-cleaner) class.
    This class provides access to various text-cleaning functions.

    **Parameters**

    | Parameter | Type   | Description                                                               |
    | --------- | ------ | ------------------------------------------------------------------------- |
    | locale    | string | The locale for which to instantiate the class. For example, 'de' or 'en'. |
    | options   | object | See below for the config options.                                         |

    **Config Options**

    ```json theme={null}
    {
        // additional characters which won't be cleaned by cleanDisallowedSymbols
        additionalAllowedCharacters: string[], 

        // additional symbols which are replace (for example, "minus": "-")
        additionalMappedSymbols: { [key: string]: string }, 

        // additional phrases which are replaced
        additionalSpecialPhrases: { [key: string]: string }, 

        // additions to the phonetic alphabet (for example, "cognigy": "c")
        additionalPhoneticAlphabet: { [key: string]: string } 
    }
    ```

    **Returns**

    Instance of the Text Cleaner class.
  </Accordion>
</AccordionGroup>


# Code
Source: https://docs.cognigy.com/ai/for-developers/code/overview

Execute custom JavaScript or TypeScript code within Cognigy Flows using Code Nodes with IntelliSense support.

<a href="/release-notes/2025.24"><Badge>Updated in 2025.24</Badge></a>

<Frame>
  <img alt="Code Node editor with JavaScript and TypeScript support" />
</Frame>

## Description

Code Nodes enable a Flow creator to execute custom JavaScript or TypeScript code within the Flow.
The Code Node editor has full IntelliSense and exposes all Cognigy objects and api functions.
For better performance, the TypeScript code is transpiled when the Node is saved.
If the transpilation fails,
the code is saved and an error icon appears in the top right corner of the Code Node,
indicating that there is a potential issue with the code.

The execution of the Code Node is synchronous, the Flow continues after the Code Node has finished executing.

Just as in other Nodes, `input`, `context`, and `profile` can be accessed and modified from Code Nodes.
If the script crashes, it is stopped and throws an error. In case of an uncaught error, a Flow execution is stopped.

The `api` object provides access to most internal Cognigy functions within the Code Node. For more information, refer to [API Functions](/ai/for-developers/code/api-functions).

The `getTextCleaner` function allows you to get access to an instance of the [Text Cleaner](/ai/for-developers/code/modules).

The `lastConversationEntries` array contains the user inputs and agent outputs for the past 10 turns of the conversation in the following format:

```json theme={null}
[
    {
        source: "user",
        text: "hello"
    },
    {
        source: "bot",
        text: "hello and welcome to Cognigy"
    }
]
```

For convenience, the [`lodash`](https://lodash.com/) and [`moment`](https://momentjs.com/) libraries are exposed for use within Code Nodes.

## Limitations

* The maximum number of characters in the code editor is 200K.
* The maximum code execution time is 1 second. If the limit is exceeded, an error is returned, and the message can be accessed in the `input.codeNodeError.message` Input object. The maximum execution time isn't configurable. To run code that takes longer than one second to execute, we recommend using custom [Extensions](/ai/for-developers/extensions).
* The maximum number of API calls per Code Node is 100. If this limit is exceeded, you can access the returned error message in the Input object via `input.codeNodeError.message`.

## Auto-Complete in the Code Node Editor

The Code Node offers a comprehensive IntelliSense feature.
This feature provides auto-complete suggestions for various objects and actions,
making coding more efficient and reducing the likelihood of errors.

When you use the `input.` operator within a Code Node, you'll see a list of all available properties of the `input` object. These properties may include text, channel, language, and more attributes.

For example:

* `sessionId`
* `userID`
* `entryPoint`
* `flowName`
* `conditionalEntryPointWasExecuted`

Additionally, if you've run a Flow before, the `context` object benefits from full IntelliSense support. When working with the `context` object in your code, you'll receive auto-complete suggestions based on the current context, making accessing and using the relevant information easier.

Furthermore, IntelliSense fully supports the `profile` object.
When working with the `profile` object in your code,
you'll receive auto-complete suggestions that are contextually relevant,
aligning with the structure and attributes
defined in the [Profile Schema](/ai/agents/analyze/contact-profiles#profile-schema).

## Adding Log Statements to Code Nodes

To view log statements on the [Logs page](/ai/agents/test/logs), insert `api.log()` statements into the Code Node.

Input code in a Code Node:

```js theme={null}
const testKeyAPILOG = "Test for api.log"
api.addToContext("test.contextKeyAPI", testKeyAPILOG, "simple")
api.log("debug", testKeyAPILOG);
```

Result on the Logs page:

`2023-01-12 10:27:08 <mark>debug</mark> ***Test for api.log*** { "flowId": "94311a23-b905-4e38-b121-9bffeb658783", "entrypoint": "63bff4588642adbc590be047", "userId": "user1234"`

## Code Node Error Handling

If a Code Node times out or sends too many events, the Flow execution doesn't stop.
Instead, an error message is written to the `input.codeNodeError` Input object.

## Example

### Encode the HTTP Authorization Header in Base64

To use Basic Authentication, add a Code Node above your HTTP Request Node in the Flow editor to encode credentials in Base64.

<Accordion title="Code Node">
  1. Create a Code Node that dynamically encodes credentials:

  ```js theme={null}
  context.encodedCredentials = api.base64Encode(`${context.username}:${context.password}`);
  ```

  2. In the HTTP Request Node, set the Authorization Header to:

  ```js theme={null}
  Basic {{context.encodedCredentials}}
  ```
</Accordion>

### Sending Facebook JSON

You can send Facebook JSON directly from within Code Nodes using the [output](/ai/for-developers/code/api-functions) action.

<Accordion title="JSON Sample">
  ```js theme={null}
  // build the facebook reply
  const obj = {
      "_cognigy": {
          "_facebook": {      
              "message": {
                  "text": "Hello World",
                  "quick_replies": [
                      {
                          "content_type": "text",
                          "condition": "",
                          "title": "Hi",
                          "image_url": "",
                          "payload": "Hi"
                      }
                  ]
              }
          }
      }
  };

  // output the reply
  api.output("test", obj);
  ```
</Accordion>

## More Information

* [API Functions](/ai/for-developers/code/api-functions)
* [Input object](/ai/agents/develop/ai-agent-memory/input)
* [Context object](/ai/agents/develop/ai-agent-memory/context)
* [Profile object](/ai/agents/develop/ai-agent-memory/profile)


# API and CLI
Source: https://docs.cognigy.com/ai/for-developers/developers/api-and-cli

Build custom integrations, manage AI Agents, and automate workflows via the Cognigy REST API or CLI.

<a href="/release-notes/2025.10"><Badge>Updated in 2025.10</Badge></a>

*Cognigy API and CLI* are toolsets for developers to interact with Cognigy. The API allows programmatic interaction and integration with existing systems. The CLI provides direct command-line access to manage AI Agents. Using both tools streamlines workflows, enhances productivity, and manages Cognigy Projects efficiently.

## Limitations

* For Cognigy 4.100 and earlier, you can download [Snapshots](/ai/agents/deploy/snapshots) up to 100 MB using the Cognigy CLI. In later versions, there's no maximum download size for Snapshots via the CLI.
* The `push` and `restore` CLI commands can only update resources. If you want to create a resource, use [the Cognigy API](#cognigy-ai-api).
* CLI commands don't support adding or moving Nodes in the Flow structure. For this purpose, use [the Cognigy API](#cognigy-ai-api).

## Cognigy API

For communication with the Cognigy installation, the frontend uses an HTTP-based API.
This API can also be used by developers to build custom integrations with Cognigy. We also offer a library in the form of an NPM package called the "rest-api-client" which helps to lower the burden of consuming our APIs. This library is available on NPM and is written in modern JavaScript including Typescript definitions.

[Our OpenAPI Documentation](https://api-trial.cognigy.ai/openapi)

### What are the API capabilities?

Any kind of information that is visible and any kind of action that can be performed in the frontend is queried or triggered through this API, which is 100% exposed for developers, too.

### API Base URL

The API Base URL is the endpoint for accessing the Cognigy OpenAPI.
The table outlines the base URLs for different environments.

| Environment              | API Base URL                                                                                                                                                                                                                                                                                                                                                                                                                  |
| ------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Trial (trial.cognigy.ai) | `https://api-trial.cognigy.ai`                                                                                                                                                                                                                                                                                                                                                                                                |
| SaaS                     | Choose the environment you use:<ul><li>**App (app.cognigy.ai)** - `https://api-app.cognigy.ai`</li><li>**App-US** - `https://api-app-us.cognigy.ai/openapi`</li><li>**Other SaaS** - use the `https://api-<company-name>.cognigy.cloud/openapi` pattern where `company-name` is your company name. For example, if your company name is `abc123`, the URL would be:<br />`https://api-abc123.cognigy.cloud/openapi`</li></ul> |
| On-premises              | Specify the API URL as the value for the environment variable `BACKEND_BASE_URL_WITH_PROTOCOL` in your Cognigy `values.yaml` file.                                                                                                                                                                                                                                                                                            |
| NiCE CXone Environments  | Use the `https://cognigy-api-<region>.nicecxone.com` pattern, where `<region>` matches your NiCE CXone cluster. For example, `https://cognigy-api-na1.nicecxone.com`, `https://cognigy-api-eu1.nicecxone.com`. For a full list of available regions, refer to [NiCE CXone Shared SaaS Clusters](/ai/administer/installation/ip-ranges-shared-environments#nice-cxone-shared-saas-clusters).                                   |

### Manage API Keys

An API key is used to authenticate to the Cognigy API as a user in the system without exposing the user's credentials.

To generate an API key, read [Create API Keys](/ai/administer/user-menu/my-profile).

To delete an API key, read [Revoke API Keys](/ai/administer/user-menu/my-profile).

### How to Use the API

Cognigy API includes an OpenAPI documentation that outlines all the different API operations in detail. For example, for the Cognigy trial environment, access `https://api-trial.cognigy.ai/openapi`, and for a NiCE CXone environment, access `https://cognigy-api-<region>.nicecxone.com/openapi`, such as `https://cognigy-api-na1.nicecxone.com/openapi`.

To call the API, send an HTTP request including one of the following:

* A query parameter `api_key` with your API key set as a value.
* A header `X-API-Key` with your API key set as a value.

For example, if your Cognigy installation runs at `https://api-<cognigy-domain>.ai` and your API key is `<api-key>`, you can open a browser tab and go to `https://api-<cognigy-domain>.ai/projects?api_key=<api-key>` to get a list of all the Projects assigned to you, including metadata about the resources within each Project.

This call is used by the frontend when you load the **Main** page.

### Super API-Key

Most of the APIs in Cognigy are agent-bound,
meaning that a developer using our API can only request information from Projects they have access to.
There are certain use-cases where, for example, all Flows within a certain Organization should be listed.
For such use-cases, we have introduced a feature called the "Super API-Key" with our 4.1.4 release.

<AccordionGroup>
  <Accordion title="Feature availability">
    This feature is only available for on-premises customers with separate Cognigy installations.

    This feature is, by default, disabled in all Cognigy installations. Our on-premises customers can activate this feature by setting the following environment variable to "true":

    ```text theme={null}
    FEATURE_USE_SUPERAPIKEY_API
    ```

    ```text theme={null}
    https://api-trial.cognigy.ai/openapi#post-/management/v2.0/organisations/{organisationId}/apikeys
    ```

    This API-key is not bound to a specific user and lives within the scope of an Organization. It can be used to manipulate all assets and objects within the system.
  </Accordion>

  <Accordion title="Super API-Key and TTL">
    This API-key has a default TTL (Time to Live) of 15 minutes for security reasons. Once the key has been created, it will only be valid for 15 minutes.
  </Accordion>
</AccordionGroup>

## Cognigy CLI

The Cognigy CLI is a command-line interface enabling users to manage Projects directly from the command line. For example, with this CLI, you can create Snapshots, manage Locales, and deploy AI Agents efficiently.

To start using the Cognigy CLI, refer to the [Cognigy repository on GitHub](https://github.com/Cognigy/Cognigy-CLI).

You can also explore the Cognigy Help Center, including its [Cognigy CLI](https://support.cognigy.com/hc/en-us/sections/360004745359-Cognigy-CLI) section.

### Examples

The following examples show common CLI commands and the resources they work with.

<AccordionGroup>
  <Accordion title="Clone a Project">
    Clone a Project or its resources of one type from Cognigy to your local machine.

    ```bash theme={null}
    # Clone all Project resources
    cognigy clone
    # Clone only Flows
    cognigy clone --type flows
    # Clone Flows and Endpoints
    cognigy clone --type flows endpoints
    ```

    Supported resource types: `agent` (all Project resources), `flows`, `endpoints`, `lexicons`, `aiAgents`.
  </Accordion>

  <Accordion title="Restore a Project">
    Upload a Project or its resources from your local machine to Cognigy.

    ```bash theme={null}
    # Restore all Project resources
    cognigy restore
    # Restore only Flows
    cognigy restore --type flows
    ```

    Supported resource types: `agent` (all Project resources), `flows`, `endpoints`, `lexicons`, `aiAgents`.
  </Accordion>

  <Accordion title="Pull and Push Resources">
    Download or upload a specific resource by name.

    ```bash theme={null}
    # Download a specific Flow by name
    cognigy pull flows myFlow
    # Upload a specific Flow by name
    cognigy push flows myFlow
    ```

    Supported resource types for:

    * `pull` — `flows`, `lexicons`, `endpoints`, `extensions`, `locales`, `aiAgents`.
    * `push` — `flows`, `lexicons`, `endpoints`, `aiAgents`.
  </Accordion>

  <Accordion title="Compare Local and Remote Resources">
    Compare a local resource against its remote version.

    ```bash theme={null}
    # Compare a Flow at the resource level
    cognigy diff flows myFlow
    # Compare a Flow at the Node level
    cognigy diff flows myFlow --mode node
    ```

    Supported resource types: `flows`, `lexicons`, `endpoints`, `aiAgents`.
  </Accordion>

  <Accordion title="Train a Flow">
    Trigger NLU model training for a Flow.

    ```bash theme={null}
    # Train a Flow with the default timeout
    cognigy train flow myFlow
    # Train a Flow with a custom timeout in milliseconds
    cognigy train flow myFlow --timeout 30000
    ```
  </Accordion>

  <Accordion title="Create a Snapshot or Locale">
    Creates a Snapshot or Locale in the remote Project and imports the new resource to your local Project.

    ```bash theme={null}
    # Create a Snapshot with an optional description
    cognigy create snapshot mySnapshot "Optional description"
    # Create a Locale
    cognigy create locale myLocale
    ```
  </Accordion>

  <Accordion title="Export and Import Multilingual Flow Texts">
    Export a `content.csv` file with the text fields from Nodes in the Flow across all Locales, or import updated localized content back into a Flow.

    ```bash theme={null}
    # Export Flow texts to content.csv
    cognigy exportcsv flow myFlow
    # Import updated localized content back into the Flow
    cognigy importcsv flow myFlow
    ```
  </Accordion>

  <Accordion title="Localize a Flow">
    Add or remove localization for multiple Intents and Nodes at once.

    ```bash theme={null}
    # Add localization for a Flow from a source locale
    cognigy localize flow myFlow --localeName de-DE --sourceLocale en-US
    # Remove localization for a Flow
    cognigy localize flow myFlow --localeName de-DE --reverse
    ```
  </Accordion>

  <Accordion title="Translate a Flow">
    Translate Flow content using a [machine translation provider](/ai/agents/deploy/endpoints/real-time-translation-settings).

    ```bash theme={null}
    # Translate a Flow using DeepL
    cognigy translate flow myFlow --localeName de-DE --fromLanguage en --toLanguage de --translator deepl --apiKey <key>
    ```

    Supported translation providers: `google`, `microsoft`, `deepl`.
  </Accordion>

  <Accordion title="Execute a Raw API Command">
    Pass a JSON payload directly to the Cognigy REST API client.

    ```bash theme={null}
    # Run an API command with a JSON payload
    cognigy execute <command> -d '{"key": "value"}'
    # List all available API commands
    cognigy execute --list
    ```

    Use `--list` to print all available API commands.
  </Accordion>

  <Accordion title="Run Playbooks">
    Schedule Playbook runs on Cognigy and poll for results. The output is written to `./playbookRunResults.json`.

    ```bash theme={null}
    # Run all Playbooks in the project
    cognigy run playbooks
    # Run Playbooks from a specific file
    cognigy run playbooks ./myPlaybooks.json
    ```

    Exit codes: `0` — success, `1` — failure, `2` — timeout.
  </Accordion>

  <Accordion title="Manage AI Agent Knowledge">
    Create a Knowledge Store, extract content from source documents, and ingest `.ctxt` files into the store.

    ```bash theme={null}
    # Create a Knowledge Store
    cognigy knowledge-ai create source myStore "My store" --projectId <id> --language en-US
    # Extract content from a PDF into a .ctxt file
    cognigy knowledge-ai extract pdf -i ./docs/manual.pdf -o ./output/manual.ctxt
    # Ingest .ctxt files into the Knowledge Store
    cognigy knowledge-ai ingest --knowledgeStoreId <id> --input ./output/
    ```

    Supported extraction types: `text`, `pdf`, `docx`, `csv`, `json`, `jsonl`, `epub`, `srt`, `md`, `cheerio`, `playwright`, `diffbot`, `other`.
  </Accordion>
</AccordionGroup>


# Client Libraries
Source: https://docs.cognigy.com/ai/for-developers/developers/client-libraries



Cognigy offers a variety of clients in various languages to connect with Cognigy. All of these are hosted on [the Cognigy GitHub account](https://github.com/Cognigy).

All clients except the REST API Client connect against Cognigy Endpoints. The REST API Client connects directly against the Cognigy API and can be used to interface with Cognigy Resources directly from TypeScript/JavaScript.

| Client Name                                                               | Language   | Description                                                                                                            |
| ------------------------------------------------------------------------- | ---------- | ---------------------------------------------------------------------------------------------------------------------- |
| [WebchatWidget](https://github.com/Cognigy/WebchatWidget)                 | TypeScript | A graphical chat widget you can embed into your website to provide visitors access to a Conversational AI via Cognigy. |
| [SocketClient](https://github.com/Cognigy/SocketClient)                   | TypeScript | A JavaScript/TypeScript client that can be used to connect to Cognigy Socket Endpoints                                 |
| [REST API Client](https://www.npmjs.com/package/@cognigy/rest-api-client) | TypeScript | An npm module to connect against the Cognigy API                                                                       |
| [Cognigy CLI](https://github.com/Cognigy/Cognigy-CLI)                     | TypeScript | Cognigy-CLI is a series of tools to interact with your Cognigy AI Agent projects from the command line                 |


# NiCE Cognigy MCP Server
Source: https://docs.cognigy.com/ai/for-developers/developers/cognigy-mcp-server

NiCE Cognigy MCP Server lets you build, test, and optimize LLM-based AI Agents via the Cognigy REST API.

<Warning>
  The NiCE Cognigy MCP Server is deprecated and has been replaced by the [NiCE Cognigy plugin](/ai/for-developers/developers/cognigy-plugin). For new setups, install the plugin instead. Existing MCP Server installations continue to work. New features and improvements are delivered through the plugin.
</Warning>

The *NiCE Cognigy MCP Server* is an [npm package](https://www.npmjs.com/package/@cognigy/mcp-server) that provides a [Model Context Protocol (MCP)](https://modelcontextprotocol.io/) server for connecting MCP clients like Claude, Cursor, or VS Code to the Cognigy REST API. With the NiCE Cognigy MCP server, you can create, test, and iteratively improve LLM-based AI Agents without leaving your development environment.

<div>
  <video title="NiCE Cognigy MCP Server tools executed via Claude Code CLI in terminal" aria-label="NiCE Cognigy MCP Server tools executed via Claude Code CLI in terminal">
    <source />
  </video>
</div>

## Key Benefits

* **Comprehensive API coverage.** Use workflow tools in the NiCE Cognigy MCP Server to cover the Cognigy API for end-to-end automation.
* **Fast AI Agent setup and iteration.** Build an AI Agent persona, Flow, Node, and REST Endpoint in one workflow, then test, refine, and repeat until it performs as expected.
* **Knowledge and guidance built in.** Access Knowledge Stores for RAG use cases to give your AI Agent the information it needs.
* **Reliable and safe operations.** Configure rate limiting, Zod input validation, and standardized RFC 7807 error responses for safe operations with the NiCE Cognigy MCP Server.

## Prerequisites

* A [Cognigy API key](/ai/administer/user-menu/my-profile#api-keys). To create one, go to **User Menu > My Profile > API Keys > Create New**.
* A [Cognigy API base URL](/ai/for-developers/developers/api-and-cli#api-base-url). For example, for the trial environment, use `https://api-trial.cognigy.ai`.
* An MCP-compatible client such as [Claude Desktop](https://claude.ai/download), Claude Code, Codex, [Cursor](https://www.cursor.com/), or VS Code with GitHub Copilot.
* Node.js 20 or later is required for command-line and manual setup.

## How to Set Up

<Tabs>
  <Tab title="Claude Desktop (.mcpb)">
    This is the easiest setup path and doesn't require Node.js.

    1. Download the `.mcpb` file from the [npm package page](https://www.npmjs.com/package/@cognigy/mcp-server).
    2. Double-click the file to open the installation dialog in Claude Desktop.
    3. Enter your Cognigy API base URL and API key.
  </Tab>

  <Tab title="CLI">
    Run the setup command for your MCP client in the CLI. The command prompts you to enter your Cognigy API base URL and API key, then automatically updates the MCP client configuration.

    | MCP Client        | Command                                             |
    | ----------------- | --------------------------------------------------- |
    | Claude Desktop    | `npx @cognigy/mcp-server init --client claude`      |
    | Claude Code       | `npx @cognigy/mcp-server init --client claude-code` |
    | Codex             | `npx @cognigy/mcp-server init --client codex`       |
    | Cursor            | `npx @cognigy/mcp-server init --client cursor`      |
    | VS Code (Copilot) | `npx @cognigy/mcp-server init --client vscode`      |

    Restart your MCP client after setup completes.
  </Tab>

  <Tab title="Manual">
    Add the server definition directly to your MCP client configuration file.

    <AccordionGroup>
      <Accordion title="Claude Desktop, Claude Code, and Cursor">
        Add the following to your MCP client configuration file:

        ```json theme={null}
        {
          "mcpServers": {
            "cognigy": {
              "command": "npx",
              "args": ["-y", "@cognigy/mcp-server"],
              "env": {
                "COGNIGY_API_BASE_URL": "your-api-base-url",
                "COGNIGY_API_KEY": "your-api-key-here"
              }
            }
          }
        }
        ```
      </Accordion>

      <Accordion title="Codex">
        Add the following to your Codex configuration file:

        ```toml theme={null}
        [mcp_servers.cognigy]
        command = "npx"
        args = ["-y", "@cognigy/mcp-server"]

        [mcp_servers.cognigy.env]
        COGNIGY_API_BASE_URL = "your-api-base-url"
        COGNIGY_API_KEY = "your-api-key"
        ```
      </Accordion>

      <Accordion title="VS Code (Copilot)">
        Add the following to your VS Code MCP configuration:

        ```json theme={null}
        {
          "servers": {
            "cognigy": {
              "command": "npx",
              "args": ["-y", "@cognigy/mcp-server"],
              "env": {
                "COGNIGY_API_BASE_URL": "your-api-base-url",
                "COGNIGY_API_KEY": "your-api-key"
              }
            }
          }
        }
        ```
      </Accordion>
    </AccordionGroup>

    Restart your MCP client to ensure it is connected to the NiCE Cognigy MCP server.
  </Tab>
</Tabs>

### Security

* API keys are passed through environment variables and are never logged.
* All inputs are validated with Zod schemas before reaching the API.
* Rate limiting protects the API from excessive or abusive requests.

#### Privacy

The NiCE Cognigy MCP Server sends requests only to the Cognigy API base URL. The server doesn't collect, store, or share any data. All data stays between your MCP client and your Cognigy instance. For more information, see the [Cognigy Privacy Policy](https://www.cognigy.com/privacy-policy).

## How to Use

Once the server is connected to your MCP client, you can interact with Cognigy using natural language prompts.

The following examples show prompts for common workflows and the tools the MCP client runs in response. For a full list of available tools, refer to the [Tools](https://www.npmjs.com/package/@cognigy/mcp-server) section of the NiCE Cognigy MCP Server npm package documentation.

<AccordionGroup>
  <Accordion title="Create an AI Agent">
    Ask your MCP client to create an AI Agent in a Cognigy Project:

    ```txt theme={null}
    Create a Cognigy AI Agent called "Support Bot" in Project <projectId>.
    Give it a helpful customer support persona, set up GPT-4 as the LLM,
    and return the endpoint URL so I can test it.
    ```

    This workflow uses the `setup_llm` tool to create the LLM resource and the `create_ai_agent` tool to provision the AI Agent, Flow, and REST Endpoint.
  </Accordion>

  <Accordion title="Test the Agent">
    Send a message to the AI Agent through its Endpoint:

    ```txt theme={null}
    Talk to my Support Bot at <endpointUrl> and ask "How do I reset my password?"
    ```

    This workflow uses the `talk_to_agent` tool and returns the response in your MCP client.
  </Accordion>

  <Accordion title="Improve the Agent">
    Refine the AI Agent instructions and compare responses:

    ```txt theme={null}
    Update the job description to make the response more concise and actionable.
    Talk to it again and compare the responses.
    ```

    This workflow uses a self-improvement loop: `talk_to_agent` — evaluate — `update_ai_agent` — `talk_to_agent`.
  </Accordion>

  <Accordion title="Add a Knowledge Store">
    Attach a Knowledge Store so the AI Agent can search approved content when answering questions:

    ```txt theme={null}
    Create a Knowledge Store in Project <projectId>, add the URL
    https://docs.example.com/faq as a source, then attach it to my Support Bot
    so the AI Agent can reference it to answer questions.
    ```

    This workflow uses the `manage_knowledge` tool to create the store and ingest the source, then `create_tool` to attach it to the AI Agent Node.
  </Accordion>

  <Accordion title="Review Existing Resources">
    List resources to inspect Projects, AI Agents, or conversations:

    ```txt theme={null}
    List all my Cognigy Projects and show which AI Agents exist in each one.
    ```

    This workflow uses the `list_resources` tool with resource types `project` and `agent`.

    You can also ask for recent conversations and summarize the returned data.
  </Accordion>

  <Accordion title="Export and Import Packages">
    Export a Package from the Project, list exportable Packages, and import the Package into another Project:

    ```txt theme={null}
    Upload the package at `/absolute/path/to/support-bot.zip` into Project <projectId>,
    show me the import preview, then import it using the default selections.
    ```

    <Tip>
      You can also drag and drop files into some MCP clients to upload them.
    </Tip>

    This workflow uses the `manage_packages` tool with operations: `upload_and_inspect`, `list_exportable` then `import`.
  </Accordion>
</AccordionGroup>

## More Information

* [NiCE Cognigy MCP Server on npm](https://www.npmjs.com/package/@cognigy/mcp-server)
* [Cognigy API and CLI](/ai/for-developers/developers/api-and-cli)
* [MCP Server Endpoint](/ai/agents/deploy/endpoint-reference/mcp-server)
* [AI Agents Overview](/ai/agents/overview)
* [Knowledge AI Overview](/ai/agents/develop/knowledge-ai/overview)


# NiCE Cognigy Plugin
Source: https://docs.cognigy.com/ai/for-developers/developers/cognigy-plugin

NiCE Cognigy Plugin connects Claude Code, Claude Desktop, ChatGPT, Codex, Gemini CLI, and Antigravity to the Cognigy REST API to build, test, and improve LLM-based AI Agents.

The *NiCE Cognigy Plugin* connects your client to the [Cognigy](https://www.cognigy.com) REST API. You can create, test, and improve LLM-based AI Agents without leaving your development environment.
The plugin includes a version-matched server engine, skills, agents, and scripts. The server engine runs as a [Model Context Protocol (MCP)](https://modelcontextprotocol.io/) server, exposing Cognigy capabilities to your client.

## Key Benefits

* **Comprehensive API coverage.** 16 tools cover the Cognigy API. They automate everything from creating an AI Agent to deploying it on voice and web channels.
* **Fast setup and iteration.** Create an AI Agent, Flow, AI Agent Node, and REST Endpoint in a single call. Then talk to the AI Agent, refine it, and repeat with a built-in self-improvement loop.
* **Guidance built in.** Skills load automatically when your request matches. Dedicated agents run the full build-and-test and voice go-live loops for you.
* **Reliable and safe operations.** Rate limiting, Zod input validation, and standardized RFC 7807 error responses keep operations safe.

## Prerequisites

* A [Cognigy API key](/ai/administer/user-menu/my-profile#api-keys). To create one, go to **User Menu > My Profile > API Keys > Create New**.
* A [Cognigy API base URL](/ai/for-developers/developers/api-and-cli#api-base-url). For example, for the trial environment, use `https://api-trial.cognigy.ai`.
* A [supported client](#supported-clients).
* Node.js 20 or later.

## Restrictions

* Codex subagents aren't supported because they use a different configuration file format than Cognigy's agents use. Codex users can still use NiCE Cognigy's tools and skills.

## Supported Clients

<Warning>
  Google [transitioned Gemini CLI to Antigravity CLI](https://developers.googleblog.com/an-important-update-transitioning-gemini-cli-to-antigravity-cli/). On 18 June 2026, Gemini CLI stopped serving the free tier, Google AI Pro, and Ultra. Gemini CLI still works for Code Assist Standard, Code Assist Enterprise, and Code Assist for GitHub subscriptions managed through Google Cloud. On a consumer tier, use Antigravity instead.
</Warning>

The following table shows what each client receives, if you need to finish the installation manually, and if it auto-updates:

| Client                                          | Tools | Skills | Agents | Manual step after the installer | Auto-updates                                 |
| ----------------------------------------------- | ----- | ------ | ------ | ------------------------------- | -------------------------------------------- |
| Claude Code CLI and Claude Desktop **Code** tab | Yes   | Yes    | Yes    | None                            | Enable once                                  |
| Claude Desktop chat                             | Yes   | Yes    | Yes    | Install the plugin in the app   | Automatic                                    |
| ChatGPT, Codex CLI, and Codex IDE extension     | Yes   | Yes    | No     | None                            | Automatic                                    |
| Google Gemini CLI                               | Yes   | Yes    | Yes    | None                            | Automatic                                    |
| Antigravity IDE and `agy` CLI                   | Yes   | Yes    | Yes    | None                            | Automatic                                    |
| Other hosts, such as VS Code and Cursor         | Yes   | Yes    | Yes    | Install the plugin in the host  | Manual <sup><a href="#footnote1">1</a></sup> |

## How to Set Up

<Steps>
  <Step title="Run installer">
    1. Run the installer in your terminal.

    <AccordionGroup>
      <Accordion title="Bootstrap">
        <Tabs>
          <Tab title="macOS and Linux">
            ```bash theme={null}
            bash <(curl -fsSL https://raw.githubusercontent.com/Cognigy/cognigy-plugin/main/install.sh)
            ```
          </Tab>

          <Tab title="Windows">
            Run in PowerShell as Administrator:

            ```powershell theme={null}
            irm https://raw.githubusercontent.com/Cognigy/cognigy-plugin/main/install.ps1 | iex
            ```
          </Tab>
        </Tabs>
      </Accordion>

      <Accordion title="With Node.js">
        <Tabs>
          <Tab title="Interactive">
            Skip the bootstrap and run the installer directly:

            ```bash theme={null}
            npx -y -p @cognigy/plugin-engine@latest cognigy-setup
            ```
          </Tab>

          <Tab title="Scripting and CI">
            Skip the prompts with flags:

            ```bash theme={null}
            npx -y -p @cognigy/plugin-engine@latest cognigy-setup \
              --client claude-code --client claude-desktop --client codex \
              --client gemini --client antigravity --client other-hosts \
              --api-base-url https://api-trial.cognigy.ai --api-key <key>
            ```

            The `--client` flag is repeatable. Valid values are `claude-code`, `claude-desktop`, `codex`, `gemini`, `antigravity`, and `other-hosts`. Omit it in interactive mode to pick from the menu.

            This single command already selects and configures every client you list. You can skip step 3.
          </Tab>
        </Tabs>
      </Accordion>
    </AccordionGroup>

    2. Enter your Cognigy API base URL and API key.
    3. Select your client.
  </Step>

  <Step title="Install Plugin">
    <Tabs>
      <Tab title="Claude Code">
        * If you use Claude Code, restart it or run `/reload-plugins`. You now have the tools, skills, and agents. If the `claude` CLI isn't on your PATH, the installer writes the credentials file instead and prints the commands to finish by hand:

          ```txt theme={null}
          /plugin marketplace add Cognigy/cognigy-plugin
          /plugin install cognigy@cognigy-plugin
          ```

        * If you use the Claude Desktop **Code** tab, do the following:

          1. In Claude Desktop, go to **Customize > Plugins** and select **Add > Add marketplace** in the upper-right corner.
          2. In the **URL** field, enter `Cognigy/cognigy-plugin`, then click **Sync**. After syncing, you are redirected to your personal marketplaces.
          3. In the **Cognigy** card, click **+** to install the plugin.
          4. In the local MCP warning dialog, click **Continue**.

        <Note>
          The first Claude Code session downloads the engine. If the tools don't appear, run `/mcp` and reconnect, or restart the client.

          On Windows, fully quit Claude Code. End every Claude process in Task Manager because a normal close can leave it running in the background. Then reopen the client.
        </Note>

        <Tip>
          Claude Code leaves auto-updates off for third-party marketplaces by default. Turn them on so future fixes reach you automatically. For more information, see [Get Plugin Updates](#get-plugin-updates).
        </Tip>
      </Tab>

      <Tab title="Claude Desktop chat">
        1. Restart Claude Desktop. The tools work now.
        2. In Claude Desktop, go to **Customize > Plugins** and select **Add > Add marketplace** in the upper-right corner.
        3. In the **URL** field, enter `Cognigy/cognigy-plugin`, then click **Sync**. After syncing, you are redirected to your personal marketplaces.
        4. In the **Cognigy** card, click **+** to install the plugin.
        5. In the local MCP warning dialog, click **Continue**.

        Leave the plugin's own `platform` connector unconnected. The `Cognigy` connector wired by the installer already serves the tools.

        <Note>
          Claude Desktop chat stores plugin configuration in your claude.ai account rather than a local file. It offers no field for an API key, so the plugin's own connector never receives credentials. To work around this, the installer wires a separate connector for the tools. The in-app install then adds the skills and agents.
        </Note>
      </Tab>

      <Tab title="ChatGPT and Codex">
        <Warning>
          The installer isn't optional. Only the installer supplies your Cognigy API key because the ChatGPT app has no dedicated field.

          Don't add a `[mcp_servers.cognigy]` entry to `~/.codex/config.toml`. The plugin declares its own server, so a manual entry starts a second copy of the engine and duplicates every tool in the picker. If an earlier setup left one behind, remove it with `codex mcp remove cognigy`.
        </Warning>

        The following steps cover the ChatGPT desktop app, the Codex CLI, and the IDE extension:

        1. Start a new thread. An open thread doesn't see the plugin because Codex loads plugins when a session starts.
        2. Install the plugin:

           * If you use the `codex` CLI, run `/plugins`.
           * If the `codex` CLI isn't on your PATH, the installer writes your credentials and stops. Follow these steps in the ChatGPT app:

             1. Select **Plugins** in the sidebar.
             2. Select **Add** in the top-right corner, then **Add a Marketplace**.
             3. Enter `Cognigy/cognigy-plugin` as the source, then select **Add Marketplace**.
             4. Select **Install** on the **Cognigy** plugin.
             5. Start a new thread.
      </Tab>

      <Tab title="Google Gemini CLI">
        1. Restart `gemini`. To verify, run `gemini extensions list` and look for `cognigy`.

        Gemini passes only the environment variables an extension declares to its MCP servers, so your shell environment never reaches them. The installer works around this by writing the credentials file that the engine reads directly. With a manual install, Gemini prompts for the API base URL and key instead, and stores the key in the OS keychain. To run those prompts again, use `gemini extensions config cognigy`.

        <Note>
          Gemini CLI requires plugin release 1.9.0 or later. It's also limited to Code Assist subscriptions. See the note in [Supported Clients](#supported-clients).
        </Note>
      </Tab>

      <Tab title="Antigravity">
        1. Restart Antigravity. The tools, skills, and agents are live in both the IDE and the `agy` CLI because they share one plugin directory.

        The installer doesn't need the `agy` binary, and works before Antigravity's first launch. Your global `mcp_config.json` file is never touched. The plugin carries its own `mcp_config.json` file, so installing or removing the plugin can't disturb your other MCP servers.

        Antigravity updates automatically and has no plugin update command. Every launch pulls the latest engine, and the first launch after a new release also refreshes the skills and agents. Those load on the launch after that, because Antigravity reads them at startup.

        <Note>
          Plugin releases before 1.10.1 can't refresh themselves. If you installed a version before 1.10.1, run `cognigy-setup update` once. Updates are hands-off from then on.
        </Note>

        <Note>
          If you migrated from Gemini CLI and ran `agy plugin import gemini`, check `agy plugin list`. An imported copy of the Cognigy extension may sit alongside this plugin and duplicate every tool. Remove the tools you didn't install here.
        </Note>
      </Tab>

      <Tab title="Other hosts">
        Any host that can install a Claude-format plugin or accept a hand-written MCP server entry can run the plugin. The installer's job here is only to supply the credentials that the host can't ask you for.

        When you select **Other hosts (VS Code, Cursor, …) — writes a local creds file**, the installer writes the credentials and stops. Nothing is added to any client configuration, so it's safe to combine with any other target. To install the plugin in your host, follow these steps:

        * For VS Code and Copilot:

          1. Add the following to `settings.json`:

             ```json theme={null}
             {
               "chat.plugins.enabled": true,
               "chat.plugins.marketplaces": ["Cognigy/cognigy-plugin"]
             }
             ```
          2. Open the Extensions view, search for `@agentPlugins`, install **Cognigy**, and restart VS Code.

        * For any other host, add the engine as an MCP server in the host's own configuration:

        ```json theme={null}
        {
          "mcpServers": {
            "cognigy": {
              "command": "npx",
              "args": ["-y", "-p", "cognigy-engine@npm:@cognigy/plugin-engine@latest", "cognigy-mcp"]
            }
          }
        }
        ```

        The `cognigy-engine@npm:` prefix is a required npm alias. Without this prefix, `npx` treats a local package of the same name as satisfying the request, skips the download, and the server fails to start.

        <Warning>
          Leave the server's `env` block empty, and ignore any credential fields the host shows for this plugin. Only Claude Code substitutes values into a plugin manifest. Every other host passes the placeholder through unchanged, so filling those fields changes nothing. The credentials file covers it instead.
        </Warning>
      </Tab>
    </Tabs>
  </Step>
</Steps>

### Configuration

The installer from [How to Set Up](#how-to-set-up) collects your Cognigy API base URL and API key. It saves them where each client you selected can reach them. Claude Code uses the OS keychain. Most other clients use `~/.cognigy-plugin/config.json`, readable only by you, because they can't prompt a plugin for secrets. The engine reads environment variables first and falls back to that file, so the following variables let you override the defaults.

| Variable                  | Required | Default | Description                              |
| ------------------------- | -------- | ------- | ---------------------------------------- |
| `COGNIGY_API_BASE_URL`    | Yes      | None    | Your Cognigy API base URL.               |
| `COGNIGY_API_KEY`         | Yes      | None    | Your Cognigy API key.                    |
| `LOG_LEVEL`               | No       | `info`  | One of `debug`, `info`, `warn`, `error`. |
| `RATE_LIMIT_MAX_REQUESTS` | No       | `100`   | Maximum requests per window.             |
| `RATE_LIMIT_WINDOW_MS`    | No       | `60000` | Rate limit window in milliseconds.       |

### Available Tools

The plugin provides the following tools to automate Cognigy workflows. Each tool has a type of `Read` or `Write`, indicating whether it reads data from Cognigy or writes data to it.

| Tool                   | Type  | Description                                                                                                |
| ---------------------- | ----- | ---------------------------------------------------------------------------------------------------------- |
| `create_ai_agent`      | Write | Create a complete AI Agent with an auto-provisioned Flow, AI Agent Node, and REST Endpoint.                |
| `update_ai_agent`      | Write | Update the persona, guardrails, and job configuration (role, procedures, LLM, temperature).                |
| `setup_llm`            | Write | Create an LLM resource (GPT-4, Claude, Mistral, and others) with automatic connection validation.          |
| `talk_to_agent`        | Write | Send a message to an AI Agent and get its response.                                                        |
| `list_resources`       | Read  | List Projects, AI Agents, Flows, Endpoints, LLMs, Knowledge Stores, and more.                              |
| `get_resource`         | Read  | Get detailed information about a single resource.                                                          |
| `delete_resource`      | Write | Permanently delete a resource.                                                                             |
| `manage_knowledge`     | Write | Create Knowledge Stores, add sources (URL, text, file), and list chunks for RAG.                           |
| `create_tool`          | Write | Add a tool (HTTP, knowledge, email, MCP) to an AI Agent Node.                                              |
| `update_tool`          | Write | Update an existing tool Node's configuration.                                                              |
| `manage_webchat`       | Write | Create or configure a Webchat v3 Endpoint for website deployment.                                          |
| `manage_flow_nodes`    | Write | Create, update, delete, or list Flow Nodes for conversation logic.                                         |
| `manage_packages`      | Write | List exportable resources, and upload, inspect, import, export, and download Cognigy Package zip files.    |
| `manage_voice_gateway` | Write | Create or configure a Voice Gateway Endpoint with WebRTC for browser-based voice interaction.              |
| `manage_settings`      | Write | Manage Project-level settings, including voice preview and Knowledge AI configuration.                     |
| `audit_voice_agent`    | Write | Audit a voice agent against the Go-Live Checklist. It reports by default and applies safe fixes on demand. |

### Security

* API keys are passed through environment variables and are never logged.
* All inputs are validated with Zod schemas before reaching the API.
* Rate limiting protects the API from excessive or abusive requests.

#### Privacy

The NiCE Cognigy plugin sends requests only to the Cognigy API base URL that you configure. The plugin doesn't collect, store, or share any data. All data stays between your client and your Cognigy instance. For more information, see the [Cognigy Privacy Policy](https://www.cognigy.com/privacy-policy).

## How to Use

Once the plugin is installed, you can interact with Cognigy using natural language prompts in your client. The following examples show prompts for common workflows and the tools the client runs in response.

<AccordionGroup>
  <Accordion title="Create an AI Agent">
    Ask your client to create an AI Agent in a Cognigy Project:

    ```txt theme={null}
    Create a Cognigy AI Agent called "Support Bot" in Project <projectId>.
    Give it a helpful customer support persona, set up an LLM,
    and return the endpoint URL so I can test it.
    ```

    This workflow uses the `setup_llm` tool to create the LLM resource. It then uses the `create_ai_agent` tool to provision the AI Agent, Flow, and REST Endpoint.
  </Accordion>

  <Accordion title="Test and Improve the Agent">
    Send a message to the AI Agent, refine its instructions, and compare responses:

    ```txt theme={null}
    Talk to my Support Bot at <endpointUrl> and ask "How do I reset my password?".
    Then update the job description to make the response more concise and actionable.
    Talk to it again and compare the responses.
    ```

    This workflow uses a self-improvement loop: `talk_to_agent`, then evaluate, then `update_ai_agent`, then `talk_to_agent` again.
  </Accordion>

  <Accordion title="Add a Knowledge Store">
    Attach a Knowledge Store so the AI Agent can search approved content when answering questions:

    ```txt theme={null}
    Create a Knowledge Store in Project <projectId>, add the URL
    https://docs.example.com/faq as a source, then attach it to my Support Bot
    so the AI Agent can reference it to answer questions.
    ```

    This workflow uses the `manage_knowledge` tool to create the store and ingest the source. It then uses `create_tool` to attach the store as a knowledge search tool on the AI Agent Node.
  </Accordion>

  <Accordion title="Deploy to Voice">
    Configure voice preview and create a browser-based Voice Gateway Endpoint:

    ```txt theme={null}
    Set the voice preview provider for Project <projectId> to Microsoft, then create
    a Voice Gateway endpoint for flow <flowId> named "Support Voice" and give me the
    WebRTC demo URL I can open in the browser.
    ```

    This workflow uses `manage_settings` with the `set_voice_preview` operation to configure speech. It then uses `manage_voice_gateway` to provision a Voice Gateway Endpoint with a WebRTC demo URL.
  </Accordion>

  <Accordion title="Export and Import Packages">
    Upload a Package, preview the import, and import it into a Project:

    ```txt theme={null}
    Upload the package at /absolute/path/to/support-bot.zip into Project <projectId>,
    show me the import preview, then import it using the default selections.
    ```

    <Tip>
      You can also drag and drop files into some clients to upload them.
    </Tip>

    This workflow uses the `manage_packages` tool with the operations `upload_and_inspect` and then `import`.
  </Accordion>

  <Accordion title="Review Existing Resources">
    List resources to inspect Projects, AI Agents, or conversations:

    ```txt theme={null}
    List all my Cognigy Projects and show which AI Agents exist in each one.
    ```

    This workflow uses the `list_resources` tool with the resource types `project` and `agent`. You can also ask for recent conversations and summarize the returned data.
  </Accordion>
</AccordionGroup>

## Get Plugin Updates

Most clients [update themselves](#supported-clients). Claude Code leaves auto-updates off for third-party marketplaces. Turn them on once in `/plugin > Marketplaces > cognigy-plugin`. Otherwise, update manually with `/plugin update cognigy@cognigy-plugin`.

The installer also works as a manager. Run the same `cognigy-setup` command with a subcommand:

```bash theme={null}
npx -y -p @cognigy/plugin-engine@latest cognigy-setup status      # Show what's installed and the latest available version
npx -y -p @cognigy/plugin-engine@latest cognigy-setup update      # Pull the latest release for every installed client
npx -y -p @cognigy/plugin-engine@latest cognigy-setup uninstall   # Remove the plugin and connector (add --purge to also clear ~/.cognigy-plugin)
```

Add `--client <name>` to any subcommand to target a single client.

## More Information

* [NiCE Cognigy Plugin on GitHub](https://github.com/Cognigy/cognigy-plugin)
* [Cognigy API and CLI](/ai/for-developers/developers/api-and-cli)
* [MCP Server Endpoint](/ai/agents/deploy/endpoint-reference/mcp-server)
* [AI Agents Overview](/ai/agents/overview)
* [Knowledge AI Overview](/ai/agents/develop/knowledge-ai/overview)

<span>
  <sup>1</sup> The host owns the plugin version.
</span>


# Extension Framework
Source: https://docs.cognigy.com/ai/for-developers/developers/extension-framework

Build custom Cognigy Extensions with TypeScript, including Extension Nodes for integrations and Knowledge Connectors for external knowledge databases.

The *Extension Framework* provides developers with the ability to create custom Extensions with Extension Nodes and Knowledge Connectors that go beyond the default and marketplace Nodes.

## Custom Extensions

Custom Extensions can offer:

* Extension Nodes that perform new actions, integrate with external systems, process data, and improve conversational scenarios for users.
* Knowledge Connectors that offer new integration with external knowledge databases and improve RAG capabilities.

## Development Requirements

Developing custom Extensions requires:

* Knowledge of TypeScript/JavaScript.
* Installation of Node.js and NPM packages.

## Getting Started

To start implementing a custom Extension, refer to the [Extension's npm package Readme](https://www.npmjs.com/package/@cognigy/extension-tools/v/0.16.5-rc2).


# Extensions
Source: https://docs.cognigy.com/ai/for-developers/extensions

Extend Cognigy Flows with custom Nodes, third-party API integrations, and Knowledge Connectors via Extensions.

<a href="/release-notes/2026.9"><Badge>Updated in 2026.9</Badge></a>

*Extensions* in Cognigy are packages that enhance your Flows with additional capabilities. An Extension can include *Extension Nodes*, *Knowledge Connectors*, or both.

* [Extension Nodes](#extension-nodes) let you:
  * Integrate with third-party APIs.
  * Run custom logic for lightweight operations.
  * Build convenience nodes using npm modules.
* [Knowledge Connectors](#knowledge) let you:
  * Import documents or data as [Knowledge Sources](/ai/agents/develop/knowledge-ai/knowledge-source/overview).
  * Improve AI Agents' RAG capabilities.

You can install Extensions from the [Marketplace](https://www.cognigy.com/marketplace#/), develop your [custom Extensions](#custom-extensions), or modify existing ones.

## Limitations

* Extensions have a default 20-second timeout. You can change the timeout for on-premises installations. Change the environment variable as described in the [Cognigy Helm Chart](https://github.com/Cognigy/cognigy-ai-helm-chart/blob/main/values.yaml).
* Extensions can make up to 10 API calls per execution, except for `api.log()`. For more details, see the [Best Practices for Developing Custom Extensions](#best-practices-for-developing-custom-extensions).
* By default, objects passed as arguments to Extension API calls have a size limit of 62 KB.
* The [Extension runtime environment](#cognigy-serverless) is cleaned up after 15 minutes of inactivity.

## Working with Extensions

<Tabs>
  <Tab title="GUI">
    In **Manage > Extensions**, you can install, update, and uninstall Extensions from the Marketplace or [custom Extensions](#custom-extensions). Also, you can [mark your custom Extensions as trusted](#make-extensions-trusted).

    The Extension Marketplace is also available on [Cognigy's website](https://www.cognigy.com/marketplace#/).
  </Tab>

  <Tab title="CLI">
    You can pull Extensions using the [Cognigy CLI](https://github.com/Cognigy/cognigy-cli).
  </Tab>

  <Tab title="API">
    You can view, delete, upload, and update Extensions using the [Cognigy API](https://api-trial.cognigy.ai/openapi#tag--Extensions-v2.0). Also, you can [mark your custom Extensions as trusted](#make-extensions-trusted).
  </Tab>
</Tabs>

### Extension Nodes

You can add Extension Nodes to your Flows in the [Flow editor](/ai/agents/develop/projects-and-flows/editor).

### Knowledge Connectors

[Knowledge Connectors](/ai/agents/develop/knowledge-ai/knowledge-connectors) allow you to import Knowledge Sources from external knowledge databases to a [Knowledge Store](/ai/agents/develop/knowledge-ai/knowledge-store) and schedule regular knowledge import in **Build > Knowledge**.

### Custom Extensions

Cognigy allows anyone to extend the capabilities of Cognigy by developing their own Extensions. Custom Extensions are JavaScript or TypeScript modules that you can use to create fully customized Extension Nodes and Knowledge Connectors. After developing a custom Extension, you can install it by uploading it in **Manage > Extensions**.

### Best Practices for Developing Custom Extensions

Extensions are designed for lightweight operations. Follow these best practices:

<Accordion title="Best Practices">
  | Practice                          | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
  | --------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
  | Keep Extensions lean              | Strip development dependencies before packaging. Run either `npm ci --omit=dev` or `npm ci --production`. Development dependencies, such as test frameworks, type definitions, linters, and bundlers, increase startup time, consume memory for the lifetime of the running environment, and increase overhead on every execution, even when the Extension is warm.<br /><br />Heavy Extensions increase costs on cold start, under concurrent load, and in memory-constrained environments.                    |
  | Comply with the API call limit    | Design your Extension to:<ul><li>Manage its functionality with a buffer [below the limits](#limitations).</li><li>Compute values locally.</li><li>Consolidate writes whenever possible.</li></ul>For example, instead of calling `addToContext()` three times for three keys, build the object containing the three keys locally, and pass it once in a single API call.<br /><br />The `api.log()` function has no call limit but should still be used at appropriate log levels instead of on every variable. |
  | Use module-level singletons       | Initialize HTTP clients, database connections, or configuration objects at the module's top level. This approach is efficient, because these resources are initialized once on startup and are reused for all subsequent warm executions.                                                                                                                                                                                                                                                                       |
  | Scope execution data              | Avoid storing mutable data related to a single execution at the module level. Sharing data related to a single execution across concurrent executions can lead to unintended data sharing and bugs. Examples of such data include user input, context values, and execution results. Create and store such data inside the Extension Node function instead.                                                                                                                                                     |
  | Handle errors explicitly          | Wrap external calls and async operations in `try`/`catch` blocks and expose meaningful errors with `api.log()` or by writing them to the Context object.<br /><br />Unhandled exceptions don't crash the execution environment. The runtime catches unhandled exceptions and returns an error to the calling Flow. However, the Flow doesn't receive any output, and the execution counts against the timeout budget.                                                                                           |
  | Store temporary data in memory    | Use in-memory variables for temporary data within a single execution. The filesystem is read-only. `fs.writeFile()`, `fs.mkdir()`, and similar operations will throw an error at runtime.                                                                                                                                                                                                                                                                                                                       |
  | Store persistent data externally  | Store critical data in an external store, such as a database or file storage. Data in module-level variables persists only while the execution environment is active. After 15 minutes of inactivity, or on redeployment, the environment is cleaned up. Critical data stored in module-level variables is lost after the inactivity period or on redeployment.                                                                                                                                                 |
  | Comply with the object size limit | Validate argument [sizes](#limitations) before your Extension passes large payloads (for example, document content, API responses, or arrays) to API methods or Context objects. Alternatively, split data across smaller keys.                                                                                                                                                                                                                                                                                 |
</Accordion>

#### Publish Custom Extensions

Additionally, you can publish your custom Extension on the Marketplace so other Cognigy users can access it.

<AccordionGroup>
  <Accordion title="Develop Custom Extensions">
    The following materials provide in-depth information on developing a custom Extension:

    * [Readme](https://www.npmjs.com/package/@cognigy/extension-tools) of the `@cognigy/extension-tools` package for requirements.
    * [Cognigy Sessions Episode](https://support.cognigy.com/hc/en-us/articles/360019893139-Cognigy-Sessions-Extensions) about Extensions for a technical deep dive.
    * An [Extension Node](https://github.com/Cognigy/Extensions/tree/master/docs/example/src/nodes) example that shows how to build an Extension Node.
    * A [Knowledge Connector](https://github.com/Cognigy/Extensions/blob/master/docs/example/src/knowledge-connectors/simpleKnowledgeConnector.ts) example that shows how to build a Knowledge Connector.
    * [GitHub repository](https://github.com/Cognigy/Extensions/tree/master/extensions) for more Extension examples.

    <Check>
      [Cognigy Hammer](https://github.com/tgbv/cognigy-hammer/), created by the Cognigy community, is an Extension development suite designed for Cognigy. Cognigy Hammer offers several tools and features to assist in the development of Cognigy Extensions. Note that Cognigy Hammer isn't a product of Cognigy and doesn't qualify for enterprise support.
    </Check>

    **Best practices for developing custom Extensions**

    To guarantee the [performance](#execution-performance-and-security) of a custom Extension, make sure the Extension code:

    * Undergoes review by experienced developers.
    * Avoids very complex logic and use cases. Keep Extensions small in size and with few Nodes, for example, fewer than 20.
    * Contains only production- or runtime-relevant dependencies in the final build.
    * Is tested both independently and in Flows before production rollout.
    * Provides error handling across all potential error cases, for example, using try/catch blocks and resolving promises.

    **Code Example for Error Handling**

    ```javascript theme={null}
    function: async ({ cognigy, config }) => {
        const { api } = cognigy;
        const { contextKey, inputKey } = config;

        const endpoint = `API_TO_CALL`;

        try {
            const result = await axios.get(endpoint);

            // store in context
            api.addToContext(contextKey, result.data, 'simple');

            // store in input 
            api.addToInput(inputKey, result.data);
        }
        catch (error) {
            api.addToContext(contextKey, { error: error.message }, 'simple');
        }
    }
    ```
  </Accordion>

  <Accordion title="Publish Custom Extensions">
    If you want to publish a custom Extension on the Marketplace, follow the approval procedure in the [Extensions GitHub repository](https://github.com/Cognigy/Extensions).
  </Accordion>
</AccordionGroup>

## Execution, Performance, and Security

Before Cognigy 2026.9.0, all Extension code was considered untrusted and was executed in a secure, isolated environment. This additional security layer introduced some overhead during startup. As a result, Extensions typically ran slower than default Nodes in Flows. However, you could still [mark Extensions as trusted](#make-extensions-trusted) to execute them in the trusted environment.

After Cognigy 2026.9.0, all organizations will be gradually migrated to a more modern, unified execution runtime for all Extensions, called [Cognigy Serverless](#cognigy-serverless). This new runtime provides high security, stability, and execution isolation for all Extensions. Additionally, Cognigy Serverless provides a significant performance boost for all Extensions. [Extension API functions](/ai/for-developers/code/api-functions) are unchanged, and Extensions are expected to continue working as before. In rare cases, Extensions relying on nonstandard or unintended runtime behaviors may have their execution blocked under the new service. For more details, see [Unsupported Extension Design Patterns](#unsupported-extension-design-patterns).

The migration to Cognigy Serverless starts with the release of [Cognigy 2026.9](/release-notes/2026.9) for Cognigy SaaS installations. The migration is performed in the background, and users don't need to take any action.

<Warning>
  Cognigy Serverless is required starting with Cognigy 2026.21.0. The legacy Extension runtime for trusted and untrusted Extensions is removed in this release.

  * **SaaS installations**. Cognigy performs the migration for you. If Cognigy hasn't contacted you, your environments and Extensions are already running on Cognigy Serverless, and no action is required.
  * **On-premises installations**. [Install Cognigy Serverless](/ai/administer/installation/serverless-installation) before upgrading to 2026.21.0. Otherwise, all Extensions will stop working after the upgrade.
</Warning>

### Cognigy Serverless

*Cognigy Serverless* is the execution runtime that powers all Marketplace and [custom Extensions](#custom-extensions) with further enhanced security for the AI era, stability, and execution isolation.

In this runtime, each organization has an isolated environment for Extensions that is created on demand and automatically cleaned up after a period of inactivity. This approach provides consistent behavior and strong isolation across all organization environments.

#### Unsupported Extension Design Patterns

Specific Extension design patterns may have unintentionally worked in the previous runtime. Cognigy Serverless makes sure that previous Extension limitations are enforced to guarantee performance and security.

<Accordion title="Unsupported Patterns">
  | Pattern                                                  | Reason for failure                                                                          | Recommended alternative                                                    |
  | -------------------------------------------------------- | ------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------- |
  | Writing to the local filesystem                          | Read-only filesystem                                                                        | Use external storage on a public IP address.                               |
  | Connecting to services on private IP address ranges      | Blocked private network access                                                              | Use a publicly accessible endpoint or a SaaS equivalent.                   |
  | Storing data permanently in module-level variables       | The environment is cleaned up after 15 minutes of inactivity.                               | Write to an external database or storage service.                          |
  | Accessing internal Cognigy services directly over HTTP   | Blocked private network access                                                              | Use [Cognigy API functions](/ai/for-developers/code/api-functions).        |
  | Installing npm packages dynamically at runtime           | Package managers and filesystem write access aren't available in the execution environment. | Include all dependencies in `package.json` before uploading the Extension. |
  | Spawning child processes (`child_process.spawn`, `exec`) | Shell access and system binaries aren't available in the execution environment.             | Implement the logic natively in Node.js.                                   |
</Accordion>

### Make Extensions Trusted

<Danger>
  Never trust Extension code without a thorough review. Extensions can use external npm packages, which may contain malicious code or routines. An Extension can expose sensitive information when executed in the standard environment.
</Danger>

A trusted Extension is considered safe to run in the standard environment. This means that trusted Extensions may offer lower latency but need to be handled more carefully to avoid [performance issues](#custom-extensions).

To make Extensions trusted and let them run in the standard environment, you have two options:

<Accordion title="Make Extensions Trusted">
  * Mark an Extension as trusted in **Manage > Extensions**. Trusted Extensions display the <img alt="trust-extensions" /> icon. Only admins and users with the `extension_trust_admin` [role](/ai/administer/access/members) can mark Extensions as trusted and update them.
  * For on-premises installations:
    1. Set the `FEATURE_ALLOW_TRUSTED_CODE_CONFIGURATION` environment variable to `true` by adding the following code to your `config-map_patch.yaml` in the `kubernetes` repository where the deployment manifest files are stored:
    ```YML theme={null}
    - op: add
      path: /data/FEATURE_ALLOW_TRUSTED_CODE_CONFIGURATION
      value: "true"
    ```
    2. Use the Cognigy API [PATCH](https://api-trial.cognigy.ai/openapi#patch-/v2.0/extensions/-extensionId-) request to update the `trustedCode` property of an Extension.
</Accordion>

## Install Extensions for All Organizations

On-premises customers can install Extensions across all organizations in their installation. To do so, add the `FEATURE_ADDITIONAL_SYSTEM_WIDE_EXTENSIONS_PATH` environment variable to `values.yaml` under the `cognigyEnv` mapping key and enter the path to the Extension.

## Cache Extensions in your Local Directory

You can cache Extensions in your local directory to improve loading performance.

<Accordion title="Maximum cache storage">
  By default, when Extensions exceed the maximum cache directory size, the last 10 Extensions are removed from the local directory. On-premises customers can change the number of Extensions that are removed when the maximum cache directory size is exceeded using the `EXCEED_DIR_SIZE_AMOUNT_TO_DROP_FROM_MAP` environment variable.

  On-premises customers can change the maximum directory size by adding the `MAX_EXTENSIONS_CACHE_DIR_SIZE_IN_MB` environment variable to `values.yaml`. By default, the maximum directory size is 512 MB.

  The cache is in the `service-execution` [Kubernetes pod](https://kubernetes.io/docs/concepts/workloads/pods/).
</Accordion>

## Dynamic Fields

You can use a dynamic selection field as a field type in Extensions. You can use this feature to dynamically fetch the content of a selection field, for example, through an external API call.

## Localization for Extensions

Extension builders can include localized UI text, such as default Node labels or Node field descriptions.

For more details, read the [Localization for Extensions](https://www.npmjs.com/package/@cognigy/extension-tools#extension-localization) documentation.

## Error Handling

If an Extension [times out or sends too many API calls](#limitations), the Flow execution doesn't stop. Instead, an error message is written to the `input.extensionError` Input object.


# Node Execution API: All Functions
Source: https://docs.cognigy.com/ai/for-developers/function-reference/all-functions



The *Node Execution API* is a toolset for developers to build and control Flows. Its functions allow you to:

* **Manage Data**. Control Contact Profiles and use the Context object.
* **Control Flows**. Set which Nodes to execute.
* **Interact and Monitor**. Send messages to the user, and log or monitor the conversation execution.
* **Run Scripts**. Execute scripts and evaluate conditions within a Flow.

## CognigyScript

The *CognigyScript* functions are used to execute and evaluate CognigyScript expressions. You can use them to render text that contains CognigyScript or to evaluate a condition to a `true` or `false` value.

### [`parseCognigyScriptText`](/ai/for-developers/function-reference/cognigyscript/parseCognigyScriptText)

**Syntax:** `parseCognigyScriptText(text)`

Renders a text containing CognigyScript expressions.

### [`parseCognigyScriptCondition`](/ai/for-developers/function-reference/cognigyscript/parseCognigyScriptCondition)

**Syntax:** `parseCognigyScriptCondition(condition)`

Evaluates a CognigyScript expression and returns a `true` or `false` value.

## Contact Profiles

The *Contact Profile* functions are used to manage a user's [Contact Profile](/ai/agents/analyze/contact-profiles). They allow you to activate or deactivate a Contact Profile, update specific fields, and merge one Contact Profile with another. When a Contact Profile is deactivated, no data can be read from or written to it.

### [`activateProfile`](/ai/for-developers/function-reference/contact-profiles/activateProfile)

**Syntax:** `activateProfile()`

Activates a Contact Profile for the user.

### [`deactivateProfile`](/ai/for-developers/function-reference/contact-profiles/deactivateProfile)

**Syntax:** `deactivateProfile(deleteData)`

Deactivates a Contact Profile for the user. When the Contact Profile is deactivated, no data can be read from or written into it.

### [`updateProfile`](/ai/for-developers/function-reference/contact-profiles/updateProfile)

**Syntax:** `updateProfile(key, value)`

Updates a value in the Contact Profile.

### [`mergeProfile`](/ai/for-developers/function-reference/contact-profiles/mergeProfile)

**Syntax:** `mergeProfile(contactId)`

Merges the current Contact Profile with another Contact Profile.

## Context Object

The *Context Object* functions are used to manage the [Context object](/ai/agents/develop/ai-agent-memory/context) of a conversation. You can add, get, remove, or delete data from the Context object, as well as reset it to its default state.

### [`addToContext`](/ai/for-developers/function-reference/context-object/addToContext)

**Syntax:** `addToContext(key, value, mode)`

Adds a value to the Context object.

### [`getContext`](/ai/for-developers/function-reference/context-object/getContext)

**Syntax:** `getContext(key)`

Retrieves the value associated with a given key in the Context object.

### [`deleteContext`](/ai/for-developers/function-reference/context-object/deleteContext)

**Syntax:** `deleteContext(key)`

Removes a value from the Context object.

### [`removeFromContext`](/ai/for-developers/function-reference/context-object/removeFromContext)

**Syntax:** `removeFromContext(key, value, mode)`

Removes data from the Context object: either a key-value entry or a specific value within an array stored under that key.

### [`resetContext`](/ai/for-developers/function-reference/context-object/resetContext)

**Syntax:** `resetContext()`

Resets the current Context to the default Context of the active Flow.

## Execution and Flow Control

The *Execution and Flow Control* functions let you manage how a Flow runs. You can stop or restart a Flow, track Node executions, run other Flows, and set which Node executes next.

### [`executeFlow`](/ai/for-developers/function-reference/execution-control/executeFlow)

**Syntax:** `executeFlow(config)`

Executes a target Flow starting from a specified Entrypoint Node.

### [`setNextNode`](/ai/for-developers/function-reference/execution-control/setNextNode)

**Syntax:** `setNextNode(nodeId, flowId)`

Sets a custom Node with the given `nodeId` to be executed after the current Node. If the target Node is in a different Flow, the `flowId` parameter is required.

### [`resetNextNodes`](/ai/for-developers/function-reference/execution-control/resetNextNodes)

**Syntax:** `resetNextNodes()`

Removes any `setNextNode` instructions that were previously set.

### [`addConditionalEntrypoint`](/ai/for-developers/function-reference/execution-control/addConditionalEntrypoint)

**Syntax:** `addConditionalEntrypoint(params)`

Adds a conditional entrypoint for the current conversation. Used by the Optional Question Node.

### *(Deprecated)* [`think`](/ai/for-developers/function-reference/execution-control/think)

**Syntax:** `think(text, data)`

Restarts Flow execution with a simulated input. As of Cognigy 4.98.0, this function is deprecated.

### [`thinkV2`](/ai/for-developers/function-reference/execution-control/thinkV2)

**Syntax:** `thinkV2(text, data)`

Restarts Flow execution with a simulated input based on defined text and data.

### [`stopExecution`](/ai/for-developers/function-reference/execution-control/stopExecution)

**Syntax:** `stopExecution()`

Stops processing the current input after this Node.

### [`getExecutionAmount`](/ai/for-developers/function-reference/execution-control/getExecutionAmount)

**Syntax:** `getExecutionAmount(nodeId)`

Retrieves the number of times a Node with the given `nodeId` has been executed during this conversation.

### [`setExecutionAmount`](/ai/for-developers/function-reference/execution-control/setExecutionAmount)

**Syntax:** `setExecutionAmount(nodeId, value)`

Updates the number of executions for a Node with the given `nodeId`. This is useful for tracking how many times a Node has been executed.

### [`resetExecutionAmount`](/ai/for-developers/function-reference/execution-control/resetExecutionAmount)

**Syntax:** `resetExecutionAmount(nodeId)`

Resets the number of executions to 0 for the Node with the given `nodeId`.

## Input and Output

The *Input and Output* functions handle communication for a conversation.
You can use `say` to send messages to the user and `addToInput` to pass data to the next Node.

### [`say`](/ai/for-developers/function-reference/input-and-output/say)

**Syntax:** `say: TSayFunction`

**Alias:** `output`

Sends a message to the user.

### [`output`](/ai/for-developers/function-reference/input-and-output/output)

**Syntax:** `output: TSayFunction`

**Alias:** `say`

Sends a message to the user.

### [`addToInput`](/ai/for-developers/function-reference/input-and-output/addToInput)

**Syntax:** `addToInput(key, value)`

Writes a value into the Input object for the conversation. This function is useful for passing data to the next Node in the Flow.

## Logging and Monitoring

The *Logging and Monitoring* functions help you monitor and debug a conversation. They let you log messages and track token usage.

### [`log`](/ai/for-developers/function-reference/logging/log)

**Syntax:** `log(level, message)`

Writes a message into the execution logs.

### [`getLLMTokenUsageForSession`](/ai/for-developers/function-reference/logging/getLLMTokenUsageForSession)

**Syntax:** `getLLMTokenUsageForSession()`

Retrieves the token usage information for the conversation session. Returns `null` if no token usage information is available.

### [`setSensitiveLoggingSettings`](/ai/for-developers/function-reference/logging/setSensitiveLoggingSettings)

**Syntax:** `setSensitiveLoggingSettings(settings)`

Updates the masking redaction settings for the active conversation.

## xApp Functions

The xApp functions are used to manage conversations with xApps.

### [`setAppState`](/ai/for-developers/function-reference/xApps/setAppState)

**Syntax:** `setAppState(xAppTemplateId, xAppTemplateData)`.

Updates the state of the xApp Session for this conversation.

## Other

These are functions that don't belong to any specific group. They let you evaluate conditions, mark conversation goals as complete, and extract data from user input using regex.

### [`evaluateRule`](/ai/for-developers/function-reference/other/evaluateRule)

**Syntax:** `evaluateRule(rule)`

Evaluates a condition based on a rule configuration that follows the `left`, `operator`, `right` pattern.
This function works similarly to the If Node.

### [`completeGoal`](/ai/for-developers/function-reference/other/completeGoal)

**Syntax:** `completeGoal(goal)`

Marks a goal as `completed` for the conversation.

### [`setRegexSlot`](/ai/for-developers/function-reference/other/setRegexSlot)

**Syntax:** `setRegexSlot(params)`

Extracts Slots from user inputs using a regular expression and writes them to `input.slots`. This function is used by the [Regex Slot Filler](/ai/agents/develop/node-reference/ai/regex-slot-filler) Node.

## More Information

* [Code Node: API Functions](/ai/for-developers/code/api-functions)


# CognigyScript Functions
Source: https://docs.cognigy.com/ai/for-developers/function-reference/cognigyscript/overview



The *CognigyScript* functions are used to execute and evaluate CognigyScript expressions. You can use them to render text that contains CognigyScript or to evaluate a condition to a `true` or `false` value.

## [`parseCognigyScriptText`](/ai/for-developers/function-reference/cognigyscript/parseCognigyScriptText)

**Syntax**

`parseCognigyScriptText(text)`

Parses and renders a text string containing CognigyScript expressions. Each CognigyScript expression, for example, `{{ cc.userName }}`, is evaluated using the current execution context.

## [`parseCognigyScriptCondition`](/ai/for-developers/function-reference/cognigyscript/parseCognigyScriptCondition)

**Syntax**

`parseCognigyScriptCondition(condition)`

Evaluates a CognigyScript expression and returns a `true` or `false` value.


# parseCognigyScriptCondition
Source: https://docs.cognigy.com/ai/for-developers/function-reference/cognigyscript/parseCognigyScriptCondition



## Syntax

`parseCognigyScriptCondition(condition)`

## Description

Evaluates a CognigyScript expression and returns a `true` or `false` value.

CognigyScript allows you to embed expressions like `{{context.value > 5}}` within strings. This function extracts and evaluates such an expression, used in logic checks for Flows, conditions, or custom functions.

This function is useful for dynamically checking conditions at run-time based on session data and information from the Input, Context, or Profile objects.

**Parameters:**

* `condition`: `string` — the CognigyScript condition to evaluate.

**Returns:** `boolean | null` — the result of the condition, or `null` if evaluation fails.

## Example

```js theme={null}
const result = parseCognigyScriptCondition("{{context.score > 70}}");

if (result === true) {
  // Proceed if the condition is true
  console.log("Score is high enough");
} else if (result === false) {
  // Condition evaluated to false
  console.log("Score is too low");
} else {
  // Evaluation failed
  console.log("Invalid condition");
}
```


# parseCognigyScriptText
Source: https://docs.cognigy.com/ai/for-developers/function-reference/cognigyscript/parseCognigyScriptText



## Syntax

`parseCognigyScriptText(text)`

## Description

Parses and renders a text string containing CognigyScript expressions. Each CognigyScript expression, for example, `{{ cc.userName }}`, is evaluated using the current execution context.

**Parameters:**

* `text`: `string` — text containing CognigyScript expressions.

**Returns:** `string` — a string with all CognigyScript expressions resolved. If an expression can't be resolved, it returns an empty string.

## Example

```js theme={null}
// Assume the execution context includes:
// cc.userName = "John Doe"
// cc.city = "Berlin"

const input = "Hello {{cc.userName}}, welcome to {{cc.city}}!";
const output = parseCognigyScriptText(input);

console.log(output);
// Output: "Hello John Doe, welcome to Berlin!"
```


# activateProfile
Source: https://docs.cognigy.com/ai/for-developers/function-reference/contact-profiles/activateProfile



## Syntax

`activateProfile()`

## Description

Activates the user's Contact Profile.
Call this function to re-enable read and write operations on a previously deactivated Contact Profile.

**Returns:** `Promise<unknown>`

## Example

```js theme={null}
// Reactivate a previously deactivated profile
await activateProfile();
```


# deactivateProfile
Source: https://docs.cognigy.com/ai/for-developers/function-reference/contact-profiles/deactivateProfile



## Syntax

`deactivateProfile(deleteData)`

## Description

Deactivates the user's Contact Profile. While the Contact Profile is deactivated, no data can be read from or written to the Contact Profile.

This function can be used to temporarily disable a Contact Profile or to permanently remove data if required.

**Parameters:**

* `deleteData`: `boolean` — if set to `true`, existing data will be erased from the Contact Profile.

**Returns:** `Promise<any>`

## Example

```js theme={null}
// Temporarily deactivate the Contact Profile without deleting any data
await deactivateProfile(false);

// Deactivate the Contact Profile and permanently delete all associated data
await deactivateProfile(true);
```


# mergeProfile
Source: https://docs.cognigy.com/ai/for-developers/function-reference/contact-profiles/mergeProfile



## Syntax

`mergeProfile(contactId)`

## Description

Merges the current Contact Profile with another one.
This function is useful for combining duplicate Contact Profiles or moving data between users.

**Parameters:**

* `contactId`: `string` — the ID of the Contact Profile to merge into.

**Returns:** `Promise<unknown>`

## Example

```js theme={null}
await mergeProfile("c-89374201a7bd5a9e4");
```


# Contact Profiles
Source: https://docs.cognigy.com/ai/for-developers/function-reference/contact-profiles/overview



The *Contact Profile* functions are used to manage a user's [Contact Profile](/ai/agents/analyze/contact-profiles). They allow you to activate or deactivate a Contact Profile, update specific fields, and merge one Contact Profile with another. When a Contact Profile is deactivated, no data can be read from or written to it.

## [`activateProfile`](/ai/for-developers/function-reference/contact-profiles/activateProfile)

**Syntax:** `activateProfile()`

Activates the user's Contact Profile.
Call this function to re-enable read and write operations on a previously deactivated Contact Profile.

## [`deactivateProfile`](/ai/for-developers/function-reference/contact-profiles/deactivateProfile)

**Syntax:** `deactivateProfile(deleteData)`

Deactivates the user's Contact Profile. When the Contact Profile is deactivated, no data can be read from or written to the Contact Profile.

## [`updateProfile`](/ai/for-developers/function-reference/contact-profiles/updateProfile)

**Syntax**

`updateProfile(key, value)`

Updates a value in the current user's Contact Profile.

## [`mergeProfile`](/ai/for-developers/function-reference/contact-profiles/mergeProfile)

**Syntax**

`mergeProfile(contactId)`

Merges the current Contact Profile with another one.


# updateProfile
Source: https://docs.cognigy.com/ai/for-developers/function-reference/contact-profiles/updateProfile



## Syntax

`updateProfile(key, value)`

## Description

Updates a value in the current user's Contact Profile.

**Parameters:**

* `key`: `string` — the ID of the Contact Profile field to update.
* `value`: `unknown` — a new value for the specified Contact Profile field.

**Returns:** `Promise<void>`

## Example

```js theme={null}
// Update the user's first name
await updateProfile("firstName", "Jane");

// Update a numeric field like age
await updateProfile("age", 30);

// Update a boolean preference
await updateProfile("newsletterSubscribed", true);
```


# addToContext
Source: https://docs.cognigy.com/ai/for-developers/function-reference/context-object/addToContext



## Syntax

`addToContext(key, value, mode)`

## Description

Adds a value to the Context object.

**Parameters:**

* `key`: `string` — the key in the Context object where the value should be stored.
* `value`: `unknown` — the value to store.
* `mode`: `"simple" | "array"`:
  * `"simple"` — stores the value directly at the specified key, replacing any existing value.
  * `"array"` — adds the value to an array at the specified key, creating the array if it doesn't exist.

**Returns:** `void`

## Example

<AccordionGroup>
  <Accordion title="Simple Mode">
    ```js theme={null}
    // Initial context: {}
    addToContext("user.city", "Berlin", "simple");
    // Context after execution: { user: { city: "Berlin" } }
    // If called again with same key:
    addToContext("user.city", "Munich", "simple");
    // Context after execution: { user: { city: "Munich" } }  // Overwrites previous value
    ```
  </Accordion>

  <Accordion title="Array Mode">
    ```js theme={null}
    // Initial context: {}
    addToContext("user.purchases", "book", "array");
    // Context after execution: { user: { purchases: ["book"] } }

    // Adding more items to the same array:
    addToContext("user.purchases", "laptop", "array");
    addToContext("user.purchases", "coffee", "array");
    // Context after execution: { user: { purchases: ["book", "laptop", "coffee"] } }
    ```
  </Accordion>
</AccordionGroup>


# deleteContext
Source: https://docs.cognigy.com/ai/for-developers/function-reference/context-object/deleteContext



## Syntax

`deleteContext(key)`

## Description

Removes a value from the Context object.

**Parameters:**

* `key`: `string` — the key in the context where the value should be removed.

**Returns:** `void`

## Example

```js theme={null}
// Example usage
deleteContext("user.location");
```


# getContext
Source: https://docs.cognigy.com/ai/for-developers/function-reference/context-object/getContext



## Syntax

`getContext(key)`

## Description

Retrieves the value associated with a given key in the Context object.

**Parameters:**

* `key`: `string` — the key in the Context object from which the value should be retrieved.

**Returns:** `unknown` — the value stored at the specified key in the Context object, or `undefined` if the key doesn't exist.

## Example

```js theme={null}
const userName = getContext("user.profile.name");
log(`User name is: ${userName}`);
```


# Context Object
Source: https://docs.cognigy.com/ai/for-developers/function-reference/context-object/overview



The *Context Object* functions are used to manage the [Context object](/ai/agents/develop/ai-agent-memory/context) of a conversation. You can add, get, remove, or delete data from the Context object, as well as reset it to its default state.

## [`addToContext`](/ai/for-developers/function-reference/context-object/addToContext)

**Syntax:** `addToContext(key, value)`

Adds a value to the Context object.

## [`getContext`](/ai/for-developers/function-reference/context-object/getContext)

**Syntax:** `getContext(key)`

Retrieves the value associated with a given key in the Context object.

## [`deleteContext`](/ai/for-developers/function-reference/context-object/deleteContext)

**Syntax:** `deleteContext(key)`

Removes a value from the Context object.

## [`removeFromContext`](/ai/for-developers/function-reference/context-object/removeFromContext)

**Syntax:** `removeFromContext(key, value, mode)`

Removes data from the Context object: either a key-value entry or a specific value within an array stored under that key.

## [`resetContext`](/ai/for-developers/function-reference/context-object/resetContext)

**Syntax:** `resetContext()`

Resets the Context object to the default Context of the active Flow. This action removes all custom keys and values that were added during the conversation.


# removeFromContext
Source: https://docs.cognigy.com/ai/for-developers/function-reference/context-object/removeFromContext



## Syntax

`removeFromContext(key, value, mode)`

## Description

Removes data from the Context object: either a key-value entry or a specific value within an array stored under that key.

**Parameters:**

* `key`: `string` — the key in the Context object.
* `value`: `string` — the value to remove if the mode is set to `"array"`.
* `mode`: `"simple" | "array"`:
  * `"simple"` — deletes a key-value entry.
  * `"array"` — removes a value from the array.

**Returns:** `void`

## Example

```js theme={null}
// Delete a key-value entry
removeFromContext("user.city", "", "simple");

// Remove a specific value from an array
removeFromContext("user.purchases", "book", "array");
```


# resetContext
Source: https://docs.cognigy.com/ai/for-developers/function-reference/context-object/resetContext



## Syntax

`resetContext()`

## Description

Resets the Context object to the default Context of the active Flow. This action removes all custom keys and values that were added during the conversation.

**Returns:** `void`

## Example

```js theme={null}
// Clear all current context data and reset to default context
resetContext();
```


# addConditionalEntrypoint
Source: https://docs.cognigy.com/ai/for-developers/function-reference/execution-control/addConditionalEntrypoint



## Syntax

`addConditionalEntrypoint(params)`

## Description

Adds a conditional Entrypoint for the current conversation. This condition is used by the Optional Question Node.

If `{condition}` is met within the next `{retentionTime}` inputs, Flow execution starts at `{entrypoint}`.

**Parameters:**

* `params`: `IAddConditionalEntrypointParams` — an object that includes the Entrypoint to trigger, the duration for retaining the condition, and the condition to evaluate.

<Accordion title="IAddConditionalEntrypointParams Object Structure">
  ```js theme={null}
  interface IAddConditionalEntrypointParams {
    entrypoint: string;
    retentionTime: number;
    condition: string;
  }
  ```

  **Returns:** `void`
</Accordion>

## Example

```js theme={null}
addConditionalEntrypoint({
  entrypoint: "askEmail",
  retentionTime: 2,
  condition: "input.text.includes('@')"
});
```


# executeFlow
Source: https://docs.cognigy.com/ai/for-developers/function-reference/execution-control/executeFlow



## Syntax

`executeFlow(config)`

## Description

Executes a target Flow starting from a specified Entrypoint Node.

The behavior is controlled by `config.flowNode.isGoto`:

* `false` – the target Flow runs temporarily, then returns to the current Flow.
* `true` – the target Flow runs permanently and doesn't return to the current Flow. For this behavior, you must call [`resetNextNodes()`](/ai/for-developers/function-reference/execution-control/resetNextNodes) before calling `executeFlow()`.

Only Nodes marked as Entrypoint in the target Flow can be used.

**Parameters:**

* `config`: `IExecuteFlowNodeConfig` — the `IExecuteFlowNodeConfig` configuration object that specifies the configuration for the Flow execution.

<Accordion title="IExecuteFlowNodeConfig Object Structure">
  ```js theme={null}
  interface IExecuteFlowNodeConfig {
      flowNode: {
          flow: string;
          node: string;
          isGoto?: boolean;
      };
      parseIntents?: boolean;
      parseKeyphrases?: boolean;
      absorbContext?: boolean;
  }
  ```

  **Returns:** `Promise<void>`
</Accordion>

## Example

<Tabs>
  <Tab title="Permanent Switch (`isGoto = true`)">
    ```txt theme={null}
    Current Flow → executeFlow(target Flow) → switches → target Flow continues (no return)
    ```

    ```js theme={null}
    resetNextNodes();         // Call this function to make sure `isGoto: true` is applied
    await executeFlow({
      flowNode: {
        flow: "support-flow",     // The ID of the target Flow
        node: "node-id",          // The ID of the Entrypoint Node
        isGoto: true              // Keep execution in the target Flow
      },
      parseIntents: true,         // Enable intent parsing in the target Flow
      parseKeyphrases: false,     // Disable keyphrase parsing
      absorbContext: true         // Absorb context from the target Flow
    });
    ```
  </Tab>

  <Tab title="Temporary Run (`isGoto = false`)">
    ```txt theme={null}
    Current Flow → executeFlow(target Flow) → returns → Current Flow continues
    ```

    ```js theme={null}
    await executeFlow({
      flowNode: {
        flow: "support-flow",     // The ID of the target Flow
        node: "node-id",          // The ID of the Entrypoint Node
        isGoto: false             // Return to this Flow after execution
      },
      parseIntents: true,         // Enable intent parsing in the target Flow
      parseKeyphrases: false,     // Disable keyphrase parsing
      absorbContext: true         // Absorb context from the target Flow
    });
    ```
  </Tab>
</Tabs>


# getExecutionAmount
Source: https://docs.cognigy.com/ai/for-developers/function-reference/execution-control/getExecutionAmount



## Syntax

`getExecutionAmount(nodeId)`

## Description

Returns the number of times the Node with the given `nodeId` has been executed in the current conversation.

**Parameters:**

* `nodeId`: `string` - the Node ID.

**Returns:** `number` - the number of executions.

## Example

```js theme={null}
// Suppose you want to check if a specific Node with nodeId "c4c5c267-55d8-44ae-a123-f351d237f496" has been executed more than 3 times:
const executions = getExecutionAmount("node-1234");

if (executions > 3) {
  say("You've triggered this step several times already.");
}
```


# Execution and Flow Control
Source: https://docs.cognigy.com/ai/for-developers/function-reference/execution-control/overview



The *Execution and Flow Control* functions let you manage how a Flow runs. You can stop or restart a Flow, track node executions, run other Flows, and set which Node executes next.

## [`executeFlow`](/ai/for-developers/function-reference/execution-control/executeFlow)

**Syntax:** `executeFlow(config)`

Executes a target Flow starting from a specified Entrypoint Node.

## [`setNextNode`](/ai/for-developers/function-reference/execution-control/setNextNode)

**Syntax:** `setNextNode(nodeId, flowId)`

Sets a custom Node with the given `nodeId` to be executed immediately after the current one.

## [`resetNextNodes`](/ai/for-developers/function-reference/execution-control/resetNextNodes)

**Syntax:** `resetNextNodes()`

Clears any active overrides for the next Node in the Flow.

## [`addConditionalEntrypoint`](/ai/for-developers/function-reference/execution-control/addConditionalEntrypoint)

**Syntax:** `addConditionalEntrypoint(params)`

Adds a conditional Entrypoint for the current conversation. This condition is used by the Optional Question Node.

## *(Deprecated)* [think](/ai/for-developers/function-reference/execution-control/think)

**Syntax:** `think(text, data)`

Restarts Flow execution with a simulated input.

## [thinkV2](/ai/for-developers/function-reference/execution-control/thinkV2)

**Syntax:** `thinkV2(text, data)`

Restarts Flow execution with a simulated input using the provided text and data.

## [`stopExecution`](/ai/for-developers/function-reference/execution-control/stopExecution)

**Syntax:** `stopExecution()`

Stops processing the current input after this Node.

## [`getExecutionAmount`](/ai/for-developers/function-reference/execution-control/getExecutionAmount)

**Syntax:** `getExecutionAmount(nodeId)`

Retrieves the number of times a Node with the given `nodeId` has been executed during this conversation.

## [`setExecutionAmount`](/ai/for-developers/function-reference/execution-control/setExecutionAmount)

**Syntax:** `setExecutionAmount(nodeId, value)`

Returns the number of times the Node with the given `nodeId` has been executed in the current conversation.

## [`resetExecutionAmount`](/ai/for-developers/function-reference/execution-control/resetExecutionAmount)

**Syntax:** `resetExecutionAmount(nodeId)`

Resets the number of executions to 0 for the Node with the given `nodeId`.


# resetExecutionAmount
Source: https://docs.cognigy.com/ai/for-developers/function-reference/execution-control/resetExecutionAmount



## Syntax

`resetExecutionAmount(nodeId)`

## Description

Resets the number of executions to 0 for the Node with the given `nodeId`.

**Parameters:**

* `nodeId`: `string` — the Node ID.

**Returns:** `void`

## Example

```js theme={null}
// Suppose you have a node with the nodeId "c4c5c267-55d8-44ae-a123-f351d237f496" that should only execute once per session.
// Under certain conditions, for example, the user wants to retry, you can reset its execution count like this:

resetExecutionAmount("c4c5c267-55d8-44ae-a123-f351d237f496");
```


# resetNextNodes
Source: https://docs.cognigy.com/ai/for-developers/function-reference/execution-control/resetNextNodes



## Syntax

`resetNextNodes()`

## Description

Clears any active overrides for the next Node in the Flow.

This function is useful when:

* You've changed the next Node using a [`setNextNode()`](/ai/for-developers/function-reference/execution-control/setNextNode) function and you want to return to the default Flow behavior.
* You want to call [`executeFlow()`](/ai/for-developers/function-reference/execution-control/executeFlow) with `config.flowNode.isGoto = true` and keep the execution in the target Flow.

**Returns:** `void`

## Example

```js theme={null}
// Imagine you previously used setNextNode("fallback-node") to redirect the Flow.
// Now, you want to clear that override so the Flow continues as originally defined.

resetNextNodes();
```


# setExecutionAmount
Source: https://docs.cognigy.com/ai/for-developers/function-reference/execution-control/setExecutionAmount



## Syntax

`setExecutionAmount(nodeId, value)`

## Description

Updates the execution count for a Node with the given `nodeId`.

**Parameters:**

* `nodeId`: `string` — the Node ID.
* `value`: `number` — the new execution count to set.

**Returns:** `void`

## Example

```js theme={null}
// Set the execution count to 5 for the Node with the nodeId "c4c5c267-55d8-44ae-a123-f351d237f496"
setExecutionAmount("c4c5c267-55d8-44ae-a123-f351d237f496", 5);
```


# setNextNode
Source: https://docs.cognigy.com/ai/for-developers/function-reference/execution-control/setNextNode



## Syntax

`setNextNode(nodeId, flowId)`

## Description

Sets a custom Node with the given `nodeId` to be executed immediately after the current one.\
This function is used by If and Lookup Nodes to select child Nodes based on their configuration.

Call `setNextNode(childNodeId)` before returning to continue execution at that child Node.
The target Node must be available as an Entrypoint in the Flow.

**Parameters:**

* `nodeId`: `string` — the Node ID.
* `flowId`: `string` — the ID of the Flow containing the target Node. This parameter is optional and required if the target Node is in a different Flow.

**Returns:** `void`

## Example

```js theme={null}
// The current Node is executing inside a Flow, and you want to control which Node should be executed next based on some logic.
// In this example, the execution will jump to a Node with the nodeId "c4c5c267-55d8-44ae-a123-f351d237f496" immediately after the current Node finishes.

setNextNode("c4c5c267-55d8-44ae-a123-f351d237f496");

// To continue the conversation in a different Flow after the current Node, specify both the target nodeId and the flowId.
// The Node with nodeId "c4c5c267-55d8-44ae-a123-f351d237f496" must be marked as an Entrypoint in the target Flow with flowId "67f511b11acdc139f91946b8".
// Execution will jump directly to that Node after the current one finishes.

setNextNode("c4c5c267-55d8-44ae-a123-f351d237f496", "67f511b11acdc139f91946b8");
```


# stopExecution
Source: https://docs.cognigy.com/ai/for-developers/function-reference/execution-control/stopExecution



## Syntax

`stopExecution()`

## Description

Stops processing the current input after this Node.

**Returns:** `void`

## Example

```js theme={null}
// Stop execution after this Node to prevent further processing of the current input
stopExecution();
```


# think (Deprecated)
Source: https://docs.cognigy.com/ai/for-developers/function-reference/execution-control/think



<Warning>
  As of Cognigy 4.98.0, the `think` function is deprecated. Use `thinkV2` instead.
</Warning>

## Syntax

`think(text, data)`

## Description

Restarts Flow execution with a simulated input.

**Parameters:**

* `text`: `string` — simulated input text.
* `data`: `Record<string, unknown>` — simulated input data.

**Returns:** `void`

## Example

```js theme={null}
think("I want to book a flight", {
  userLocation: "Berlin",
  destination: "Paris"
});
```


# thinkV2
Source: https://docs.cognigy.com/ai/for-developers/function-reference/execution-control/thinkV2



## Syntax

`thinkV2(text, data)`

## Description

Restarts Flow execution with a simulated input using the provided text and data.

**Parameters:**

* `text`: `string` — simulated input text.
* `data`: `Record<string, unknown>` — simulated input data.

**Returns:** `void`

## Example

```js theme={null}
thinkV2("Check order status", {
  orderId: "A12345",
  customerType: "premium"
});
```


# addToInput
Source: https://docs.cognigy.com/ai/for-developers/function-reference/input-and-output/addToInput



## Syntax

`addToInput(key, value)`

## Description

Writes a value into the Input object for the execution.
This function allows you to store data in the Input object, which can be accessed later in the execution context.

**Parameters:**

* **key** — the key in the Input object.
* **value** — the value to store.

**Returns:** `void`

## Example

```js theme={null}
// Add a user's preferred language to the Input object
addToInput("userLanguage", "en-US");

// Later in the Flow, you can access input.userLanguage
console.log(input.userLanguage); // Output: "en-US"
```


# output
Source: https://docs.cognigy.com/ai/for-developers/function-reference/input-and-output/output



## Syntax

`output: TSayFunction`

**Alias:** `say`

## Description

Sends a message to the user.

**Parameters:**

`TSayFunction` has the following parameters:

* **text** — the message to send. Can be an empty string `""` or `null` if you're only sending structured content.
* **data** — a structured content payload containing rich UI elements like buttons, carousels, quick replies, and images. This parameter is optional.
* **settings** — a configuration object related to a handover to a human agent, such as whether the message can be forwarded and the output destination (user, human agent, or both). This parameter is optional.

**Returns:** depends on the `TSayFunction` implementation.

## Example

```js theme={null}
output("Hello, world!", { key: "value" });
```

For more details on `TSayFunction` and its examples, refer to the [`say` function documentation](/ai/for-developers/function-reference/input-and-output/say).


# Input and Output
Source: https://docs.cognigy.com/ai/for-developers/function-reference/input-and-output/overview



The *Input and Output* functions handle communication for a conversation.
You can use `say` to send messages to the user and `addToInput` to pass data to the next Node.

## [`addToInput`](/ai/for-developers/function-reference/input-and-output/addToInput)

**Syntax:** `addToInput(key, value)`

Writes a value into the Input object for the execution.
This function allows you to store data in the Input object, which can be accessed later in the execution context.

## [`output`](/ai/for-developers/function-reference/input-and-output/output)

**Syntax:** `output: TSayFunction`

**Alias:** `say`

Sends a message to the user.

## [`say`](/ai/for-developers/function-reference/input-and-output/say)

**Syntax:** `say: TSayFunction`

**Alias:** `output`

Sends a message to the user.


# say
Source: https://docs.cognigy.com/ai/for-developers/function-reference/input-and-output/say



## Syntax

`say: TSayFunction`

**Alias:** `output`

## Description

Sends a message to the user.

**Parameters:**

`TSayFunction` has the following parameters:

* **text** — the message to send. Can be an empty string `""` or `null` if you're only sending structured content.
* **data** — a structured content payload containing rich UI elements like buttons, carousels, quick replies, and images. This parameter is optional.
* **settings** — a configuration object related to a handover to a human agent, such as whether the message can be forwarded and the output destination (user, human agent, or both). This parameter is optional.

**Returns:** depends on the `TSayFunction` implementation.

## Types

```js theme={null}
interface ILiveAgentSettings {
  forwardable: boolean;
  outputDestination?: "userOnly" | "agentOnly" | "userAndAgent";
}

interface ISayNodeSettings {
  liveAgentSettings: ILiveAgentSettings;
}

type TSayFunction = (
  text: string,
  data?: IStructuredContentData,
  settings?: ISayNodeSettings
) => void;
```

### Structured Content Types

```js theme={null}
export interface IStructuredContentBase {
  type: string;
  text?: string;
  fallbackText?: string;
}

export interface IStructuredContentObject {
  _gallery?: ICarouselStructuredContent;
  _list?: IListStructuredContent;
  _quickReplies?: IQuickRepliesStructuredContent;
  _buttons?: IButtonsStructuredContent;
  _adaptiveCard?: IAdaptiveCardStructuredContent;
  _image?: IImageStructuredContent;
  _video?: IVideoStructuredContent;
  _audio?: IAudioStructuredContent;
}

export interface ITextStructuredContent extends IStructuredContentBase {
  type: 'text';
  text: string;
}

export interface IQuickRepliesStructuredContent extends IStructuredContentBase {
  type: 'quick_replies';
  quickReplies: IQuickReply[];
}

export interface IImageStructuredContent extends IStructuredContentBase {
  type: 'image';
  imageUrl: string;
  imageAltText?: string;
}

export interface IVideoStructuredContent extends IStructuredContentBase {
  type: 'video';
  videoUrl: string;
  videoAltText?: string;
  videoCaptionsUrl?: string;
}

export interface IAudioStructuredContent extends IStructuredContentBase {
  type: 'audio';
  audioUrl: string;
  audioAltText?: string;
}

export interface ICarouselStructuredContent extends IStructuredContentBase {
  type: 'carousel';
  items: ICarouselItem[];
}

export interface IListStructuredContent extends IStructuredContentBase {
  type: 'list';
  button: IUrlButton | IPostbackButton | IPhoneNumberButton;
  items: IListItem[];
  largeTopItem?: boolean;
}

export interface IButtonsStructuredContent extends IStructuredContentBase {
  type: 'buttons';
  text: string;
  buttons: (IUrlButton | IPostbackButton | IPhoneNumberButton)[];
}

export interface IAdaptiveCardStructuredContent extends IStructuredContentBase {
  type: 'adaptiveCard';
  adaptiveCard: object;
}

export type TStructuredContent =
  | ITextStructuredContent
  | IQuickRepliesStructuredContent
  | IImageStructuredContent
  | IVideoStructuredContent
  | IAudioStructuredContent
  | ICarouselStructuredContent
  | IListStructuredContent
  | IButtonsStructuredContent
  | IAdaptiveCardStructuredContent;
```

#### Supporting Types

```js theme={null}
interface IListItem {
  title: string;
  subtitle: string;
  imageUrl: string;
  imageAltText?: string;
  defaultActionUrl?: string;
  buttons: (IUrlButton | IPostbackButton | IPhoneNumberButton)[];
}

interface ICarouselItem {
  title: string;
  subtitle: string;
  imageUrl: string;
  imageAltText?: string;
  buttons: (IUrlButton | IPostbackButton | IPhoneNumberButton)[];
}

interface IQuickReply {
  title: string;
  payload: string;
  contentType: string;
  imageUrl: string;
  imageAltText?: string;
}

interface IButtonBase {
  type: string;
  title: string;
}

interface IUrlButton extends IButtonBase {
  type: 'web_url';
  url: string;
  target?: '_self' | '_blank';
}

interface IPostbackButton extends IButtonBase {
  type: 'postback';
  payload: string;
}

interface IPhoneNumberButton extends IButtonBase {
  type: 'phone_number';
  payload: string;
}
```

## Example

<AccordionGroup>
  <Accordion title="Send a Simple Text Message">
    ```js theme={null}
    say("Hello, world!");
    ```
  </Accordion>

  <Accordion title="Send Structured Content: Text and Quick Replies">
    ```js theme={null}
    say("Choose an option:", {
      _data: {
        _cognigy: {
          _default: {
            _quickReplies: {
              type: "quick_replies",
              quickReplies: [
                { title: "Yes", payload: "yes", contentType: "text", imageUrl: "", imageAltText: "" },
                { title: "No", payload: "no", contentType: "text", imageUrl: "", imageAltText: "" }
              ]
            }
          }
        }
      }
    });
    ```
  </Accordion>

  <Accordion title="Send Structured Content: Carousel">
    ```js theme={null}
    say("", {
      _data: {
        _cognigy: {
          _default: {
            _gallery: {
              type: "carousel",
              items: [
                {
                  title: "Item 1",
                  subtitle: "Description for item 1",
                  imageUrl: "https://example.com/image1.jpg",
                  buttons: [
                    { type: "web_url", title: "View", url: "https://example.com" }
                  ]
                },
                {
                  title: "Item 2",
                  subtitle: "Description for item 2",
                  imageUrl: "https://example.com/image2.jpg",
                  buttons: [
                    { type: "postback", title: "Select", payload: "select_item_2" }
                  ]
                }
              ]
            }
          }
        }
      }
    });
    ```
  </Accordion>

  <Accordion title="Send a Message Only to the Human Agent">
    ```js theme={null}
    say("This message is for the human agent only.", null, {
      liveAgentSettings: {
        forwardable: true,
        outputDestination: "agentOnly"
      }
    });
    ```
  </Accordion>
</AccordionGroup>


# getLLMTokenUsageForSession
Source: https://docs.cognigy.com/ai/for-developers/function-reference/logging/getLLMTokenUsageForSession



## Syntax

`getLLMTokenUsageForSession()`

## Description

Retrieves the token usage information for the current session.

**Returns:** `TSessionUsageInformation | null` — the number of input and output tokens used, or `null` if the data is unavailable.

<Accordion title="TSessionUsageInformation object">
  ```js theme={null}
  type TSessionUsageInformation = {
      [key: string]: {
          llmDisplayName: string;
          providerType: string;
          modelType: string;
          usage: {
              inputTokens: number;
              outputTokens: number;
          }
      }
  };
  ```
</Accordion>

## Example

```js theme={null}
const usage = getLLMTokenUsageForSession();

if (usage) {
  Object.entries(usage).forEach(([key, info]) => {
    console.log(`Model: ${info.llmDisplayName}`);
    console.log(`Provider: ${info.providerType}`);
    console.log(`Type: ${info.modelType}`);
    console.log(`Input tokens: ${info.usage.inputTokens}`);
    console.log(`Output tokens: ${info.usage.outputTokens}`);
    console.log('---');
  });
} else {
  console.log('Token usage data not available for this session.');
}
```


# log
Source: https://docs.cognigy.com/ai/for-developers/function-reference/logging/log



## Syntax

`log(level, message)`

## Description

Writes a log entry to the execution logs.

**Parameters:**

* `level`: `"fatal" | "error" | "warn" | "info" | "debug" | "trace"` — the log level.
* `message`: `string` — the log entry text.

**Returns:** `void`

## Example

```js theme={null}
// Example: Write an error log entry
log("error", "Failed to load user profile due to network error.");

// Example: Write an info log entry
log("info", "User profile loaded successfully.");
```


# Logging and Monitoring
Source: https://docs.cognigy.com/ai/for-developers/function-reference/logging/overview



The *Logging and Monitoring* functions help you monitor and debug a conversation. They let you log messages and track token usage.

## [`log`](/ai/for-developers/function-reference/logging/log)

**Syntax:** `log(level, message)`

Writes a log entry to the execution logs.

## [`getLLMTokenUsageForSession`](/ai/for-developers/function-reference/logging/getLLMTokenUsageForSession)

**Syntax:** `getLLMTokenUsageForSession()`

Retrieves the token usage information for the conversation session. Returns `null` if no token usage information is available.

## [`setSensitiveLoggingSettings`](/ai/for-developers/function-reference/logging/setSensitiveLoggingSettings)

**Syntax:** `setSensitiveLoggingSettings(settings)`

Adjusts the masking and redaction configuration for the conversation.


# setSensitiveLoggingSettings
Source: https://docs.cognigy.com/ai/for-developers/function-reference/logging/setSensitiveLoggingSettings



## Syntax

`setSensitiveLoggingSettings(settings)`

## Description

Adjusts the masking and redaction configuration for the conversation.

**Parameters:**

* `settings`: `ISensitiveLoggingSettings` — an object containing the sensitive logging settings.

<Accordion title="ISensitiveLoggingSettings Object Structure">
  ```js theme={null}
  export interface ISensitiveLoggingSettings {
      maskLogging: boolean;
      maskAnalytics: boolean;
      maskIPAddress?: boolean;
      disableConversations: boolean;
      disableIntentTrainer: boolean;
  }
  ```

  **Returns:** `void`
</Accordion>

## Example

```js theme={null}
// Example: Update sensitive logging settings for the current conversation
setSensitiveLoggingSettings({
  maskLogging: true,
  maskAnalytics: false,
  disableConversations: false,
  disableIntentTrainer: true,
});
```


# completeGoal
Source: https://docs.cognigy.com/ai/for-developers/function-reference/other/completeGoal



## Syntax

`completeGoal(goal)`

## Description

Marks a goal as completed for the current conversation. Completing a goal may influence reporting, analytics, or decision-making in the Flow.

**Parameters:**

* `goal`: `string` — the name of the goal to mark as completed.

**Returns:** `void`

## Example

```js theme={null}
// Assume the user has successfully completed the "scheduleAppointment" Flow

completeGoal("scheduleAppointment");
```


# evaluateRule
Source: https://docs.cognigy.com/ai/for-developers/function-reference/other/evaluateRule



<a href="/release-notes/2026.5"><Badge>Updated in 2026.5</Badge></a>

## Syntax

`evaluateRule(rule)`

## Description

Evaluates a condition based on a rule configuration that follows the `left`, `operator`, `right` pattern. This function works similarly to the If Node. The `left` value is compared with the `right` value using the specified operator (for example, `!=`, `>`, `includes`). If the condition evaluates to `true`, the associated logic or transition is executed.

This function is asynchronous.

**Parameters:**

* `rule`: `IRule` – a rule configuration object.

<Accordion title="IRule Object Structure">
  ```js theme={null}
  export const ruleOperands = <const>[
      "lt",
      "lte",
      "eq",
      "neq",
      "gt",
      "gte",
      "exists",
      "nexists",
      "contains",
      "ncontains",
      "isyes",
      "isno"
  ];

  export type TRuleOperand = typeof ruleOperands[number];

  export interface IRule {
      left: string;
      operand: TRuleOperand;
      right: string;
  }
  ```

  **Returns:** `Promise<boolean>` – a promise that resolves to `true` if the rule condition is met, otherwise `false`.
</Accordion>

## Example

```js theme={null}
const rule: IRule = {
  left: "hello world",
  operand: "contains",
  right: "hello"
};

const result = await evaluateRule(rule);

console.log(result); // true
```


# Other Functions
Source: https://docs.cognigy.com/ai/for-developers/function-reference/other/overview



These are functions that don't belong to any specific group. They let you evaluate conditions, mark conversation goals as complete, and extract data from user input using regex.

## [`evaluateRule`](/ai/for-developers/function-reference/other/evaluateRule)

**Syntax**

`evaluateRule(rule)`

Evaluates a condition based on a rule configuration that follows the `left`, `operator`, `right` pattern. This function works similarly to the If Node.

## [`completeGoal`](/ai/for-developers/function-reference/other/completeGoal)

**Syntax**

`completeGoal(goal)`

Marks a goal as `completed` for the conversation.

## [`setRegexSlot`](/ai/for-developers/function-reference/other/setRegexSlot)

**Syntax**

`setRegexSlot(params)`

Extracts Slots from user inputs using a regular expression and writes them to `input.slots`. This function is used by the [Regex Slot Filler](/ai/agents/develop/node-reference/ai/regex-slot-filler) Node.


# setRegexSlot
Source: https://docs.cognigy.com/ai/for-developers/function-reference/other/setRegexSlot



## Syntax

`setRegexSlot(params)`

## Description

Extracts Slots from user inputs using a regular expression and writes them to `input.slots`. This function is used by the [Regex Slot Filler](/ai/agents/develop/node-reference/ai/regex-slot-filler) Node.

**Parameters:**

* `params`: `IRegexSlotFillerParams` – a configuration object with regex, flags, and Slot information.

<Accordion title="IRegexSlotFillerParams Object Structure">
  ```js theme={null}
  export interface INodeExecutionCognigyObject extends IExecutionObjects {
      api: INodeExecutionAPI;
  }

  export type TNodeChildConfigs = {
      id: string;
      type: string;
      config: {
          [key: string]: unknown
      };
  };

  export interface ICognigyNodeFunctionParams {
      cognigy: INodeExecutionCognigyObject;
      childConfigs: TNodeChildConfigs[];
      config: object;
      nodeId: string;
  }

  export interface IRegexSlotFillerConfig {
      regex: string;
      flags: string;
      slot: string;
  }

  interface IRegexSlotFillerParams extends
      ICognigyNodeFunctionParams {
      config: IRegexSlotFillerConfig;
  }
  ```

  **Returns:** `Promise<void>`
</Accordion>

## Example

```js theme={null}
const params: IRegexSlotFillerParams = {
  cognigy: {
    input: {
      text: "My ZIP code is 90210",
      // ...other Cognigy input properties
    },
    context: {},
    session: {},
    api: cognigyApi // typically injected by the platform
  },
  childConfigs: [],
  config: {
    regex: "\\b\\d{5}\\b",
    flags: "g",
    slot: "zipCode"
  },
  nodeId: "node-abc-123"
};

await setRegexSlot(params);
```


# xApps Functions
Source: https://docs.cognigy.com/ai/for-developers/function-reference/xApps/overview



The xApp functions are used to manage conversations with xApps.

## [`setAppState`](/ai/for-developers/function-reference/xApps/setAppState)

**Syntax:** `setAppState(xAppTemplateId, xAppTemplateData)`

Updates the state of the xApp Session for this conversation.


# setAppState
Source: https://docs.cognigy.com/ai/for-developers/function-reference/xApps/setAppState



## Syntax

`setAppState(xAppTemplateId, xAppTemplateData)`

## Description

Updates the state of the xApp Session for the current conversation. Only works if an xApp Session was started beforehand.

**Parameters:**

* `xAppTemplateId`: `string` — the URL to a hosted xApp template, `generic-html`, or `null`.
  * If a URL is provided, the xApp template is loaded from that URL.
  * If `generic-html` is provided, the xApp page SDK is loaded from the host, and the HTML code defined in `xAppTemplateData.html` is displayed.
  * If `null` is provided, the active xApp page is closed.
* `xAppTemplateData`: `Record<string, unknown>` — the data payload for the xApp page.

**Returns:** `void`

## Examples

<AccordionGroup>
  <Accordion title="Load an xApp Template">
    ```js theme={null}
    // Load an xApp Template
    // loads the xApp template hosted at `https://my-xapp-template-url` as an xApp page
    // - injects `{ my: "xapp-template-data" }` to the xApp
    // - waits for the xApp page to load, then displays it to the end user
    setAppState("https://my-xapp-template-url", { my: "xapp-template-data" })
    ```
  </Accordion>

  <Accordion title="Load a Generic HTML xApp page">
    ```js theme={null}
    // Load generic HTML
    /* javascript
    const html = `
    <html>
      <body>
        Hello, World!
        <button type="button" onclick="SDK.submit({});">continue</button>
        <script src="/sdk/app-page-sdk.js"></script>
      </body>
    </html>
     */
    // loads the HTML code defined in `html` as an xApp page.
    // loads the xApp page SDK from its host
    setAppState("generic-html", { html })
    ```
  </Accordion>

  <Accordion title="Close an Active xApp Page">
    ```js theme={null}
    // Close active xApp page
    setAppState(null, {})
    ```
  </Accordion>
</AccordionGroup>


# Functions
Source: https://docs.cognigy.com/ai/for-developers/functions

Execute long-running JavaScript functions in Cognigy for async processes like HTTP API calls to third-party systems.

*Functions* are JavaScript functions that you can define and execute in Cognigy. Use Functions to assist with asynchronous processes, such as interacting with third-party systems through an HTTP API.

The execution of Functions is independent of Flows, but you can trigger them from a Flow. You can also use the [inject and notify](/ai/agents/deploy/endpoints/inject-and-notify) API endpoints in a Function to send results back into a Flow.

You can access the `parameters` and `api` objects from the Function's arguments.

## Limitations

* The maximum run-time of a Function instance is 15 minutes. For on-premises installations, you can change this limit via the `FUNCTION_EXECUTION_MAX_EXECUTION_TIME_IN_MINUTES` environment variable in the `values.yaml` file. For long-running operations that exceed this limit, set up a dedicated middleware service, such as AWS Lambda or Azure Functions, and call back using the [Inject API](/ai/agents/deploy/endpoints/inject-and-notify) once the process completes. This approach is more scalable than Cognigy Functions for time-intensive tasks.
* You can define a Function only in JavaScript or TypeScript.

## Working with Functions

<Tabs>
  <Tab title="GUI">
    You can view, create, edit, and delete Functions in **Build > Functions**. To define, trigger, and monitor a Function, use the **Code** and **Instances** tabs.

    Also, you can trigger a Function from the Flow editor by using the [Trigger Function Node](/ai/agents/develop/node-reference/service/trigger-function).
  </Tab>

  <Tab title="API">
    You can view, create, edit, and delete Functions using the [Cognigy API](https://api-trial.cognigy.ai/openapi#tag--Functions-v2.0). Using this API, you can also monitor, trigger, and stop a Function instance.
  </Tab>
</Tabs>

## Examples

<Note>
  You can use HTTP requests to interact with third-party systems.
</Note>

#### GET Request

```js theme={null}
export default async ({ parameters, api }: IFunctionExecutionArguments) => {
    const response = await api.httpRequest({
        method: "get",
        url: "https://example.com/api/v2",
        headers: {
            "Accept": "application/json"
        }
    });
}
```

#### POST Request

```js theme={null}
export default async ({ parameters, api }: IFunctionExecutionArguments) => {
    const response = await api.httpRequest({
        method: "post",
        url: "https://api.github.com/repos/octocat/Hello-World/issues",
        headers: {
            "Accept": "application/vnd.github.v3+json",
            "Authorization": "Bearer <your-github-token>"
        },
        data: {
            "title": "Found a bug",
            "body": "I'm having a problem with this.",
            "labels": ["bug"]
        }
    });
}
```

<Note>
  Functions can interact with Flows through the [Inject and Notify](/ai/agents/deploy/endpoints/inject-and-notify) APIs.

  The following examples assume that you pass `userId` and `sessionId` through the function's parameters.
</Note>

#### Injecting Text into Flows

```js theme={null}
export default async ({ parameters, api }: IFunctionExecutionArguments) => {
    const { userId, sessionId } = parameters;

    api.inject({
        userId,
        sessionId,
        text: "This text was injected through a Function"
    });
}
```

#### Sending Notifications

```js theme={null}
export default async ({ parameters, api }: IFunctionExecutionArguments) => {
    const { userId, sessionId } = parameters;

    api.notify({
        userId,
        sessionId,
        text: "This text was injected through a Function"
    });
}
```

Find more Function examples in the Cognigy's [functions](https://github.com/Cognigy/Functions?tab=readme-ov-file) repository.

## Use Case

Functions can help you handle HTTP requests that exceed the timeout in the [HTTP Request Node](/ai/agents/develop/node-reference/service/http-request). The HTTP Request Node allows direct access to API responses from the Flow, which is not possible with Functions. However, Functions allow you to work with slower APIs or long-running processes.

## More Information

* [Inject and Notify](/ai/agents/deploy/endpoints/inject-and-notify)
* [Trigger Function Node](/ai/agents/develop/node-reference/service/trigger-function)


# createKnowledgeChunk
Source: https://docs.cognigy.com/ai/for-developers/knowledge-connector/createKnowledgeChunk



## Syntax

`api.createKnowledgeChunk(params)`

## Description

Creates a [Knowledge Chunk](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk) in a [Knowledge Source](/ai/agents/develop/knowledge-ai/knowledge-source/overview). A Knowledge Chunk is a discrete piece of knowledge containing text and optional Chunk metadata that can be searched and retrieved during AI Agent interactions.

**Parameters:**

* `params`: `CreateKnowledgeChunkParams` — configuration object for the Knowledge Chunk. The object contains the following parameters:

| Parameter           | Type   | Required | Description                                                                                               |
| ------------------- | ------ | -------- | --------------------------------------------------------------------------------------------------------- |
| `knowledgeSourceId` | String | Yes      | The ID of the Knowledge Source.                                                                           |
| `text`              | String | Yes      | The content of the Knowledge Chunk.                                                                       |
| `data`              | Object | No       | Metadata for the chunk. An object where keys are strings and values can be strings, numbers, or booleans. |

**Returns:** `Promise<{}>` — a promise that resolves to an empty object when the Knowledge Chunk is created.

## Example

```js theme={null}
await api.createKnowledgeChunk({
    knowledgeSourceId: "ks_abc123def456",
    text: "To reset your password, click 'Forgot Password' on the login page and follow the instructions sent to your email.",
    data: {
        category: "account-management",
        priority: 5,
        lastUpdated: "2025-11-01",
        verified: true
    }
});
```


# createKnowledgeSource
Source: https://docs.cognigy.com/ai/for-developers/knowledge-connector/createKnowledgeSource



<Warning>
  As of Cognigy 2026.5.0, the `createKnowledgeSource` function is deprecated. For Knowledge Connectors, use [`upsertKnowledgeSource`](/ai/for-developers/knowledge-connector/upsertKnowledgeSource) to create or update Knowledge Sources that also support knowledge synchronization.
</Warning>

## Syntax

`api.createKnowledgeSource(params)`

## Description

Creates a [Knowledge Source](/ai/agents/develop/knowledge-ai/knowledge-source/overview) in a [Knowledge Store](/ai/agents/develop/knowledge-ai/knowledge-store). A Knowledge Source is a container for [Knowledge Chunks](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk) to organize and manage your AI Agents' knowledge.

**Parameters:**

* `params`: `CreateKnowledgeSourceParams` — configuration object for the Knowledge Source. The object contains the following parameters:

| Parameter     | Type   | Required | Description                                                                 |
| ------------- | ------ | -------- | --------------------------------------------------------------------------- |
| `name`        | String | Yes      | The name of the Knowledge Source. Must be unique for each Knowledge Source. |
| `description` | String | No       | A description of the Knowledge Source.                                      |
| `tags`        | Array  | No       | Tags associated with the Knowledge Source, used to filter searches.         |
| `chunkCount`  | Number | Yes      | The number of Knowledge Chunks in the Knowledge Source.                     |

**Returns:** `Promise<{ knowledgeSourceId: string }>` — a promise that resolves to an object containing the ID of the created Knowledge Source.

## Example

```js theme={null}
const result = await api.createKnowledgeSource({
  name: "Customer Support FAQ",
  description: "Frequently asked questions from the support portal",
  tags: ["support", "faq", "customer-service"],
  chunkCount: 75
});

console.log(result.knowledgeSourceId);
```


# deleteKnowledgeSource
Source: https://docs.cognigy.com/ai/for-developers/knowledge-connector/deleteKnowledgeSource



## Syntax

`api.deleteKnowledgeSource(params)`

## Description

Deletes a [Knowledge Source](/ai/agents/develop/knowledge-ai/knowledge-source/overview) and the relative [Knowledge Chunks](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk).

**Parameters:**

* `params`: `DeleteKnowledgeSourceParams` — configuration object containing the source ID. The object contains the following parameters:

| Parameter           | Type   | Description                     |
| ------------------- | ------ | ------------------------------- |
| `knowledgeSourceId` | string | The ID of the Knowledge Source. |

**Returns:** `Promise<{}>` — a promise that resolves to an empty object when the deletion is complete.

## Example

```js theme={null}
await api.deleteKnowledgeSource({
  knowledgeSourceId: "ks_abc123def456"
});

console.log("Knowledge source deleted successfully");
```


# Knowledge Connector API
Source: https://docs.cognigy.com/ai/for-developers/knowledge-connector/overview



The *Knowledge Connector API* is a toolset for developers to create custom [Knowledge Connectors](/ai/agents/develop/knowledge-ai/knowledge-connectors) and integrate external knowledge databases into Cognigy. The toolset is intended be used within the [Knowledge Connector implementation](https://www.npmjs.com/package/@cognigy/extension-tools#user-content-knowledge-connector). This API provides functions to create and delete [Knowledge Sources](/ai/agents/develop/knowledge-ai/knowledge-source/overview) and the relative [Knowledge Chunks](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk).

## Use Cases

* **Create and Update Knowledge Sources** — detect documents that have been created or changed and import them from cloud-based document repositories, for example, SharePoint, Confluence.
* **API Integration** — automate knowledge imports from REST APIs.

## Knowledge Connector API Functions

### [`upsertKnowledgeSource`](/ai/for-developers/knowledge-connector/upsertKnowledgeSource)

**Syntax:** `api.upsertKnowledgeSource(params)`

Creates or updates a Knowledge Source with the specified name, description, Source Tags, and number of Knowledge Chunks. Supports knowledge synchronization.

### [`createKnowledgeSource`](/ai/for-developers/knowledge-connector/createKnowledgeSource) (deprecated)

**Syntax:** `api.createKnowledgeSource(params)`

Creates a new Knowledge Source with the specified name, description, Source Tags, and number of Knowledge Chunks.

### [`deleteKnowledgeSource`](/ai/for-developers/knowledge-connector/deleteKnowledgeSource)

**Syntax:** `api.deleteKnowledgeSource(params)`

Deletes a Knowledge Source.

### [`createKnowledgeChunk`](/ai/for-developers/knowledge-connector/createKnowledgeChunk)

**Syntax:** `api.createKnowledgeChunk(params)`

Creates a Knowledge Chunk in a Knowledge Source, containing text content and optional Chunk metadata.


# upsertKnowledgeSource
Source: https://docs.cognigy.com/ai/for-developers/knowledge-connector/upsertKnowledgeSource



## Syntax

`api.upsertKnowledgeSource(params)`

## Description

Creates or updates a [Knowledge Source](/ai/agents/develop/knowledge-ai/knowledge-source/overview) in a [Knowledge Store](/ai/agents/develop/knowledge-ai/knowledge-store). Use `upsertKnowledgeSource` in [Knowledge Connectors](/ai/agents/develop/knowledge-ai/knowledge-connectors) to allow knowledge synchronization.

The API compares `externalIdentifier`, `chunkCount`, and `contentHashOrTimestamp` to detect changes to the external source. If there are any changes, the Knowledge Source and the relative Knowledge Chunks are updated in Cognigy.

**Parameters:**

* `params`: `UpsertKnowledgeSourceParams` — configuration object. The object contains the following parameters:

| Parameter                | Type   | Required | Description                                                                                                                                          |
| ------------------------ | ------ | -------- | ---------------------------------------------------------------------------------------------------------------------------------------------------- |
| `name`                   | String | Yes      | The name of the Knowledge Source.                                                                                                                    |
| `description`            | String | No       | A description of the Knowledge Source.                                                                                                               |
| `tags`                   | Array  | No       | Source Tags associated with the Knowledge Source, used to filter searches.                                                                           |
| `chunkCount`             | Number | Yes      | The number of Knowledge Chunks in the Knowledge Source.                                                                                              |
| `contentHashOrTimestamp` | String | Yes      | Used to detect content changes, for example, with a SHA-256 hash. If it changes, the Knowledge Source and the relative Knowledge Chunks are updated. |
| `externalIdentifier`     | String | No       | Stable identifier for the Knowledge Source. The default value is the Knowledge Source name.                                                          |

**Returns:** `Promise<{ knowledgeSourceId: string, ... } | null>` — resolves to the Knowledge Source when changes to the external source are detected and Knowledge Chunks need to be updated, or `null` when there are no changes, however, in this case the description and Source Tags are still updated.

## Example

The following Knowledge Connector uses `upsertKnowledgeSource` to create or update a source, then adds Knowledge Chunks when the source is returned:

```js expandable theme={null}
import * as crypto from "node:crypto";
import { createKnowledgeConnector } from "@cognigy/extension-tools";

export const myKnowledgeImport = createKnowledgeConnector({
	type: "myKnowledgeImport",
	label: "My Knowledge Import",
	summary: "My own Knowledge Extension",
	fields: [
		{
			key: "name",
			label: "Name",
			type: "text",
			params: {
				required: true,
			},
		},
		{
			key: "color",
			label: "Color",
			type: "text",
			params: {
				required: true,
			},
		},
	] as const,
	function: async ({ config, api, sources }) => {
		const text = `The color of ${config.name} is ${config.color}`;
		const contentHash = crypto.hash("sha256", text, "hex");

		const knowledgeSource = await api.upsertKnowledgeSource({
			name: config.name,
			description: "This is my own knowledge source",
			tags: ["example"],
			chunkCount: 1,
			contentHashOrTimestamp: contentHash,
		});

		if (knowledgeSource) {
			await api.createKnowledgeChunk({
				knowledgeSourceId: knowledgeSource.knowledgeSourceId,
				text,
				data: {},
			});
		}

		const createdSources = [config.name];
		for (const source of sources) {
			if (!createdSources.includes(source.externalIdentifier)) {
				await api.deleteKnowledgeSource({
					knowledgeSourceId: source.knowledgeSourceId,
				});
			}
		}
	},
});
```


# Execution Finished Transformer
Source: https://docs.cognigy.com/ai/for-developers/transformers/execution-finished-transformer



<Warning>
  In Cognigy 2025.25.0, transformers will start using the `axios` library for handling HTTP requests internally instead of `request-promise`. This change occurs to improve security and maintainability. The following HTTP request options will still be supported:

  * **URL configuration** — `uri`, `url`
  * **Request method** — `method`
  * **Request headers and configuration** — `headers`
  * **Request body** — `form`, `json`
  * **Query parameters** — `qs`, `qsStringifyOptions`
  * **Authentication** — `auth`
  * **Connection and network** — `proxy`, `maxRedirects`, `gzip`, `encoding`

  If you use other HTTP request options, review your transformer code to ensure it is compatible with the `axios` library before the release of Cognigy 2025.25.0.
</Warning>

The *Execution Finished Transformer* converts the Flow output after a Flow execution has finished. The return value of the execution finished transformer is sent to the Endpoint. This transformer applies only to REST-based Endpoints using the output transformer.

You can configure the execution finished transformer in the `handleExecutionFinished` function in the [Endpoint settings or via CLI](/ai/for-developers/transformers/overview#working-with-transformers).

## Transformer Function Arguments

The following table shows an overview of the function arguments:

| Argument        | Description                                                                            |
| --------------- | -------------------------------------------------------------------------------------- |
| endpoint        | The configuration object for the [Endpoint](#endpoint-configuration-object).           |
| outputs         | The `outputs` array from the output transformer. This array contains the Flow outputs. |
| processedOutput | The Flow output processed into the format that the Endpoint expects.                   |
| userId          | The unique ID of the user.                                                             |
| sessionId       | The unique ID of the session.                                                          |

### Endpoint Configuration Object

<Accordion title="endpoint Object Properties">
  | Property              | Type      | Description                                                                                                             |
  | --------------------- | --------- | ----------------------------------------------------------------------------------------------------------------------- |
  | `_id`                 | `string`  | The Endpoint's Reference ID.                                                                                            |
  | `channel`             | `string`  | The Endpoint's [channel](/ai/agents/develop/nodes/channels).                                                            |
  | `URLToken`            | `string`  | The URL Token, found in the [Input object](/ai/agents/deploy/endpoints/inject-and-notify#how-to-use-inject-and-notify). |
  | `name`                | `string`  | The Endpoint's name.                                                                                                    |
  | `flowId`              | `string`  | The Reference ID of the specified Flow to which the Endpoint points to.                                                 |
  | `entrypoint`          | `string`  | The Entrypoint of the Endpoint.                                                                                         |
  | `active`              | `boolean` | The flag to activate or deactivate the Endpoint.                                                                        |
  | `nluConnectorId`      | `string`  | ID of the selected NLU connector for the Endpoint.                                                                      |
  | `useAnalytics`        | `boolean` | The flag to collect analytics data for the Endpoint.                                                                    |
  | `storeDataPayload`    | `boolean` | The flag to store data payloads into analytics for the Endpoint.                                                        |
  | `useConversations`    | `boolean` | The flag to collect conversations history for the Endpoint.                                                             |
  | `maskIPAddress`       | `boolean` | The flag to mask sensitive IP addresses in input object and analytics data.                                             |
  | `maskAnalytics`       | `boolean` | The flag to mask sensitive data in analytics for the Endpoint.                                                          |
  | `maskLogging`         | `boolean` | The flag to mask sensitive data in logs for the Endpoint.                                                               |
  | `useContactProfiles`  | `boolean` | The flag to use Contact Profiles for the Endpoint.                                                                      |
  | `useDashbotAnalytics` | `boolean` | The flag to use Dashbot for collecting analytics data.                                                                  |
  | `dashbotApikey`       | `string`  | The API key of the Dashbot bot for analytics collection.                                                                |
  | `dashbotPlatform`     | `string`  | The selected platform of the Dashbot bot for analytics collection.                                                      |
  | `settings`            | `object`  | Optional Endpoint-specific settings. For example, Facebook Page token.                                                  |
  | `handoverSettings`    | `object`  | Settings to configure a handover provider.                                                                              |
  | `createdAt`           | `number`  | Unix timestamp when the Endpoint was created.                                                                           |
  | `lastChanged`         | `number`  | Unix timestamp when the Endpoint was last modified.                                                                     |
  | `createdBy`           | `string`  | Email of the user who created the Endpoint.                                                                             |
  | `lastChangedBy`       | `string`  | Email of the user who last modified the Endpoint.                                                                       |
</Accordion>

## Return Values

The return value of the execution finished transformer depends on the Endpoint type you are using. There is no validation of the execution finished transformer return value. If the execution finished transformer returns a falsy value, the output isn't sent to the Endpoint.

The execution finished transformer takes the `outputs` array from the output transformer and sends it to the Endpoint. For example, if you're using the execution finished transformer in a [REST Endpoint](/ai/agents/deploy/endpoint-reference/rest), the return value format must meet the REST Endpoint response format (`text`, `data`, `userId`, `sessionId`).

Here is an example of the correct return format for a REST Endpoint:

```js theme={null}
handleExecutionFinished: async ({ processedOutput, outputs, userId, sessionId, endpoint, response }) => {

    const text = outputs.reduce((processedOutput, output) => `${processedOutput}. ${output.text}`, "").trim();
    const data = outputs.reduce((acc, output) => ({ ...acc, ...output.data }), {});

    return {
        text,
        data,
        userId,
        sessionId
    };
}
```


# Inject Transformer
Source: https://docs.cognigy.com/ai/for-developers/transformers/inject-transformer



<Warning>
  In Cognigy 2025.25.0, transformers will start using the `axios` library for handling HTTP requests internally instead of `request-promise`. This change occurs to improve security and maintainability. The following HTTP request options will still be supported:

  * **URL configuration** — `uri`, `url`
  * **Request method** — `method`
  * **Request headers and configuration** — `headers`
  * **Request body** — `form`, `json`
  * **Query parameters** — `qs`, `qsStringifyOptions`
  * **Authentication** — `auth`
  * **Connection and network** — `proxy`, `maxRedirects`, `gzip`, `encoding`

  If you use other HTTP request options, review your transformer code to ensure it is compatible with the `axios` library before the release of Cognigy 2025.25.0.
</Warning>

The *Inject Transformer* converts the webhook payload from an external service before you call an [Inject API request](/ai/agents/deploy/endpoints/inject-and-notify), which sends the data to the Flow. With the inject transformer, you don't need a service that translates the webhook payload of the external service into the correct format for the Inject API request. You can handle this conversion in the inject transformer by parsing the payload to return only the necessary values.

You can configure the inject transformer in the `handleInject` function in the [Endpoint settings or via CLI](/ai/for-developers/transformers/overview#working-with-transformers).

## Restrictions

* The inject transformer is supported only for webhook- and socket-based Endpoints.

- The return value is validated against a set of rules. If these rules aren't met, the transformer throws an error. The following rules apply:

  * `userId` is a string with up to 256 characters.
  * `sessionId` is a string with up to 256 characters.
  * `text` is a string with up to 10,000 characters.
  * `data` is an object.

## Transformer Function Arguments

The following table shows an overview of the function arguments:

| Argument | Description                                                                  |
| -------- | ---------------------------------------------------------------------------- |
| endpoint | The configuration object for the [Endpoint](#endpoint-configuration-object). |
| request  | The Express request object with a JSON-parsed body.                          |
| response | The Express response object.                                                 |

### Endpoint Configuration Object

<Accordion title="endpoint Object Properties">
  | Property              | Type      | Description                                                                                                             |
  | --------------------- | --------- | ----------------------------------------------------------------------------------------------------------------------- |
  | `_id`                 | `string`  | The Endpoint's Reference ID.                                                                                            |
  | `channel`             | `string`  | The Endpoint's [channel](/ai/agents/develop/nodes/channels).                                                            |
  | `URLToken`            | `string`  | The URL Token, found in the [Input object](/ai/agents/deploy/endpoints/inject-and-notify#how-to-use-inject-and-notify). |
  | `name`                | `string`  | The Endpoint's name.                                                                                                    |
  | `flowId`              | `string`  | The Reference ID of the specified Flow to which the Endpoint points to.                                                 |
  | `entrypoint`          | `string`  | The Entrypoint of the Endpoint.                                                                                         |
  | `active`              | `boolean` | The flag to activate or deactivate the Endpoint.                                                                        |
  | `nluConnectorId`      | `string`  | ID of the selected NLU connector for the Endpoint.                                                                      |
  | `useAnalytics`        | `boolean` | The flag to collect analytics data for the Endpoint.                                                                    |
  | `storeDataPayload`    | `boolean` | The flag to store data payloads into analytics for the Endpoint.                                                        |
  | `useConversations`    | `boolean` | The flag to collect conversations history for the Endpoint.                                                             |
  | `maskIPAddress`       | `boolean` | The flag to mask sensitive IP addresses in input object and analytics data.                                             |
  | `maskAnalytics`       | `boolean` | The flag to mask sensitive data in analytics for the Endpoint.                                                          |
  | `maskLogging`         | `boolean` | The flag to mask sensitive data in logs for the Endpoint.                                                               |
  | `useContactProfiles`  | `boolean` | The flag to use Contact Profiles for the Endpoint.                                                                      |
  | `useDashbotAnalytics` | `boolean` | The flag to use Dashbot for collecting analytics data.                                                                  |
  | `dashbotApikey`       | `string`  | The API key of the Dashbot bot for analytics collection.                                                                |
  | `dashbotPlatform`     | `string`  | The selected platform of the Dashbot bot for analytics collection.                                                      |
  | `settings`            | `object`  | Optional Endpoint-specific settings. For example, Facebook Page token.                                                  |
  | `handoverSettings`    | `object`  | Settings to configure a handover provider.                                                                              |
  | `createdAt`           | `number`  | Unix timestamp when the Endpoint was created.                                                                           |
  | `lastChanged`         | `number`  | Unix timestamp when the Endpoint was last modified.                                                                     |
  | `createdBy`           | `string`  | Email of the user who created the Endpoint.                                                                             |
  | `lastChangedBy`       | `string`  | Email of the user who last modified the Endpoint.                                                                       |
</Accordion>

## Return Values

The inject transformer returns a valid user ID, session ID, text, and data. These return values need to match the format of the payload of the Endpoint you are using. The payload format is specific to the Endpoint type you use, for example, Facebook (Meta) Messenger. Read the documentation of the [specific Endpoint](/ai/agents/deploy/endpoint-reference/overview) to know how to format the payload.

If the inject transformer returns a falsy value, the output is discarded and not sent to the Inject API request.

Here is an example of the correct return format for a webhook-based Endpoint:

```js theme={null}
handleInject: async ({ endpoint, request, response }) => {
    const webhookPayload = request.body;
    
    // Parse the webhook payload and extract necessary data
    const userId = webhookPayload.user.id;
    const sessionId = webhookPayload.session.id;
    const text = webhookPayload.message.text;
    const data = {
        source: webhookPayload.source,
        timestamp: webhookPayload.timestamp
    };
    
    return {
        userId,
        sessionId,
        text,
        data
    };
}
```


# Input Transformer
Source: https://docs.cognigy.com/ai/for-developers/transformers/input-transformer



<Warning>
  In Cognigy 2025.25.0, transformers will start using the `axios` library for handling HTTP requests internally instead of `request-promise`. This change occurs to improve security and maintainability. The following HTTP request options will still be supported:

  * **URL configuration** — `uri`, `url`
  * **Request method** — `method`
  * **Request headers and configuration** — `headers`
  * **Request body** — `form`, `json`
  * **Query parameters** — `qs`, `qsStringifyOptions`
  * **Authentication** — `auth`
  * **Connection and network** — `proxy`, `maxRedirects`, `gzip`, `encoding`

  If you use other HTTP request options, review your transformer code to ensure it is compatible with the `axios` library before the release of Cognigy 2025.25.0.
</Warning>

The *Input Transformer* converts the user input after it's sent to an Endpoint. This allows you to manipulate the user input data before it is sent to the Flow, for example, to communicate with external systems, to implement integrations with a custom channel, and much more.

You can configure the input transformer in the `handleInput` function in the [Endpoint settings or via CLI](/ai/for-developers/transformers/overview#working-with-transformers).

## Restrictions

* The return value is validated against a set of rules. If these rules aren't met, the transformer throws an error. The following rules apply:

  * `userId` is a string with up to 256 characters.
  * `sessionId` is a string with up to 256 characters.
  * `text` is a string with up to 10,000 characters.
  * `data` is an object.

## Transformer Function Arguments

Depending on the Endpoint type in which you want to use the transformer, the function arguments vary:

<Tabs>
  <Tab title="Webhook- and REST-based Endpoints">
    | Argument | Description                                                                  |
    | -------- | ---------------------------------------------------------------------------- |
    | endpoint | The configuration object for the [Endpoint](#endpoint-configuration-object). |
    | request  | The Express request object with a JSON-parsed body.                          |
    | response | The Express response object.                                                 |
  </Tab>

  <Tab title="Socket-based Endpoints">
    | Argument | Description                                                                                                                                              |
    | -------- | -------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | endpoint | The configuration object for the [Endpoint](#endpoint-configuration-object).                                                                             |
    | payload  | The payload object contains the userId, sessionId, text, and data that was sent through the socket. The payload also contains the channel of the client. |
  </Tab>
</Tabs>

### Endpoint Configuration Object

<Accordion title="endpoint Object Properties">
  | Property              | Type      | Description                                                                                                             |
  | --------------------- | --------- | ----------------------------------------------------------------------------------------------------------------------- |
  | `_id`                 | `string`  | The Endpoint's Reference ID.                                                                                            |
  | `channel`             | `string`  | The Endpoint's [channel](/ai/agents/develop/nodes/channels).                                                            |
  | `URLToken`            | `string`  | The URL Token, found in the [Input object](/ai/agents/deploy/endpoints/inject-and-notify#how-to-use-inject-and-notify). |
  | `name`                | `string`  | The Endpoint's name.                                                                                                    |
  | `flowId`              | `string`  | The Reference ID of the specified Flow to which the Endpoint points to.                                                 |
  | `entrypoint`          | `string`  | The Entrypoint of the Endpoint.                                                                                         |
  | `active`              | `boolean` | The flag to activate or deactivate the Endpoint.                                                                        |
  | `nluConnectorId`      | `string`  | ID of the selected NLU connector for the Endpoint.                                                                      |
  | `useAnalytics`        | `boolean` | The flag to collect analytics data for the Endpoint.                                                                    |
  | `storeDataPayload`    | `boolean` | The flag to store data payloads into analytics for the Endpoint.                                                        |
  | `useConversations`    | `boolean` | The flag to collect conversations history for the Endpoint.                                                             |
  | `maskIPAddress`       | `boolean` | The flag to mask sensitive IP addresses in input object and analytics data.                                             |
  | `maskAnalytics`       | `boolean` | The flag to mask sensitive data in analytics for the Endpoint.                                                          |
  | `maskLogging`         | `boolean` | The flag to mask sensitive data in logs for the Endpoint.                                                               |
  | `useContactProfiles`  | `boolean` | The flag to use Contact Profiles for the Endpoint.                                                                      |
  | `useDashbotAnalytics` | `boolean` | The flag to use Dashbot for collecting analytics data.                                                                  |
  | `dashbotApikey`       | `string`  | The API key of the Dashbot bot for analytics collection.                                                                |
  | `dashbotPlatform`     | `string`  | The selected platform of the Dashbot bot for analytics collection.                                                      |
  | `settings`            | `object`  | Optional Endpoint-specific settings. For example, Facebook Page token.                                                  |
  | `handoverSettings`    | `object`  | Settings to configure a handover provider.                                                                              |
  | `createdAt`           | `number`  | Unix timestamp when the Endpoint was created.                                                                           |
  | `lastChanged`         | `number`  | Unix timestamp when the Endpoint was last modified.                                                                     |
  | `createdBy`           | `string`  | Email of the user who created the Endpoint.                                                                             |
  | `lastChangedBy`       | `string`  | Email of the user who last modified the Endpoint.                                                                       |
</Accordion>

## Return Values

The input transformer returns the user ID, session ID, text, and data, which are subsequently sent to the Flow. These values are extracted from the Endpoint payload. The payload format is specific to the Endpoint type you use, for example, Facebook (Meta) Messenger. Read the documentation of the [specific Endpoint](/ai/agents/deploy/endpoint-reference/overview) to know how to format the payload.

<Accordion title="Example of Transformer Return Values">
  ```js theme={null}
  return {
      userId: payload.userId,
      sessionId: payload.sessionId,
      text: payload.text,
      data: { "test": 1 }
  };
  ```
</Accordion>

### Undefined Return Values

If the transformer returns `undefined` for `userId`, `sessionId`, `text`, or `data`, this means that the transformer uses the value from the Endpoint.

<Accordion title="Example of Undefined Return Values">
  The following example overwrites `text` and `data`, but uses the `userId` and `sessionId` from the Endpoint:

  ```js theme={null}
  return {
      userId: undefined,
      sessionId: undefined,
      text: payload.text,
      data: { "test": 1 }
  };
  ```
</Accordion>

### Endpoints Event Messages

The Webchat and the Socket.IO Endpoints output event messages that indicate user activity, such as whether the user is connected (`user-connected`) or disconnected (`user-disconnected`). These event messages neither trigger Flows nor are counted. The event messages inform [handover providers](/ai/escalate/handovers) about user activity, allowing human agents to determine if the user is still engaged in the conversation. You can access the event messages with the `data._cognigy.event` property in the return value of the input transformer.

<Accordion title="Example of Return Values with Event Messages">
  ```js theme={null}
  return {
      "userId": payload.userId,
      "sessionId": payload.sessionId,
      "text": payload.text,
      "data": {
          "_cognigy": {
                "event": {
                  "type": "user-connected"
              }
          }
      }
  }
  ```
</Accordion>

If you want to access event messages in the input transformer, don't modify them. The event messages are only recognized if their data payload follows a specific format. The following code snippet shows how to use event messages in the input transformer:

<Accordion title="Example of Input Transformer with Event Messages">
  ```js theme={null}
  {
  handleInput: async ({ payload, endpoint }) => {
    if (!!payload.data?._cognigy?.event) {
  // pass on "event messages" without modification
        return payload;
      }

      // rest of your input transformer code
    }
  }
  ```
</Accordion>

### Transformers and Conversation Counts

Conversations in Cognigy are counted if the input transformer returns any result but falsy.


# Notify Transformer
Source: https://docs.cognigy.com/ai/for-developers/transformers/notify-transformer



<Warning>
  In Cognigy 2025.25.0, transformers will start using the `axios` library for handling HTTP requests internally instead of `request-promise`. This change occurs to improve security and maintainability. The following HTTP request options will still be supported:

  * **URL configuration** — `uri`, `url`
  * **Request method** — `method`
  * **Request headers and configuration** — `headers`
  * **Request body** — `form`, `json`
  * **Query parameters** — `qs`, `qsStringifyOptions`
  * **Authentication** — `auth`
  * **Connection and network** — `proxy`, `maxRedirects`, `gzip`, `encoding`

  If you use other HTTP request options, review your transformer code to ensure it is compatible with the `axios` library before the release of Cognigy 2025.25.0.
</Warning>

The *Notify Transformer* converts the webhook payload from an external service before you call a [Notify API request](/ai/agents/deploy/endpoints/inject-and-notify), which sends the data directly to the user. With the notify transformer, you don't need a service that translates the webhook payload of the external service into the correct format for the Notify API request. You can handle this conversion in the notify transformer by parsing the payload to return only the necessary values.

You can configure the notify transformer in the `handleNotify` function in the [Endpoint settings or via CLI](/ai/for-developers/transformers/overview#working-with-transformers).

## Restrictions

* The notify transformer is supported only for webhook- and socket-based Endpoints.

- The return value is validated against a set of rules. If these rules aren't met, the transformer throws an error. The following rules apply:

  * `userId` is a string with up to 256 characters.
  * `sessionId` is a string with up to 256 characters.
  * `text` is a string with up to 10,000 characters.
  * `data` is an object.

## Transformer Function Arguments

The following table shows an overview of the function arguments:

| Argument | Description                                                                  |
| -------- | ---------------------------------------------------------------------------- |
| endpoint | The configuration object for the [Endpoint](#endpoint-configuration-object). |
| request  | The Express request object with a JSON-parsed body.                          |
| response | The Express response object.                                                 |

### Endpoint Configuration Object

<Accordion title="endpoint Object Properties">
  | Property              | Type      | Description                                                                                                             |
  | --------------------- | --------- | ----------------------------------------------------------------------------------------------------------------------- |
  | `_id`                 | `string`  | The Endpoint's Reference ID.                                                                                            |
  | `channel`             | `string`  | The Endpoint's [channel](/ai/agents/develop/nodes/channels).                                                            |
  | `URLToken`            | `string`  | The URL Token, found in the [Input object](/ai/agents/deploy/endpoints/inject-and-notify#how-to-use-inject-and-notify). |
  | `name`                | `string`  | The Endpoint's name.                                                                                                    |
  | `flowId`              | `string`  | The Reference ID of the specified Flow to which the Endpoint points to.                                                 |
  | `entrypoint`          | `string`  | The Entrypoint of the Endpoint.                                                                                         |
  | `active`              | `boolean` | The flag to activate or deactivate the Endpoint.                                                                        |
  | `nluConnectorId`      | `string`  | ID of the selected NLU connector for the Endpoint.                                                                      |
  | `useAnalytics`        | `boolean` | The flag to collect analytics data for the Endpoint.                                                                    |
  | `storeDataPayload`    | `boolean` | The flag to store data payloads into analytics for the Endpoint.                                                        |
  | `useConversations`    | `boolean` | The flag to collect conversations history for the Endpoint.                                                             |
  | `maskIPAddress`       | `boolean` | The flag to mask sensitive IP addresses in input object and analytics data.                                             |
  | `maskAnalytics`       | `boolean` | The flag to mask sensitive data in analytics for the Endpoint.                                                          |
  | `maskLogging`         | `boolean` | The flag to mask sensitive data in logs for the Endpoint.                                                               |
  | `useContactProfiles`  | `boolean` | The flag to use Contact Profiles for the Endpoint.                                                                      |
  | `useDashbotAnalytics` | `boolean` | The flag to use Dashbot for collecting analytics data.                                                                  |
  | `dashbotApikey`       | `string`  | The API key of the Dashbot bot for analytics collection.                                                                |
  | `dashbotPlatform`     | `string`  | The selected platform of the Dashbot bot for analytics collection.                                                      |
  | `settings`            | `object`  | Optional Endpoint-specific settings. For example, Facebook Page token.                                                  |
  | `handoverSettings`    | `object`  | Settings to configure a handover provider.                                                                              |
  | `createdAt`           | `number`  | Unix timestamp when the Endpoint was created.                                                                           |
  | `lastChanged`         | `number`  | Unix timestamp when the Endpoint was last modified.                                                                     |
  | `createdBy`           | `string`  | Email of the user who created the Endpoint.                                                                             |
  | `lastChangedBy`       | `string`  | Email of the user who last modified the Endpoint.                                                                       |
</Accordion>

## Return Values

The notify transformer returns a valid user ID, session ID, text, and data. These return values need to match the format of the payload of the Endpoint you are using. The payload format is specific to the Endpoint type you use, for example, Facebook (Meta) Messenger. Read the documentation of the [specific Endpoint](/ai/agents/deploy/endpoint-reference/overview) to know how to format the payload.

If the notify transformer returns a falsy value, the output is discarded and not sent to the Notify API request.

Here is an example of the correct return format for a webhook-based Endpoint:

```js theme={null}
handleNotify: async ({ endpoint, request, response }) => {
    const webhookPayload = request.body;
    
    // Parse the webhook payload and extract necessary data
    const userId = webhookPayload.user.id;
    const sessionId = webhookPayload.session.id;
    const text = webhookPayload.message.text;
    const data = {
        source: webhookPayload.source,
        timestamp: webhookPayload.timestamp,
        priority: webhookPayload.priority || 'normal'
    };
    
    return {
        userId,
        sessionId,
        text,
        data
    };
}
```


# Output Transformer
Source: https://docs.cognigy.com/ai/for-developers/transformers/output-transformer



<Warning>
  In Cognigy 2025.25.0, transformers will start using the `axios` library for handling HTTP requests internally instead of `request-promise`. This change occurs to improve security and maintainability. The following HTTP request options will still be supported:

  * **URL configuration** — `uri`, `url`
  * **Request method** — `method`
  * **Request headers and configuration** — `headers`
  * **Request body** — `form`, `json`
  * **Query parameters** — `qs`, `qsStringifyOptions`
  * **Authentication** — `auth`
  * **Connection and network** — `proxy`, `maxRedirects`, `gzip`, `encoding`

  If you use other HTTP request options, review your transformer code to ensure it is compatible with the `axios` library before the release of Cognigy 2025.25.0.
</Warning>

The *Output Transformer* converts the Flow output. This transformer allows you to manipulate the Flow output before it's sent to the Endpoint.

You can configure the output transformer in the `handleOutput` function in the [Endpoint settings or via CLI](/ai/for-developers/transformers/overview#working-with-transformers).

The output transformer works differently for the following [Endpoint types](/ai/for-developers/transformers/overview#endpoint-types):

<Tabs>
  <Tab title="REST-based Endpoints">
    When using output transformers in REST Endpoints, you can't send multiple Flow outputs to the user. However, you can still have multiple Say Nodes in a Flow that are triggered by each user input.

    In this case, the `handleOutput` function stores the Say Node outputs in the `outputs` array. You can process the stored outputs with the [execution finished transformer](/ai/for-developers/transformers/execution-finished-transformer), which concatenates them into one output when the Flow execution has finished. This approach allows you to manipulate each Flow output in the `handleOutput` function and then concatenate them into one output with the execution finished transformer.

    ```mermaid theme={null}
    graph LR
    A[Flow output 1] --> B[Output Transformer]
    C[Flow output 2] --> B[Output Transformer]
    D[Flow output 3] --> B[Output Transformer]
    B --> E[outputs Array]
    E --> F[Execution Finished Transformer]
    F -- concatenated messages --> G[Endpoint]
    G --> H[User]
    ```
  </Tab>

  <Tab title="Webhook- and Socket-based Endpoints">
    When using output transformers in Webhook and Socket Endpoints, each Flow output is sent to the user. The return value of the `handleOutput` function is sent directly to the user and must comply with the [channel](/ai/agents/develop/nodes/channels) format.

    In this case, the `handleOutput` function takes an additional argument: `processedOutput`. This argument contains the output that would be sent as the correct format corresponding to the specific channel.

    ```mermaid theme={null}
    graph LR
    A[Flow output 1] --> B[Output Transformer]
    C[Flow output 2] --> B[Output Transformer]
    D[Flow output 3] --> B[Output Transformer]
    B -- message 1 --> E[Endpoint]
    B -- message 2 --> E[Endpoint]
    B -- message 3 --> E[Endpoint]
    E -- message 1 --> F[User]
    E -- message 2 --> F[User]
    E -- message 3 --> F[User]
    ```
  </Tab>
</Tabs>

## Transformer Function Arguments

Depending on the Endpoint type you are using, the function arguments vary:

<Tabs>
  <Tab title="REST-based Endpoints">
    | Argument  | Description                                                                  |
    | --------- | ---------------------------------------------------------------------------- |
    | endpoint  | The configuration object for the [Endpoint](#endpoint-configuration-object). |
    | output    | The raw output from the Flow.                                                |
    | userId    | The unique ID of the user.                                                   |
    | sessionId | The unique ID of the session.                                                |
  </Tab>

  <Tab title="Webhook- and Socket-based Endpoints">
    | Argument        | Description                                                                                                                                                           |
    | --------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | endpoint        | The configuration object for the [Endpoint](#endpoint-configuration-object).                                                                                          |
    | output          | The raw output from the Flow.                                                                                                                                         |
    | processedOutput | The processed output from the Flow to comply with the Endpoint format. This is the data that the Endpoint sends to the user if the output transformer isn't executed. |
    | userId          | The unique ID of the user.                                                                                                                                            |
    | sessionId       | The unique ID of the session.                                                                                                                                         |
  </Tab>
</Tabs>

### Endpoint Configuration Object

<Accordion title="endpoint Object Properties">
  | Property              | Type      | Description                                                                                                             |
  | --------------------- | --------- | ----------------------------------------------------------------------------------------------------------------------- |
  | `_id`                 | `string`  | The Endpoint's Reference ID.                                                                                            |
  | `channel`             | `string`  | The Endpoint's [channel](/ai/agents/develop/nodes/channels).                                                            |
  | `URLToken`            | `string`  | The URL Token, found in the [Input object](/ai/agents/deploy/endpoints/inject-and-notify#how-to-use-inject-and-notify). |
  | `name`                | `string`  | The Endpoint's name.                                                                                                    |
  | `flowId`              | `string`  | The Reference ID of the specified Flow to which the Endpoint points to.                                                 |
  | `entrypoint`          | `string`  | The Entrypoint of the Endpoint.                                                                                         |
  | `active`              | `boolean` | The flag to activate or deactivate the Endpoint.                                                                        |
  | `nluConnectorId`      | `string`  | ID of the selected NLU connector for the Endpoint.                                                                      |
  | `useAnalytics`        | `boolean` | The flag to collect analytics data for the Endpoint.                                                                    |
  | `storeDataPayload`    | `boolean` | The flag to store data payloads into analytics for the Endpoint.                                                        |
  | `useConversations`    | `boolean` | The flag to collect conversations history for the Endpoint.                                                             |
  | `maskIPAddress`       | `boolean` | The flag to mask sensitive IP addresses in input object and analytics data.                                             |
  | `maskAnalytics`       | `boolean` | The flag to mask sensitive data in analytics for the Endpoint.                                                          |
  | `maskLogging`         | `boolean` | The flag to mask sensitive data in logs for the Endpoint.                                                               |
  | `useContactProfiles`  | `boolean` | The flag to use Contact Profiles for the Endpoint.                                                                      |
  | `useDashbotAnalytics` | `boolean` | The flag to use Dashbot for collecting analytics data.                                                                  |
  | `dashbotApikey`       | `string`  | The API key of the Dashbot bot for analytics collection.                                                                |
  | `dashbotPlatform`     | `string`  | The selected platform of the Dashbot bot for analytics collection.                                                      |
  | `settings`            | `object`  | Optional Endpoint-specific settings. For example, Facebook Page token.                                                  |
  | `handoverSettings`    | `object`  | Settings to configure a handover provider.                                                                              |
  | `createdAt`           | `number`  | Unix timestamp when the Endpoint was created.                                                                           |
  | `lastChanged`         | `number`  | Unix timestamp when the Endpoint was last modified.                                                                     |
  | `createdBy`           | `string`  | Email of the user who created the Endpoint.                                                                             |
  | `lastChangedBy`       | `string`  | Email of the user who last modified the Endpoint.                                                                       |
</Accordion>

## Return Values

The output transformer return value depends on the Endpoint type in which you use the transformer. There is no validation of the output transformer return value. If the output transformer returns a falsy value, the output is discarded and not stored in the `outputs` array.

<Tabs>
  <Tab title="REST-based Endpoints">
    The `handleOutput` function returns a value and stores it in the `outputs` array. You can access this array with the `handleExecutionFinished` function. You can set the format of the `outputs` array items in the `handleOutput` function. By default, the `outputs` array entries have `text` and `data` properties. You can change the output format as follows:

    <Accordion title="Modify the Return Value Format">
      ```js theme={null}
      handleOutput: async ({ processedOutput, output, endpoint, userId, sessionId }) => {
      const modifiedOutput = {
          text: "text",
          data: {}
          // ...
      };

      return modifiedOutput;
      }
      ```
    </Accordion>
  </Tab>

  <Tab title="Webhook- and Socket-based Endpoints">
    The output transformer must return a value that can be sent directly to the specific channel. For example, if you're using the output transformer in a Sunshine Conversations Endpoint, the return value format must meet the [Sunshine Conversations message format](https://developer.zendesk.com/documentation/conversations/messaging-platform/programmable-conversations/sending-messages/).

    <Accordion title="Format Example for Sunshine Conversations Endpoints">
      ```js theme={null}
      handleOutput: async ({ processedOutput, output, endpoint, userId, sessionId }) => {
          const requestPayload = {
              author: {
                  type: "business"
              },
              content: {
                  type: "text",
                  text: "Hello world"
              }
          };
          
          return requestPayload;
      }
      ```
    </Accordion>
  </Tab>
</Tabs>


# Transformers
Source: https://docs.cognigy.com/ai/for-developers/transformers/overview

Use Endpoint Transformers to manipulate data between Flows and Endpoints, convert payloads, and create custom integrations.

<a href="/release-notes/2026.17"><Badge>Updated in 2026.17</Badge></a>

<Warning>
  In Cognigy 2025.25.0, transformers will start using the `axios` library for handling HTTP requests internally instead of `request-promise`. This change occurs to improve security and maintainability. The following HTTP request options will still be supported:

  * **URL configuration** — `uri`, `url`
  * **Request method** — `method`
  * **Request headers and configuration** — `headers`
  * **Request body** — `form`, `json`
  * **Query parameters** — `qs`, `qsStringifyOptions`
  * **Authentication** — `auth`
  * **Connection and network** — `proxy`, `maxRedirects`, `gzip`, `encoding`

  If you use other HTTP request options, review your transformer code to ensure it is compatible with the `axios` library before the release of Cognigy 2025.25.0.
</Warning>

*Endpoint Transformers* are functions that convert the text and data sent between the [Endpoint](/ai/agents/deploy/endpoints/overview) and the [Flow](/ai/agents/develop/projects-and-flows/overview) as well as through the [Inject and Notify API requests](/ai/agents/deploy/endpoints/inject-and-notify). These functions allow you to manipulate the data from the user input, the Flow output, and third-party API services.

Transformers fully support TypeScript and JavaScript, including [npm modules](#npm-modules).

You can use transformers to:

* Convert request payloads from an Endpoint before sending them to a Flow, for example, to encrypt data.
* Convert the Flow output before sending it to an Endpoint.
* Translate messages from the user into the Flow language.
* Customize the default and custom Endpoint interfaces.
* Convert a channel output to be compatible with another channel. For example, convert a [Webchat output](/ai/agents/deploy/endpoint-reference/webchat) into a [WhatsApp-compatible output](/ai/agents/deploy/endpoint-reference/whatsapp).
* Create [custom Endpoints](#create-a-custom-endpoint) for additional integrations and handle webhook events from the custom Endpoints.

Depending on the [type of the Endpoint](#endpoint-types) in which you use the transformers, they might work differently.

To learn more about Endpoint transformers, watch this webinar:

<iframe title="YouTube video player" />

## Restrictions

* You can't send HTTP requests from Endpoint transformers to other Endpoints.
* Endpoint transformers operate only on data exchanged between the Endpoint and the Flow. They don't modify the prompt or response of an LLM Prompt Node or AI Agent Node.

## Limitations

* You can only send one HTTP request per execution from a transformer.
* HTTP requests from transformer functions have a 5-second timeout. When the request times out, Flow execution throws an error and logs it. You can change or remove the timeout:
  * For SaaS installations, contact [Cognigy technical support](/help/get-help).
  * For on-premises installations, change the timeout by setting it in milliseconds in `TRANSFORMER_FUNCTION_EXECUTION_TIMEOUT_MS` or remove it by setting `FEATURE_DISABLE_TRANSFORMER_HTTP_TIMEOUT` to `true`.

## Endpoint Types

There are three different types of Endpoints based on the underlying technology.

<Tabs>
  <Tab title="REST-based Endpoints">
    REST-based Endpoints receive messages from the user and send a single reply. For these Endpoints, you can add more than one Say Node in the Flow, but all the outputs are processed only after the Flow execution has finished. The outputs are concatenated into one message and then sent to the user. The REST-based Endpoints are:

    * [Alexa (deprecated)](/release-notes/deprecations-and-removals)
    * [Dialogflow (deprecated)](/release-notes/deprecations-and-removals)
    * Twilio
    * Twilio SMS
    * Microsoft Teams
    * Userlike
    * AudioCodes
    * REST
  </Tab>

  <Tab title="Webhook-based Endpoints">
    Webhook-based Endpoints receive user messages and reply after each input. For these Endpoints, you can process each output after the user input. The webhook-based Endpoints are:

    * Facebook (Meta) Messenger
    * Slack
    * Microsoft Bot Framework
    * Sunshine Conversations
    * LINE
    * RingCentral Engage
    * Intercom
    * WhatsApp
    * Webhook
  </Tab>

  <Tab title="Socket-based Endpoints">
    Socket-based Endpoints receive messages from the user and can reply after each input. For these Endpoints, you can process each output after the user input. The socket-based Endpoints are:

    * Webchat
    * Socket
  </Tab>
</Tabs>

## Transformer Functions

The transformer function determines the transformer you use in the [Endpoint settings](/ai/agents/deploy/endpoints/overview). The following table provides an overview of the transformer functions:

| Transformer Function    | Transformer                                                                          | Description                                                                                                                                                                                                                                                                                                        |
| ----------------------- | ------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| handleInput             | [Input](/ai/for-developers/transformers/input-transformer)                           | Executes when a user input is sent to an Endpoint, before the Flow is executed. The return value is sent to the Flow.                                                                                                                                                                                              |
| handleOutput            | [Output](/ai/for-developers/transformers/output-transformer)                         | Executes when a Flow sends an output to an Endpoint. For webhook- and socket-based Endpoints, the return value of each output is sent to the user. For REST-based Endpoints, the return value is stored and can be sent to the user after the Flow execution has finished with the execution finished transformer. |
| handleExecutionFinished | [Execution finished](/ai/for-developers/transformers/execution-finished-transformer) | Executes when the Flow execution has finished. The return value is sent to the user. In this case, all outputs are concatenated and sent to the user. Applies only to REST-based Endpoints in which you use an output transformer.                                                                                 |
| handleNotify            | [Notify](/ai/for-developers/transformers/notify-transformer)                         | Executes when you use the [Notify API request](/ai/agents/deploy/endpoints/inject-and-notify). The return value is sent directly to the user.                                                                                                                                                                      |
| handleInject            | [Inject](/ai/for-developers/transformers/inject-transformer)                         | Executes when you use the [Inject API request](/ai/agents/deploy/endpoints/inject-and-notify). The return value is sent to the Flow.                                                                                                                                                                               |

## Working with Transformers

<Tabs>
  <Tab title="GUI">
    You can configure the transformer functions in the **Transformer Functions** section of the Endpoint settings. You need to activate the transformer functions individually and provide the corresponding code in the **Transformer** editor.
  </Tab>

  <Tab title="API">
    You can configure transformer settings using the [Cognigy API](https://api-trial.cognigy.ai/openapi#post-/v2.0/endpoints).
  </Tab>

  <Tab title="CLI">
    You can configure the transformer functions locally and deploy them with the [Cognigy CLI](https://github.com/Cognigy/Cognigy-CLI). To do this, follow these steps:

    1. In **Deploy > Endpoints**, select an Endpoint and activate the [transformer functions](#transformer-settings) you want to use in the Endpoint settings.
    2. [Clone your AI Agent](https://github.com/Cognigy/Cognigy-CLI?tab=readme-ov-file#command-clone) to your local machine.
    3. Make the changes to the transformer functions in the `transformers.ts` file in your Endpoint's folder, for example, `<root-directory>/agent/endpoints/Webchat/transformer.ts`, where Webchat is the name of your Endpoint.
    4. [Restore the AI Agent](https://github.com/Cognigy/Cognigy-CLI?tab=readme-ov-file#command-restore) to Cognigy.
  </Tab>
</Tabs>

### Transformer Settings

| Setting                               | Type        | Description                                                                                                      |
| ------------------------------------- | ----------- | ---------------------------------------------------------------------------------------------------------------- |
| Abort on Error in Transformer         | Toggle      | Terminates transformer execution when an error occurs during execution.                                          |
| Enable Input Transformer              | Toggle      | Activates the Input transformer function.                                                                        |
| Enable Output Transformer             | Toggle      | Activates the Output transformer function.                                                                       |
| Enable Execution Finished Transformer | Toggle      | Activates the Execution Finished transformer function.                                                           |
| Enable Notify Transformer             | Toggle      | Activates the Notify transformer function.                                                                       |
| Enable Inject Transformer             | Toggle      | Activates the Inject transformer function.                                                                       |
| Transformer                           | Code editor | Includes the code for each transformer function. By default, includes one instance of each transformer function. |

### Available JavaScript Modules

#### npm Modules

The following npm modules are available for use in a transformer:

* `uuid` — for generating UUIDs.
* `crypto` — for encrypting data.
* `_` (Lodash) — for manipulating arrays, objects, and strings.
* `moment` — for manipulating dates and times.
* `xmljs` — for manipulating XML data.

#### Text Cleaner Class

Transformers can use the `getTextCleaner` function to retrieve an instance of the [Text Cleaner](/ai/platform-features/nlu/text-cleaner) class, which allows you to preprocess user input.

### Modify the User ID in Transformers

Endpoint transformers can modify the [user ID](/ai/agents/analyze/user-and-session-identification). If a transformer changes the user ID, Cognigy uses the updated value consistently throughout the entire processing pipeline, including:

* Flow execution
* Notify API requests
* xApps
* Live Follow
* Logs

To avoid exposing personally identifiable information, for example, phone numbers or email addresses, and to support GDPR compliance, hash the user ID.

When implementing user ID transformation:

* Use a deterministic method: the same input must always produce the same output.
* Use simple string prefixing, such as `"hashed_" + userId`, only for testing or development. Prefixing doesn't provide real security. In production, use a cryptographic hash function, such as SHA-256.

<Accordion title="Example: Hashing the User ID (SHA-256)">
  ```javascript id="9k2lpx" theme={null}
  const crypto = require("crypto");

  function hashUserId(userId) {
      return crypto
          .createHash("sha256")
          .update(userId)
          .digest("hex");
  }

  handleInput: async ({ payload }) => {
      return {
          userId: hashUserId(payload.userId),
          sessionId: payload.sessionId,
          text: payload.text,
          data: payload.data
      };
  };
  ```
</Accordion>

### Send HTTP Requests

You can send HTTP requests to external services from a transformer. All transformer functions can access the `httpRequest` module. Call the module as a function that takes a single configuration object as argument. You can use the following formats for the configuration object:

<Tabs>
  <Tab title="Axios">
    The configuration object follows the format described in the [Axios Request Config documentation](https://axios-http.com/docs/req_config). For example:

    ```js theme={null}
    handleInput: async ({ payload }) => {
      const result = await httpRequest({
        url: "<url>",
        method: "POST",
        data: {
          userId: payload.userId,
          text: payload.text
        }
      });

      console.log(JSON.stringify(result)); // Registers results on the Logs page
      return {
        userId: payload.userId,
        sessionId: payload.sessionId,
        text: payload.text,
        data: payload.data
      };
    },
    ```
  </Tab>

  <Tab title="request-promise (deprecated)">
    <Warning>
      The `request-promise` library is deprecated. Review your transformer code to ensure it is compatible with the `axios` library.
    </Warning>

    The configuration object follows the format described in the [npm module documentation](https://www.npmjs.com/package/request-promise). For example:

    ```js theme={null}
    handleInput: async ({ payload }) => {
      const result = await httpRequest({
        uri: "<uri>",
        method: "POST",
        body: {},
        json: true,
      });

      console.log(JSON.stringify(result)); // Registers results on the Logs page

      return {
        userId: payload.userId,
        sessionId: payload.sessionId,
        text: payload.text,
        data: payload.data
      };
    },
    ```
  </Tab>
</Tabs>

### Session Storage

The session storage is a JSON object available to all transformer functions that persists throughout a session. You can use the session storage to store variables for later use in other transformer functions or to store the conversation state. To access the stored data, use the `getSessionStorage` function inside the transformer functions. The `getSessionStorage` function takes the `userId` and `sessionId` as arguments and returns a promise. The returned promise resolves with the `sessionStorage` object, which can be manipulated as a regular object.

#### Limitations

* Session storages are limited to the scope of an Endpoint. You can't access a session storage from transformer functions across Endpoints.
* Session storages don't support mutating complex objects, such as arrays, directly. Mutate them locally, then reassign them to the session storage object, as in [the following example](#example).

#### Example

Here is an example of session storage usage:

```js theme={null}
const handleOutput: async ({ userId, sessionId }) => {
    const sessionStorage = await getSessionStorage(userId, sessionId);

    sessionStorage.test = "test";

    // Mutating complex objects
    // Instead of changing the objects directly inside sessionStorage,
    // change them locally and then assign them to the sessionStorage
    let tempArr = sessionStorage.arr || [];
    tempArr.push(1);
    tempArr.push(2);
    sessionStorage.arr = tempArr;
}
```

### Stop Transformer Execution

Every transformer function expects a certain return value that is further processed in an Endpoint or Flow. In certain cases, you might want to stop the execution completely, for example, if the output from a Flow is forwarded to another platform. To stop the transformer execution, set the transformer function to return a falsy value, such as `null` or `void`.

### Error Handling in Transformers

By default, transformers log errors on the [Logs](/ai/agents/test/logs) page and continue with the Flow execution. This means that if an error occurs during the transformer execution, including [timeout](#limitations), the user input is still processed.

If the transformer is essential to the pipeline, for example, when building a custom Endpoint integration, you might want to prevent the Flow execution after an error in the transformer execution. To do this, activate the **Abort on Error in Transformer** setting in the [transformer settings](#transformer-settings).

Transformer error log entries include `traceId` and `sessionId` in their metadata when these values are available. You can use them to search the [Logs](/ai/agents/test/logs) page and correlate a failed transformer execution with a specific request and session.

<Warning>
  * When an error occurs during the transformer execution, the logs list the line number where the error occurred. The line number might be incorrect because the logs don't account for lines used for interfaces and comments.
  * Errors caused by a return value validation error don't show a line number in the logs.
</Warning>

### Log Transformer Events

With the following methods, you can log transformer events on the [Logs page](/ai/agents/test/logs):

* `console.log` — as a `debug` log entry.
* `console.info` — as an `info` log entry.
* `console.error` — as an `error` log entry.

### Create a Custom Endpoint

With transformers, you can create an integration with a custom Endpoint that Cognigy doesn't support by default. To create a custom Endpoint, you need to know the Endpoint type of the channel. Then, you can use a transformer to call the custom Endpoint service.

## Transformer Examples

You can find examples of various transformers in the [Cognigy GitHub repository](https://github.com/Cognigy/TransformersSamples).


# About the Cognigy Platform
Source: https://docs.cognigy.com/ai/overview/about-cognigy-ai

Discover how Cognigy orchestrates AI-powered customer experiences across voice and chat channels.

*The Cognigy platform* is an AI customer experience orchestration platform that integrates with any contact center or CRM system. It enhances existing infrastructure with advanced conversational AI capabilities, streamlining customer interaction. Cognigy empowers organizations to deliver personalized, seamless, and efficient customer experiences on all channels, managing customer journeys intelligently for maximum engagement and satisfaction.

<Frame>
  <img alt="Cognigy platform products overview" />
</Frame>

## Cognigy Platform Products

### [Cognigy](/ai/overview/ai-workforce)

The Cognigy interface provides conversation designers with a range of powerful tools and features
to create and manage voice and chat assistants.
It serves as an Integrated Development Environment (IDE)
that simplifies the development process and enables designers to create engaging user experiences.
The interface helps conversation designers build and refine conversational flows, test and analyze user interactions,
and quickly deploy voice and chat assistants across multiple channels.

### [Voice Gateway](/voice-gateway/index)

Bridges telephony systems with Cognigy, enabling natural voice interactions and integration with IVR systems for enhanced customer service.
This product has a separate installation and is compatible only with Cognigy.

### [xApps](/xApps/index)

Cognigy's xApps are pre-built application components that can be easily integrated into conversational flows. These xApps extend the functionality of the platform, enabling quick deployment of features like payment processing, appointment booking, and more, without the need for custom development.
This product is provided together with Cognigy.

### [Live Agent](/live-agent/index)

Facilitates collaboration between human agents and virtual assistants, enabling smooth escalation of customer inquiries for efficient support.
This product has a separate installation and is compatible only with Cognigy.

### [Webchat](/webchat/index)

Customizable chat widget for real-time customer engagement on websites, providing intelligent and personalized assistance to drive conversion rates.
This product is free of charge.

### [Insights](/insights/index)

Offers advanced analytics and reporting within Cognigy, enabling data-driven decision-making to optimize conversational AI strategies and improve customer engagement.
This product is provided together with Cognigy.

### [Agent Copilot](/agent-copilot/index)

The Agent Copilot feature acts as an assistant to both customers and service agents, offering suggestions, automating routine tasks, and facilitating smoother interactions. This not only enhances the user experience but also supports agents in delivering high-quality service.
This product is provided together with Cognigy.

## [Cognigy Installations](/ai/administer/installation/about)

Cognigy installations are offered as Cognigy SaaS: Cognigy operates on and maintains cloud-hosted instances for our customers.


# Cognigy AI Agents
Source: https://docs.cognigy.com/ai/overview/ai-workforce



*AI Agents* are virtual assistants powered by artificial intelligence that can interact with users through text or voice, handling tasks traditionally performed by humans.

*Cognigy AI Workforce* is a collection of AI Agents that you can design using the Cognigy platform to automate tasks and enhance various business processes.
Cognigy AI Agents can handle a wide range of tasks, from answering simple customer inquiries to providing complex technical support.

Key benefits of an AI Workforce include:

* **24/7 Availability**. AI Agents are always available, ensuring continuous customer support.
* **Scalability**. AI Agents can handle increased workloads without compromising quality.
* **Consistency**. AI Agents provide consistent and accurate responses.
* **Efficiency**. Automating routine tasks reduces operational costs and frees up human resources.

Cognigy AI Workforce goes beyond simply offering virtual assistants.
It offers intelligent AI Agents to support both end users and human agents:

* [Self-Service AI Agents](#self-service-ai-agents)
* [Human Agent Assistants](#human-agent-assistants)

### Self-Service AI Agents

Cognigy enables the creation of autonomous AI Agents that can manage customer inquiries, resolve common issues, and guide users through processes without human intervention.

Using the tools on the Cognigy platform, you can create AI Agents customized to meet the specific needs of your business.
Here are some examples of AI Agents you can build with Cognigy:

* [Chat-Based AI Agents](#chat-based-ai-agents)
* [Voice AI Agents](#voice-ai-agents)
* [Digital AI Agents](#digital-ai-agents)
* [LLM-Driven AI Agents](#llm-driven-ai-agents)

#### Chat-Based AI Agents

Chat-based AI Agents specialize in text-based interactions, primarily used for customer support, answering FAQs, and handling tasks, including booking appointments and providing product information through chat platforms, websites, and messaging apps.

To build a chat AI Agent using Cognigy, you need to start by creating a [Project](/ai/agents/develop/projects-and-flows/projects) and a [Flow](/ai/agents/develop/projects-and-flows/overview).
Once created, you can add at least [basic Nodes](/ai/agents/develop/node-reference/overview) such as Say and Question Nodes,
which form the building blocks of the AI Agent's conversational flow.
After that, you need to choose an appropriate Endpoint for the channel you want to deploy the AI Agent on.
For example, select the [Webchat v3 Endpoint](/webchat/v3/configuration) if you want to connect the AI Agent to the [Cognigy Webchat v3 widget](/webchat/v3/overview), the [WhatsApp Endpoint](/ai/agents/deploy/endpoint-reference/whatsapp) to connect it to WhatsApp Messenger, or the [Slack Endpoint](/ai/agents/deploy/endpoint-reference/slack) to connect it to the Slack platform. To choose the best one for your needs, refer to the [full list of available Endpoints](/ai/agents/deploy/endpoint-reference/overview).

#### Voice AI Agents

Voice AI Agents handle spoken conversations; they use speech-to-text recognition and text-to-speech technology, which makes them ideal for phone-based customer support.

To build a voice AI Agent using Cognigy,
you need to start by creating a [Project](/ai/agents/develop/projects-and-flows/projects) and a [Flow](/ai/agents/develop/projects-and-flows/overview).
However, in addition to the [basic Nodes](/ai/agents/develop/node-reference/overview),
you also need
to add [voice Nodes](/ai/agents/develop/node-reference/voice/voice-gateway/overview)
that allow the AI Agent to interact with users through voice commands.
After that, you need to select a suitable voice Endpoint for the channel where you want to deploy the AI Agent.
For seamless connectivity, we recommend using the [Voice Gateway](/ai/agents/deploy/endpoint-reference/voice-gateway) Endpoint.
To enable voice interactions,
you'll require credentials from a speech provider and a Session Initiation Protocol (SIP) trunk.
Once you have configured the SIP Trunk and Speech provider on the Voice Gateway Self-Service Portal,
you can deploy your voice AI Agent in production.
For more information, read the [Voice Gateway](/voice-gateway/getting-started) documentation.

#### Digital AI Agents

By integrating text and voice capabilities,
you can develop sophisticated digital AI Agents that engage customers across multiple channels,
providing omnichannel support on platforms such as websites, apps, and social media.
For instance, an AI Agent may start a conversation through text and then suggest transitioning to a call with the contact center if needed.
To enhance the capability of your AI Agent,
Cognigy offers a [marketplace of prebuilt Extensions](/ai/for-developers/extensions) and [Extension framework](/ai/for-developers/developers/extension-framework)
to connect AI Agents with external systems, such as CRM and task management tools, or to build extensions from scratch.

#### LLM-Driven AI Agents

By using advanced AI technologies like [Generative AI](/ai/agents/develop/gen-ai-and-llms/generative-ai), [Natural Language Understanding (NLU)](/ai/platform-features/nlu/overview), and [Knowledge AI](/ai/agents/develop/knowledge-ai/overview),
you can empower AI Agents — whether chat, voice, or digital — to understand complex requests,
retrieve information from multiple sources, and deliver personalized responses that enhance user experiences.

Also, you can make your AI Agent [fully driven by Large Language Models (LLMs)](/ai/agents/overview), tailoring its personality to align with your brand and customer preferences.

### Human Agent Assistants

Cognigy AI Agents can assist human agents by offering real-time support, accessing knowledge bases, and analyzing customer sentiment.
By assisting human agents, AI Agents improve efficiency, speed up resolution times, and enhance customer service quality.

Cognigy offers a comprehensive suite of human assistants, each tailored to meet specific needs in your contact center solution:

* [Agent Copilot](#agent-copilot)
* [Agent Copilot Whisper](#agent-copilot-whisper)
* [Agent Copilot Bot](#agent-copilot-bot)

#### Agent Copilot

[Cognigy Agent Copilot](/agent-copilot/index) is an intelligent assistant that helps human agents work more effectively.
It provides real-time suggestions, handles routine tasks, and manages extra workload during busy times.
With an easy-to-use interface and customizable widgets,
such as tools for quick knowledge searches or sentiment analysis, it offers human agents fast access to key information.
By boosting the productivity of each human agent, Agent Copilot helps address labor shortages and combines AI and human strengths to deliver better customer service.

#### Agent Copilot Whisper

[Cognigy Agent Copilot Whisper](/live-agent/assistants/agent-copilot-whisper) is a real-time AI assistant that provides direct, in-conversation support to human agents. It listens to the ongoing conversation and offers suggestions, recommendations, and relevant information directly within the chat interface, empowering agents to provide faster and more accurate responses.

Agent Copilot Whisper is available only in the Cognigy Live Agent interface.

#### Agent Copilot Bot

[Cognigy Agent Copilot Bot](/live-agent/assistants/agent-copilot-bot) is a dedicated chat-based assistant designed to support human agents.
It provides a personalized workflow, enabling AI Agents to access additional information and resources as needed.

Agent Copilot Bot is available only in the Cognigy Live Agent interface.

## More Information

* [Key Features](/ai/overview/key-features)
* [Build Your AI Agent in Minutes](/ai/agents/quickstart)


# Cognigy Glossary
Source: https://docs.cognigy.com/ai/overview/glossary



Cognigy terms listed in alphabetical order. All terms in the glossary are given for informational purposes.

## A

**Agent Copilot**

A Cognigy product that offers a variety of features in a workspace and empowers human agents to provide faster and more accurate customer support. You can customize the Agent Copilot workspace with widgets that display the most relevant information for chat and voice use cases. When embedded into your contact center's desktop, Agent Copilot enhances the overall efficiency of your contact center. For more information, read the [Agent Copilot](/agent-copilot/overview) documentation.

**AI Agent**

A Cognigy bot that autonomously handles calls and chats from customers and employees. AI Agents use large language models (LLMs) and natural language understanding (NLU) to understand user inputs and search inquiries. For more information, read the [AI Agents](/ai/agents/overview) documentation.

## C

**Channel**

A Cognigy resource that allows you to interact with the AI Agent through multiple interfaces. Each channel is associated with one or more Cognigy [Endpoints](#e). Channels include Webchat, Facebook Messenger, Slack, WhatsApp, and other messaging or communication platforms. For more information, read the [Channels](/ai/agents/develop/nodes/channels) documentation.

**Child Node**

A Node that is part of the logic of another Node. Child Nodes can perform actions, for example, [AI Agent tools](/ai/agents/develop/node-reference/ai/ai-agent#tool), or represent conditional paths, for example, the Then and Else Nodes of an [If Node](/ai/agents/develop/node-reference/logic/if).

**Chunk Editor**

An interface in Cognigy for editing and managing [Knowledge Chunks](#k). You can edit the text, add new information, delete sections, or rearrange the order of the Knowledge Chunks to ensure their accuracy and relevance. For more information, read the [Knowledge AI](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk-editor) documentation.

**Client Libraries**

Applications developed in different programming languages that connect with Cognigy, hosted on Cognigy's GitHub pages. For example, the REST API client allows direct interaction with Cognigy resources using TypeScript or JavaScript. For more information, read the [Client Libraries](/ai/for-developers/developers/client-libraries) documentation.

**Cognigy**

Cognigy's conversational AI platform, designed for enterprises to create, deploy, and orchestrate chat and voice AI Agents at scale. Cognigy leverages advanced natural language understanding (NLU) technology and LLMs to develop conversational interfaces that can accurately understand and interpret user queries. For more information, read [About the Cognigy Platform](/ai/overview/about-cognigy-ai).

**Cognigy API**

An HTTP-based application programming interface (API) provided by Cognigy that allows communication with the Cognigy installation. With the Cognigy API, developers can access various functionalities and resources in Cognigy, control the behavior of AI Agents, and develop custom integrations. For more information, read the [API and CLI](/ai/for-developers/developers/api-and-cli) article and the [Cognigy API](https://api-trial.cognigy.ai/openapi) documentation.

**Cognigy NLU**

A natural language understanding engine that powers AI Agents in Cognigy. Cognigy NLU allows AI Agents to understand and interpret human language and provide accurate responses. For more information, read the [NLU Overview](/ai/platform-features/nlu/overview) documentation.

**Cognigy Products**

Licensed applications and services that Cognigy offers, such as Cognigy, Cognigy Voice Gateway, Cognigy Live Agent, Cognigy Insights, and others. Cognigy Products are listed in the Cognigy order form. For more information, select a [product line](/index) and explore all features and platform capabilities.

**Cognigy SaaS**

Cognigy's delivery model, which operates on and maintains cloud-hosted instances for customers.

**CognigyScript**

A JavaScript superset developed for Cognigy. CognigyScript lets you configure AI Agents to access and manipulate the Input, Context, and Profile objects dynamically. For more information, read the [CognigyScript](/ai/platform-features/cognigyscript) documentation.

**Connections**

A Cognigy resource that securely stores and manages credentials for integrating external services or APIs with the platform. Connections let AI Agents access and interact with external systems, such as databases, APIs, or third-party services, while ensuring sensitive information, such as API keys or access tokens, is securely encrypted. For more information, read the [Connections](/ai/platform-features/connections) article.

**Contact Profiles**

The stored data about the users that interact with your AI Agents. Contact Profiles persistently store user data and help to personalize conversations, even across different channels. You can access Contact Profiles from a Flow through the [Profile](/ai/agents/develop/ai-agent-memory/profile) object and manage them with specific Nodes. For more information, read the [Contact Profiles](/ai/agents/analyze/contact-profiles) article.

**Context Object**

A JSON object that stores session-specific data. The Context object is generated when a user starts a conversation with an AI Agent, which triggers the start of a session. When the user disconnects from Cognigy, the session data persists in the [Context](/ai/agents/develop/ai-agent-memory/context) object and is retrieved upon reconnection with the same session ID. This method allows for multichannel conversations by accessing the Context object when the user connects from a different channel. You can also use CognigyScript to access the Context object and generate more context-aware AI Agent responses. For more information, read the [Context](/ai/agents/develop/ai-agent-memory/context) documentation.

**Conversation**

A billable unit of interaction between a user and an AI Agent or human agent. For more information about billed conversations, read the [Billing](/ai/administer/billing) documentation.

**Conversational AI**

A technology that allows computers to interact with humans in natural language. Conversational AI uses AI and natural language processing to simulate humanlike conversations and is used in various industries for tasks like customer service and information retrieval.

## D

**Data Redaction**

A Cognigy feature that provides an easy way to detect and redact any data, including [personally identifiable information (PII)](https://en.wikipedia.org/wiki/Personal_data), to ensure compliant [logging](/ai/agents/test/logs) and [analytics](/ai/agents/analyze/overview). Data redaction is important in the context of conversational AI due to privacy, trust, compliance, and responsible AI practices.

## E

**Endpoints**

A Cognigy resource that allows seamless interaction between AI Agents and external systems or platforms so that AI Agents can send and receive messages to and from external services. Each Endpoint represents a specific communication channel or interface through which the AI Agent can connect and engage with users or other systems. For more information, read the [Endpoints](/ai/agents/deploy/endpoints/overview) documentation.

**Extension**

Custom-built JavaScript modules that extend the platform's capabilities beyond its core features. With Extensions, you can create custom Nodes and integrate Knowledge AI with external knowledge databases to enhance the AI Agent's responses and interactions. For more information, read the [Extensions](/ai/for-developers/extensions) documentation.

**Extension Node**

A custom Node that extends the Flows' capabilities. Extension Nodes are created using the [Extension Framework](/ai/for-developers/extensions) and can be used in Flows to perform new actions, integrate with external systems, process data, and improve conversational scenarios.

## F

**Flow**

A conversational structure that defines how AI Agents interact with users. Flows are composed of Nodes with different functionalities, such as generating output and collecting backend information. When an input comes in through an Endpoint, the Flow is executed, enabling smooth interactions with users. For more information, read the [Flows](/ai/agents/develop/projects-and-flows/overview) documentation.

**Flow Editor**

A graphical interface where you can create and organize the conversational paths and define the interaction logic by using Nodes that represent different actions, conditions, prompts, and responses to design a [Flow](#f).

**Functions**

A Cognigy resource that you can define with JavaScript code and run to help with asynchronous processes, for example, interaction with third-party systems through HTTP APIs. Functions are independent of Flows, but you can trigger them from a Flow. For more information, read the [Functions](/ai/for-developers/functions) documentation.

## G

**Generative AI**

A type of artificial intelligence that generates new content, such as images, video, audio, and text, using machine learning algorithms. Generative AI produces new content using LLMs, which are trained on large datasets. Cognigy integrates with multiple LLM providers to improve real-time customer interactions with AI Agents through personalized responses, advanced language understanding, and sentiment analysis. For more information, read the [Generative AI](/ai/agents/develop/gen-ai-and-llms/generative-ai) documentation and see [Wikipedia](https://en.wikipedia.org/wiki/Generative_artificial_intelligence).

## H

**Handover**

A Cognigy feature that allows AI Agents to transfer conversations to human agents when necessary, combining automation with human support for improved customer service. You can integrate handover into a Flow using handover providers and the Handover to Human Agent Node. This approach allows you to connect AI Agents to various contact center platforms, such as Live Agent, Genesys, 8x8, Salesforce, and many others. For more information, read the [Handover](/ai/agents/develop/node-reference/service/handover-to-agent) documentation.

**Handover Provider**

A Cognigy resource that acts as a connector between an AI Agent and contact center platforms. You can transfer conversations to contact center platforms using the Handover to Human Agent Node. For more information, read the [Handover Providers](/ai/escalate/handover-reference/overview) documentation.

## I

**Inbox**

A container for incoming conversations routed to Live Agent. For more information, read the [Inbox](/live-agent/settings/inboxes) documentation.

**Input Object**

A JSON object that serves as the short-term AI Agent memory. The Input object is generated each time the user sends a message or provides a voice input to an AI Agent. The Input object stores general data about the user input, such as time and type of input, Endpoint-specific data, and results from the NLU Intent mapping process, such as the recognized Intent and Slots. You can use CognigyScript to access the Input object and generate more context-aware AI Agent responses. For more information, read the [Input](/ai/agents/develop/ai-agent-memory/input) documentation.

**Insights**

A Cognigy product that gathers and displays analytics data in Cognigy. Insights reports and explorers provide a range of key metrics for understanding the deployment and performance of your AI Agents. For more information, read the [Cognigy Insights](/insights/overview) documentation.

**Intents**

Patterns or categories you can define in Cognigy to allow AI Agents to better understand user input during conversational interactions. Intents represent the intentions or goals behind what a user says or types. Based on the Intent, AI Agents can take appropriate actions. For example, if a user says `Book a flight to New York for next Friday`, the Intent might be `Book a flight`. For more information, read the [Intents](/ai/platform-features/nlu/intents/overview) documentation.

**Intent Trainer**

A Cognigy feature that analyzes collected conversation records, processes them, and creates Intents to improve the comprehension of user input. For more information, read the [Intent Trainer](/ai/platform-features/nlu/intents/intent-trainer) documentation.

**Interaction Panel**

A Cognigy feature that provides a user interface for testing AI Agents in Cognigy. For more information, read the [Interaction Panel](/ai/agents/test/interaction-panel/overview) documentation.

## K

**Keyphrase**

A specific word or phrase you can define in a [Lexicon](#l) for the AI Agent to recognize, and that can be assigned to Slots and Synonyms. When Keyphrases are detected, they are stored in the Input object for further use, a process known as Slot mapping. For more information, read the [Keyphrases](/ai/platform-features/nlu/slots/user-defined/lexicon) and the [Slot mapping](/ai/platform-features/nlu/slots/system-defined) documentation.

**Knowledge AI**

A Cognigy feature that provides AI Agents with RAG capabilities to generate context-aware, tailored answers to user questions. Knowledge AI combines Knowledge Sources, embedding, chat, and completion models to improve user interactions with AI Agents. For more information, read the [Knowledge AI](/ai/agents/develop/knowledge-ai/overview) documentation.

**Knowledge Chunk**

A Knowledge AI component that represents a unit of information extracted from a Knowledge Source and can include metadata. Knowledge AI searches through Knowledge Chunks to answer user questions. For more information, read the [Knowledge AI](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk) documentation.

**Knowledge Connector**

A Knowledge AI component that links your Knowledge Store to external systems such as CRMs, help centers, or databases. You can use prebuilt Knowledge Connectors from the Marketplace, such as Confluence and Diffbot, or create custom ones using the [Extension framework](/ai/for-developers/extensions). For more information, read the [Knowledge AI](/ai/agents/develop/knowledge-ai/knowledge-connectors) documentation.

**Knowledge Store**

A Knowledge AI component that contains and organizes Knowledge Sources. The Knowledge Store helps streamline knowledge management by grouping related Knowledge Sources, making it easier to search and retrieve relevant information during run-time. For more information, read the [Knowledge AI](/ai/agents/develop/knowledge-ai/knowledge-store) documentation.

**Knowledge Source**

A Knowledge AI component that contains the source of information to generate Knowledge Chunks and can include metadata. You can create Knowledge Sources from content types, such as PDF, CTXT, web pages, or via [*Knowledge Connectors*](/ai/agents/develop/knowledge-ai/knowledge-connectors) that link to external knowledge bases. For more information, read the [Knowledge AI](/ai/agents/develop/knowledge-ai/knowledge-source/overview) documentation.

## L

**Large Language Model (LLM)**

A type of generative AI model that is designed to generate humanlike text based on the user input and the conversation history. LLMs allow AI Agents to understand and respond to questions in a natural way. For more information, read the [Large Language Models (LLMs)](/ai/agents/develop/gen-ai-and-llms/llms) documentation.

**Lexicon**

A collection of Keyphrases that you can attach to Flows to help AI Agents understand user queries. For more information, read the [Lexicons](/ai/platform-features/nlu/slots/user-defined/lexicon) documentation.

**Live Agent**

A Cognigy product that allows human agents to join and assist in conversations handled by the AI Agents. With Live Agent, AI Agents can escalate user queries to human agents, providing a more personalized and efficient customer support experience. For more information, read the [Live Agent](/live-agent/overview) documentation.

**Localization**

The process of adapting a product or a service, such as AI Agents, to a specific region or country, taking language and culture into account. Cognigy includes a localization feature that allows AI Agents to interact with end users in different languages. For more information, read the [Localization](/ai/platform-features/translation-and-localization/localization) documentation.

**Logs**

A Cognigy feature that allows you to track every input and output message in a conversation with an AI Agent along with other metadata, like timestamp, Endpoint, and more. Logs' main purpose is to help you troubleshoot AI Agents by reviewing errors and identifying where they occur. For more information, read the [Logs](/ai/agents/test/logs) documentation.

## M

**Members**

A Cognigy feature used to manage access rights through roles at the Project level. These roles complement the Global Roles assigned in the Access Control panel. For more information, read the [Members](/ai/administer/access/members) documentation.

**Mocking**

A Cognigy feature that allows you to replace a Node's execution behavior with custom mock code for testing purposes. With Mocking, you can test AI Agent scenarios without modifying Flows, avoid costly or time-consuming requests to external services, and ensure consistent testing interactions with deterministic data. For more information, read the [Mocking](/ai/agents/test/mocking) documentation.

## N

**NLU Connectors**

A Cognigy feature that provides built-in support for several third-party NLU engines. You can use NLU connectors to connect with third-party NLU engines, such as Google Dialogflow and IBM Watson Assistant. For more information, read the [NLU Connectors](/ai/platform-features/nlu/external/nlu-connector-reference/all-nlu-connectors) documentation.

**NLU Model**

A machine learning model used for natural language understanding tasks in Cognigy. The NLU model allows AI Agents to understand and interpret natural language input from users. For more information, read the [Intents](/ai/platform-features/nlu/intents/overview) documentation.

**Node**

A fundamental building block of Flows that defines the logic of AI Agents. Each Node represents a specific action, decision point, or interaction during the Flow execution. For more information, read the [Nodes](/ai/agents/develop/nodes/overview) documentation.

**Node Editor**

A graphical interface where you can set the parameters of a Node to define the Flow behavior. For more information, read the [Nodes](/ai/agents/develop/nodes/overview) documentation.

## O

**OData**

A RESTful API endpoint Cognigy provides that allows you to retrieve analytics records of AI Agents' conversations. For more information, read the [OData](/ai/agents/analyze/odata) documentation.

**Organization**

An independent instance of Cognigy within an environment, isolated for a group of users or licenses.

## P

**Package**

A Cognigy feature that allows you to export and import Project resources, such as Flows or Lexicons, transfer them to other Projects, or share them with other teams. For more information, read the [Packages](/ai/platform-features/packages) documentation.

**Playbook**

A Cognigy resource that represents conversation scripts to automatically run tests on AI Agents. You can create Playbooks and add Assertions to check the AI Agent's responses and other aspects of the Flow. For more information, read the [Playbooks](/ai/agents/test/playbooks) documentation.

**Profile Object**

A JSON object that serves as the long-term memory of the AI Agent. The Profile object stores persistent information about the user, displayed in the user's Contact Profile. When a user disconnects from Cognigy, their data remains in the Profile object and is retrieved upon reconnection with the same user ID. This allows for multichannel conversations by accessing the Profile object when the user connects from a different channel. You can also use CognigyScript to access the Profile object and generate more context-aware AI Agent responses. For more information, read the [Profile](/ai/agents/develop/ai-agent-memory/profile) documentation.

**Project**

A workspace where you can create, test, and deploy AI Agents. Projects include different resources, such as Flows, Lexicons, and Endpoints, and allow for collaboration with multiple team members. For more information, read the [Projects](/ai/agents/develop/projects-and-flows/projects) documentation.

## S

**Session**

The complete interaction between a user and AI Agent or human agent. The session begins when a user sends their initial chat or voice message to the AI Agent, and ends when the user, AI Agent, or human agent completes or terminates the interaction. A session can contain an unlimited number of user inputs and can be carried out across different channels.

**Simulator**

An LLM-powered tool that allows you to model and test full conversational scenarios for LLM-based AI Agents.

**Slot**

A specific data type or custom phrases that Cognigy NLU can recognize. Slots allow AI Agents to extract structured information from user input and use it dynamically in conversations. For more information, read the [Slots](/ai/platform-features/nlu/slots/system-defined) documentation.

**Snapshot**

A Cognigy resource that represents an immutable version of an AI Agent that you can publish to users, or use to create backups or to revert to an earlier state. For more information, read the [Snapshots](/ai/agents/deploy/snapshots) documentation.

**State**

A Cognigy feature that limits which Intents the AI Agent can recognize during a part of a Flow.

## T

**Task**

Specific activities that run in the background and for which you get notifications, such as training Intents, importing Snapshots, or creating a new Project. The Task panel provides an overview of ongoing and completed Tasks. For more information, read the [Tasks](/ai/agents/develop/projects-and-flows/projects#task-panel) documentation.

**Tenant**

An independent instance of Cognigy within an environment, also known as *organization*, isolated for a group of users or licenses.

**Tokens**

A Cognigy resource that visually represents [CognigyScript](/ai/platform-features/cognigyscript) expressions in the Node editor and stores them for reuse. Cognigy has a selection of default Tokens that are associated with Input, Context, and Profile objects. For more information, read the [Tokens](/ai/platform-features/tokens) documentation.

## U

**User Input**

A message, including text, data, or voice input, received by Cognigy from a user or system.

**User Menu**

A menu interface that provides access to a set of views related to the user that is logged in. For more information, read the [User Menu](/ai/administer/user-menu/overview) documentation.

## V

**Voice Gateway**

A Cognigy product that allows you to deploy voice AI Agents by phone and in contact center platforms, by connecting them to Cognigy. For more information, read the [Voice Gateway](/ai/agents/deploy/endpoint-reference/voice-gateway) documentation.

## W

**Webchat**

A Cognigy product that integrates AI Agents into websites seamlessly, facilitating interactive engagement with visitors. For more information, read the [Webchat](/webchat/overview) documentation.

## X

**xApps**

A Cognigy product that allows you to build flexible micro-web applications for customer self-service. xApps can be used as a standalone application or complementary to any channel to enhance AI and human agents' capabilities. With xApps, you can overcome channel limitations that compromise user experiences. For more information, read the [xApps](/xApps/overview) documentation.

## Y

**Yes/No Intents**

A Cognigy feature that allows machine learning to recognize yes and no responses similarly to regular Intent matching. With Yes/No Intents, you can train an NLU model to understand different yes and no confirmation phrases based on a small set of example sentences. For more information, read the [Yes/No Intents](/ai/platform-features/nlu/intents/yes-no-intents) documentation.


# Key Features
Source: https://docs.cognigy.com/ai/overview/key-features



Cognigy is a leading AI-first CX platform for building, managing and deploying AI Agents at enterprise scale.

<Frame>
  <img alt="Cognigy key features for building and deploying AI Agents" />
</Frame>

## [Agentic AI](/ai/agents/overview)

Agentic AI combines the language and reasoning abilities of LLMs with real-time context, memory, tools and enterprise governance to create AI Agents that understand intent, flexibly handle digression and multi-turn conversations, and take action based on the conversation context in contrast to classic linear flows.

Cognigy allows for seamless combination of Agentic AI with deterministic, rule-based automation, enabling enterprise control where needed and adaptability for sophisticated interactions (Hybrid AI).

## [AI Agent Studio](/ai/agents/develop/projects-and-flows/editor)

Cognigy's AI Agent Studio (Flow editor) offers a low/no-code collaborative, visual interface for business users, conversation designers and developers to co-create, deploy and manage AI Agents. It provides AI-assisted design, powerful tooling for scalable testing and debugging, and 100+ channel and system connectors, accelerating development and time-to-value. You can create projects, manage resources like LLMs and collaborate in real time on Agents.

## [Knowledge AI](/ai/agents/develop/knowledge-ai/overview)

Cognigy can ingest existing knowledge sources such as documents or internal repositories like Confluence into vector databases. AI Agents then employ Retrieval Augmented Generation (RAG), enabling the AI Agent to accurately answer user queries in natural language based on the individual conversation context, instead of a standardized reply.

## [NLU](/ai/platform-features/nlu/overview)

Cognigy NLU is an intent-driven Natural Language Understanding AI solution for building deterministic or rule-based processes. Hybrid AI Cognigy enables you to combine NLU with the language understanding and generation power of LLMs. This can be useful, for example, when extracting multiple entities from a single user utterance or when understanding more complex utterances such as `me, my wife and our two kids` which an NLU can mistake as two people based on the word whereas an LLM would understand the user means four total people. Cognigy's AI Agents can be built purely with NLU, purely with Agentic AI or using both within the same project and even within the same Agent.

## [Native Voice Gateway](/voice-gateway/technical-capabilities)

Cognigy's native voice solution enables voice connectivity and supports a free choice of STT and TTS models from providers like Google, Microsoft, ElevenLabs, Deepgram and more.

## [Multimodal xApps](/xApps/use-cases)

xApps are mobile first, OS agnostic micro web applications that can be integrated in conversational experiences, enabling multimodal, transaction-ready interactions that transcend the constraints of traditional chat and voice channels. This enables customers to mix channels or transition to another within the same conversation, without losing context or having to start over. For example, a customer could send an image in the middle of a voice call, share their GPS location from a chat or make a payment while using WhatsApp.

## [Native Contact Center Integrations](/ai/escalate/handovers)

Cognigy is designed to be technology and platform agnostic, integrating with multiple CCaaS, CRM, case management solutions and more. Cognigy can sit "in front" of CCaaS solutions for example with voice and digital interactions routed between them if necessary. Cognigy's Agent Copilot can also be integrated into the agent desktop of CCaaS solutions or used standalone within a browser tab.

## [Analytics and Business Dashboard](/ai/agents/analyze/overview)

Understand and optimize how AI Agents perform across every channel using millions of data points while also measuring business impact with goals, time saved and cost savings based on your business. Insights are available in real-time, aggregated or at a granular level for deep-dives into individual customer journeys for example. Data can also be used in external analytics applications such as Power BI.


# Main Page
Source: https://docs.cognigy.com/ai/overview/user-interface/main-page

Explore the Cognigy Main page, where you can manage, pin, and create Projects.

In Cognigy's **Main** page, you can manage your Projects.

The following image provides an overview of the **Main** page:

<Frame>
  <img alt="Cognigy Main page with Project list and management options" />
</Frame>

<AccordionGroup>
  <Accordion title="1. Project List">
    In the **Project List**, you can pin or unpin a Project by hovering over a Project, clicking the <img alt="project options" /> button, and selecting **Pin Project** or **Unpin Project**.

    The <img alt="create project" /> button opens the [Project Creation Wizard](/ai/agents/develop/projects-and-flows/projects).
  </Accordion>

  <Accordion title="2. Top Bar">
    |          Button         | Description                                                                                                                                              |
    | :---------------------: | -------------------------------------------------------------------------------------------------------------------------------------------------------- |
    |    <img alt="help" />   | Opens the Help Center search widget, where you can search for answers to potential questions in the [Cognigy Help Center](https://support.cognigy.com/). |
    | <img alt="user-menu" /> | Opens the [User Menu](/ai/administer/user-menu/overview).                                                                                                |
  </Accordion>

  <Accordion title="3. Welcome Dashboard">
    | Section     | Description                                                       |
    | ----------- | ----------------------------------------------------------------- |
    | News        | Includes Cognigy updates.                                         |
    | Get Started | Links to the [Cognigy Help Center](https://support.cognigy.com/). |
  </Accordion>
</AccordionGroup>


# Project Page
Source: https://docs.cognigy.com/ai/overview/user-interface/project-page

Explore the Cognigy Project page and its navigation sidebar.

The **Project** page is the landing page of your Project, where you have access to all of Cognigy's capabilities to build your AI Agents.

The following image shows a Project page:

<Frame>
  <img alt="Cognigy Project page showing the left-side menu and project resources" />
</Frame>

<AccordionGroup>
  <Accordion title="1. Left-Side Menu">
    |          Menu Item         | Name                   | Description                                                                                                                                                                                                                        |
    | :------------------------: | ---------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    |   <img alt="Main Page" />  | **Main Page**          | Returns to the **Main** page.                                                                                                                                                                                                      |
    | <img alt="Project Page" /> | **Project Page**       | Overview of your Project.                                                                                                                                                                                                          |
    |     <img alt="build" />    | **Build**              | Configure the building blocks of your AI Agent.                                                                                                                                                                                    |
    |    <img alt="deploy" />    | **Deploy**             | Configure Endpoints to third-party systems and manage Snapshots for your production and test environments.                                                                                                                         |
    |     <img alt="test" />     | **Test**               | Use Playbooks and logs to test your AI Agent.                                                                                                                                                                                      |
    |     <img alt="tweak" />    | **Tweak**              | Use the [Intent Trainer](/ai/platform-features/nlu/intents/intent-trainer) to analyze user input and improve your AI Agent.                                                                                                        |
    |   <img alt="insights" />   | **Insights**           | Opens **Insights**, where you can have an overview of key metrics that help you understand your AI Agents' performance.                                                                                                            |
    |    <img alt="manage" />    | **Manage**             | Configure your Project settings, such as [languages](/ai/platform-features/translation-and-localization/localization), [Tokens](/ai/platform-features/tokens), [Contact Profiles](/ai/agents/analyze/contact-profiles) and others. |
    |     <img alt="help" />     | **Help Center search** | Opens the Help Center search widget, where you can search for answers to potential questions on the [Cognigy Help Center](https://support.cognigy.com/).                                                                           |
    |   <img alt="User Menu" />  | **User Menu**          | Opens the [User Menu](/ai/administer/user-menu/overview).                                                                                                                                                                          |
  </Accordion>

  <Accordion title="2. Top Bar">
    |            UI Element           | Type   | Description                                                                                                        |
    | :-----------------------------: | ------ | ------------------------------------------------------------------------------------------------------------------ |
    |    <img alt="project-list" />   | List   | Allows you to switch between Projects.                                                                             |
    |       <img alt="search" />      | Button | Opens the search field.                                                                                            |
    |     <img alt="task-menu" />     | Button | Opens the [Task Panel](/ai/agents/develop/projects-and-flows/projects#task-panel) and displays Task notifications. |
    | <img alt="interaction-panel" /> | Button | Opens the [Interaction Panel](/ai/agents/test/interaction-panel/overview) for testing AI Agents.                   |
  </Accordion>

  <Accordion title="3. Dashboard">
    | Section            | Description                                                                                                                                                  |
    | ------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Conversation Chart | Displays the number of conversations in the last 30 days and the **Show more in Insights** button, that redirects to the **Insights** or **Analytics** page. |
    | Project Member     | Displays the Project members.                                                                                                                                |
    | Recently Changed   | Displays the last changed Project.                                                                                                                           |
    | Explore Cognigy    | Redirects you to the Getting Started page on the [Cognigy Help Center](https://support.cognigy.com/).                                                        |
  </Accordion>
</AccordionGroup>


# Resource Pages
Source: https://docs.cognigy.com/ai/overview/user-interface/resource-pages

Resource pages let you manage Cognigy resources like Flows, Endpoints, and Knowledge Stores.

Resource pages are pages where you can view and manage resources to build AI Agents, for example, Flows, Knowledge Stores, Endpoints, and many others. You can access resource pages in the [left-side menu](/ai/overview/user-interface/project-page) of the **Project** page.

The following image provides an overview of the Flows page:

<Frame>
  <img alt="Overview of the Flows resource page in Cognigy" />
</Frame>

<AccordionGroup>
  <Accordion title="Controls">
    |           UI Element          | Type   | Description                                                                                                                                                                                                              |
    | :---------------------------: | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | <img alt="create resource" /> | Button | You can create resources by clicking **+** in the upper-left corner.                                                                                                                                                     |
    |      <img alt="search" />     | Field  | You can search for resources by name.                                                                                                                                                                                    |
    |   <img alt="more options" />  | Menu   | With the right-side menu, you can perform other actions, for example, edit, delete, export as a Package, clone, copy Reference ID, as well as other resource-specific options. This menu is available for each resource. |
  </Accordion>

  <Accordion title="Notifications">
    | Notification Type | Description                                                                                                              | Example                                   |
    | ----------------- | ------------------------------------------------------------------------------------------------------------------------ | ----------------------------------------- |
    | Success message   | Displays when a Task is completed successfully.                                                                          | <img alt="Success notification banner" /> |
    | Warning message   | Display when a Task is completed with a warning. Click <img alt="expand warning message" /> to view the warning message. | <img alt="Warning notification banner" /> |
    | Error message     | Displays when a Task fails. Click <img alt="expand error message" /> to view the error message.                          | <img alt="Error notification banner" />   |
  </Accordion>
</AccordionGroup>


# CognigyScript
Source: https://docs.cognigy.com/ai/platform-features/cognigyscript



*CognigyScript* is a superset of JavaScript that provides access to the [Input](/ai/agents/develop/ai-agent-memory/input), [Context](/ai/agents/develop/ai-agent-memory/context), and [Profile](/ai/agents/develop/ai-agent-memory/profile) objects. With CognigyScript expressions, you can retrieve values from JSON properties, and use these values in AI Agent conversations.

In most cases, you wrap CognigyScript expressions in double curly braces and quotation marks `{{ }}`, for example, `{{input.text}}`. This syntax enables Cognigy to identify and execute the CognigyScript code embedded in the Flow. However, CognigyScript evaluates conditions as standard JavaScript. You don't need to wrap the CognigyScript expression in `{{ }}`. For example, `context.orders === 3` evaluates to true if the `orders` variable in the Context is equal to `3`.

You can use CognigyScript to:

* Add dynamic and executable content to chat and voice conversations.
* Retrieve information from [Cognigy objects](/ai/agents/develop/ai-agent-memory/overview#lifespan-of-cognigy-objects) and reuse it in chat and voice conversations.
* Run scripts to manipulate data dynamically.
* Use CognigyScript in JSON arguments to make requests to an external system via the HTTP Request Node.

## Working with CognigyScript

You can use CognigyScript in:

* Text fields
* JSON editors
* Code Nodes

The type of field in which you use CognigyScript determines the CognigyScript syntax.

You can also store CognigyScript in [Tokens](/ai/platform-features/tokens).

### Text Fields

For text fields, wrap the CognigyScript expression in `{{ }}`.

<Note>
  If the CognigyScript expression is invalid, CognigyScript returns an empty string.
</Note>

<Accordion title="Example">
  `{{input.text.toUpperCase()}}` returns the text the client sent in uppercase format.
</Accordion>

### JSON Editors

The CognigyScript syntax in JSON editors depends on where you place the CognigyScript expression.

#### Inline CognigyScript

For inline CognigyScript in a JSON object, use `{{ }}`.

<Accordion title="Example">
  ```json theme={null}
  {
      "customer_orders": "{{context.orders}}"
  }
  ```

  The code attempts to fetch the `orders` variable from the Context object and to assign it to `customer_orders`. If `context.orders` doesn't exist, the code skips `customer_orders`.
</Accordion>

#### Inside JSON Arguments

To run CognigyScript within a JSON argument, use `{ "$cs": { "script": "x", "type": "t"}}`, where `x` is the script and `t` is the return type (optional). For instance, you can convert a string such as `"6"` into a number or an object into a string.

<Accordion title="Example">
  ```json theme={null}
  {
      "customer_orders": {
          "$cs": {
              "script": "context.orders",
              "type": "object"
          }
      }
  }
  ```

  The code attempts to fetch the `orders` variable from the Context object and to assign it to `customer_orders` as an object. If `context.orders` doesn't exist, the code skips `customer_orders`.
</Accordion>

### Code Nodes

In a Code Node, you don't need to wrap the CognigyScript expression in `{{ }}`. You can use CognigyScript as standard JavaScript. The `input`, `context`, `profile`, and `actions` variables are exposed by default, as well as `_` and `moment`.

<Accordion title="Example">
  ```js theme={null}
  const ordercount = context.orders;
  switch (ordercount) {
      case 0:
          input.ordertext = "You ordered no items";
          api.output('Hello', {'action': 1}); // outputs the text 'hello' with data {'action': 1}
          break;
      case 1:
          input.ordertext = "You ordered one item";
          break;
      default:
          input.ordertext = "You ordered many items";
  }
  ```
</Accordion>

## Additional Examples

<Accordion title=" Examples of CognigyScript Expressions">
  The following table shows additional examples of CognigyScript expressions:

  | Example                                                             | Description                                                                                                               |
  | ------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------- |
  | `{{input.slots.SLOT_NAME[0].keyphrase}}`                            | Extracts the first Keyphrase found for the Slot with the name `SLOT_NAME`.                                                |
  | `{{JSON.stringify(input.data)}}`                                    | Displays a JSON object as text. You can use this CognigyScript expression for debugging purposes.                         |
  | `{{input.slots.DATE[0].start.plain}}`                               | Returns the date information if the user sent a date. Example: July 23, 2021                                              |
  | `{{moment(input.slots.DATE[0].start.ISODate).format('MMM Do YY')}}` | Formats a given date that the user provided in the last message.                                                          |
  | `{{input.currentTime.hour}}`                                        | Gets the hour of the current incoming message. You can use this CognigyScript expression to greet the user appropriately. |
</Accordion>

## More Information

* [Input](/ai/agents/develop/ai-agent-memory/input)
* [Context](/ai/agents/develop/ai-agent-memory/context)
* [Profile](/ai/agents/develop/ai-agent-memory/profile)


# Connections
Source: https://docs.cognigy.com/ai/platform-features/connections

Manage Connections in Cognigy to securely store credentials for external services, APIs, LLMs, and voice providers used by AI Agents.

*Connections* are configurations that empower AI Agents to securely interact with external services, systems, or APIs. These configurations involve the encryption, storage, and management of credentials for seamless communication between Cognigy and external resources.

In contrast to other resources in Cognigy, you can't create a Connection from scratch on the Connections page. You create Connections while adding a Node that requires credentials or adding a provider for [LLM](/ai/agents/develop/gen-ai-and-llms/providers/all-providers), [translation](/ai/platform-features/translation-and-localization/auto-translation), [voice](/ai/agents/test/voice-preview), [handover](/ai/escalate/handovers), or [file storage](/ai/agents/deploy/endpoints/file-storage).

## Working with Connections

<Tabs>
  <Tab title="GUI">
    You can view, edit, and delete Connections in **Build > Connections**. Also, you can copy their Reference ID or export Connections as a [Package](/ai/platform-features/packages).
  </Tab>

  <Tab title="API">
    You can view, edit, and delete Connections using the [Cognigy API](https://api-trial.cognigy.ai/openapi#get-/v2.0/connections).
  </Tab>
</Tabs>

<Warning>
  If you delete a Connection, features relying on this Connection stop working. When you delete a Connection, the Flow that uses the deleted Connection in a Node isn't updated automatically. Executing the Flow that uses the deleted Connection hangs on that Node and can cause error messages.
</Warning>

## More Information

* [Extensions](/ai/for-developers/extensions)
* [Nodes](/ai/agents/develop/nodes/overview)
* [HTTP Request Node](/ai/agents/develop/node-reference/service/http-request)


# Data Redaction
Source: https://docs.cognigy.com/ai/platform-features/data-redaction



<a href="/release-notes/2026.9"><Badge>Updated in 2026.9</Badge></a>

The *Data Redaction* feature provides an easy way to detect and redact any data, including [personally identifiable information (PII)](https://en.wikipedia.org/wiki/Personal_data), to ensure compliant [logging](/ai/agents/test/logs) and [analytics](/ai/agents/analyze/overview). This feature is important in the context of conversational AI due to privacy, trust, compliance, and responsible AI practices.

## Key Benefits

* **Enhanced Data Handling**. Ensures data protection during the processing, logging, and storage of conversations containing sensitive data. Compared to data masking, which hides or selectively excludes information at the [Endpoint](/ai/agents/deploy/endpoints/data-protection-and-analytics) or Flow level ([Blind Mode Node](/ai/agents/develop/node-reference/analytics/blind-mode)), data redaction fully removes the sensitive data from all logs and analytics for stronger data protection.
* **Easy Configuration**. Includes a set of [default data types](#default-data-types) you can select to redact, such as credit card numbers or email addresses.
* **Customization**. Lets you add [custom data types](#custom-data-type) to redact. You can use regular expressions (regex) to define patterns for custom data types to detect.

## Prerequisites

For on-premises installations, you control whether data redaction is activated in the `values.yaml` file. By default, data redaction is activated. Depending on which Cognigy version you use, you can deactivate data redaction as follows:

<Tabs>
  <Tab title="Cognigy 2026.9 and later">
    In the `values.yaml` file, configure the `commonConfigs.cognigyRollout.pii-data-redaction.yaml` key:

    <AccordionGroup>
      <Accordion title="Deactivate data redaction for all organizations">
        To deactivate data redaction for all organizations, set `percentage` to `0`.

        ```yaml theme={null}
        commonConfigs:
          cognigyRollout:
            pii-data-redaction.yaml: |
              organisations:
                percentage: 0
                disabled: []
        ```
      </Accordion>

      <Accordion title="Deactivate data redaction for specific organizations">
        To deactivate data redaction for specific organizations, assign organization IDs to the `disabled` array.

        ```yaml theme={null}
        commonConfigs:
          cognigyRollout:
            pii-data-redaction.yaml: |
              organisations:
                percentage: 100
                disabled:
                  - "<organization-id-1>"
                  - "<organization-id-2>"
        ```
      </Accordion>
    </AccordionGroup>
  </Tab>

  <Tab title="Cognigy 2026.8 and earlier">
    In the `values.yaml` file, set the following environment variables:

    * `FEATURE_PII_DATA_REDACTION_ROLLOUT` — the percentage of organizations for which data redaction is activated. Set it to `"100"` to activate data redaction for all organizations, or to `"0"` to disable it for all. Values in between activate the feature for that percentage of organizations. The system automatically and consistently determines which organizations are included.
    * `FEATURE_DISABLE_PII_DATA_REDACTION_WHITELIST` — a comma-separated list of organization IDs to explicitly exclude from data redaction. This setting takes precedence over `FEATURE_PII_DATA_REDACTION_ROLLOUT`. For example, `FEATURE_DISABLE_PII_DATA_REDACTION_WHITELIST="<organization-id-1>,<organization-id-2>"`.
  </Tab>
</Tabs>

## Limitations

* Log entries with `data` payloads over 2.5 MB aren't redacted. They're replaced with a placeholder message. For on-premises installations, you can configure the maximum payload size using the `PII_REDACTION_DEFAULT_MAX_DATA_SIZE_IN_MEGABYTES` environment variable in `values.yaml`. Setting this value higher than 2.5 MB may cause performance issues. Increase it at your own risk.

## Restrictions

* Data redaction uses regex patterns to detect and redact data, which can:
  * Cause false positive matches. For example, the credit card redaction pattern may cause false positives for 3- or 4-digit numbers that are sent in the message input but aren't CVC or CVV numbers.
  * Skip patterns that don't match the [default types](#default-data-types) or a [custom data type](#custom-data-type) you define. For example, if the user enters a phone number with `/` or misspells a company name set to be redacted in a custom data type.
* Custom data types support only Google's [RE2 syntax](https://github.com/google/re2) library for regex patterns.
* Depending on the Cognigy version you use, data redaction works differently with [Slots](/ai/platform-features/nlu/slots/overview):

  <AccordionGroup>
    <Accordion title="Cognigy 2026.3 and later versions">
      Data redaction works for patterns detected by Slots only if the Slot matches the entire data redaction pattern. If a Slot matches only part of the pattern, the data isn't redacted as expected.

      For example, if a credit card number is split across multiple Slots, such as [a number Slot and a date Slot](/ai/platform-features/nlu/slots/system-defined#list-of-system-defined-slots), the credit card number isn't fully redacted.
      Make sure you use Slots that match the exact data redaction pattern, or deactivate the Slots that cause the pattern to split either at the [Flow](/ai/agents/develop/projects-and-flows/overview#system-slots) or the [Project](/ai/administer/access/project-settings#nlu-settings) level.

      If you use [Regex Slots](/ai/platform-features/nlu/slots/user-defined/regex), you need to remove the [Slot Filler](/ai/platform-features/nlu/slot-fillers) that contains the Regex Slot.
    </Accordion>

    <Accordion title="Cognigy 2026.2 and earlier versions">
      Data redaction doesn't work for patterns detected by Slots. To redact such data, deactivate Slots at the [Flow](/ai/agents/develop/projects-and-flows/overview#system-slots) or the [Project](/ai/administer/access/project-settings#nlu-settings) level.

      If you use [Regex Slots](/ai/platform-features/nlu/slots/user-defined/regex), you need to remove the [Slot Filler](/ai/platform-features/nlu/slot-fillers) that contains the Regex Slot.
    </Accordion>
  </AccordionGroup>

## How to Use

To use the data redaction feature, first check the [default data types](#default-data-types). If none are suitable or if an existing type doesn't match your data due to its pattern rules, you can create a [custom data type](#custom-data-type).

### Default Data Types

The data redaction feature includes the following default data types, covering common forms of PII:

| Data Type              | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             | Examples                                                                                                                                                               | Redaction Pattern                                                   |
| ---------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------- |
| Credit Card Number     | Redacts:<ul><li>16-digit card numbers starting with `3` (Amex), `4` (Visa), `5` (Mastercard), or `6` (Discover). Digits may be separated by spaces or hyphens.</li><li>Card verification values (CVV) or card validation codes (CVC) with 3 or 4 digits. CVC and CVV numbers are redacted only after a credit card number has been recognized and if they are included in the same message as the credit card number.</li></ul>The credit card number redaction is activated by default to protect credit card data and ensure compliance with PCI DSS, GDPR, and CCPA. | <ul><li>Credit card numbers: `4111111111111111`, `4111 1111 1111 1111`, `5555-5555-5555-4444`, `3782-822463-10005`</li><li>CVV or CVC numbers: `123`, `1234`</li></ul> | <ul><li>`[CREDIT CARD REDACTED]`</li><li>`[CVV REDACTED]`</li></ul> |
| Email Address          | Redacts email addresses. Must contain a valid local part, an `@` symbol, and a valid domain. By default, the email address redaction is deactivated.                                                                                                                                                                                                                                                                                                                                                                                                                    | `test@example.com`, `admin@company.org`, `user@domain.com`                                                                                                             | `[EMAIL REDACTED]`                                                  |
| Phone Number           | Redacts phone numbers in U.S. national, U.S. local, or international formats. By default, the phone number redaction is deactivated.                                                                                                                                                                                                                                                                                                                                                                                                                                    | `(555) 123-4567`, `+1 (555) 123-4567`, `555-123-4567`, `5551234567`, `+491575313119`, `+447911123456`                                                                  | `[PHONE REDACTED]`                                                  |
| Social Security Number | Redacts 9-digit U.S. Social Security numbers in `XXX-XX-XXXX` format. By default, the Social Security number redaction is deactivated.                                                                                                                                                                                                                                                                                                                                                                                                                                  | `123-45-6789`, `987-65-4320`                                                                                                                                           | `[SSN REDACTED]`                                                    |
| IPv4 Address           | Redacts IPv4 addresses in `X.X.X.X` format, where each octet is 0–255. Supports private ranges. By default, the IPv4 address redaction is deactivated.                                                                                                                                                                                                                                                                                                                                                                                                                  | `192.168.1.1`, `255.255.255.255`, `10.0.0.1`                                                                                                                           | `[IPv4 REDACTED]`                                                   |
| IPv6 Address           | Redacts IPv6 addresses, including full, shortened, or loopback (`::1`) forms. By default, the IPv6 address redaction is deactivated.                                                                                                                                                                                                                                                                                                                                                                                                                                    | `2001:0db8:85a3:0000:0000:8a2e:0370:7334`, `::1`, `2001:db8::1`, `fe80::1`                                                                                             | `[IPv6 REDACTED]`                                                   |

To activate the data redaction for default data types, follow these steps:

1. In **Manage > Settings**, go to the **Data Redaction Settings** section.
2. Click the subsection of the data type you want to redact and configure the following:
   * **Activate Redaction** — toggle on to activate the redaction for the data type.
   * **Redaction Pattern** — select one of the following options:
     * **Default** — use the default redaction pattern. Each data type has a default redaction pattern, for example, the redaction pattern for credit card numbers is `[CREDIT CARD REDACTED]`.
     * **Custom** — enter a custom redaction pattern. The custom redaction pattern should be enclosed in square brackets and must not contain spaces or special characters. For example, `[PROTECTED_INFO]`.
3. Under **Scope**, select where the data type is redacted:
   * **Logs** — redact the data type on the Logs page.
   * **Analytics** — redact the data type in Insights and the Cognigy OData endpoint.

### Custom Data Type

If you need to redact data that isn't included in the default data types, you can add custom data types.

To add a custom data type, follow these steps:

1. In **Manage > Settings**, click the **Data Redaction Settings** section to expand it.
2. At the bottom of the section, click **+ Custom Data Type** and configure the following:
   * **Name** — enter a unique name.
   * **Regex** — enter a regular expression to detect the data type pattern.
   * **Custom redaction pattern** — enter a custom redaction pattern. The custom redaction pattern should be enclosed in square brackets and must not contain spaces or special characters. For example, `[PROTECTED_INFO]`.
3. Under **Scope**, select where the data type is redacted:
   * **Logs** — redact the data type on the Logs page.
   * **Analytics** — redact the data type in Insights and the Cognigy OData endpoint.

#### Examples

The following table includes examples of regex patterns and redaction patterns for custom data types.

| Custom Data Type     | Regex                                        | Redaction Pattern             |
| -------------------- | -------------------------------------------- | ----------------------------- |
| ZIP Code             | `\b\d{5}-\d{4}\b`                            | `[REDACTED_ZIP_CODE]`         |
| German License Plate | `\b[A-ZÄÖÜ]{1,3}-[A-Z]{1,2}\s\d{1,4}[HE]?\b` | `[REDACTED_DE_LICENSE_PLATE]` |
| US P.O. Box          | `\b(?:P\.?O\.?\|POB)\s?Box?\s?\d+\b`         | `[REDACTED_US_PO_BOX]`        |

## More Information

* [Logs](/ai/agents/test/logs)
* [Insights](/insights/overview)
* [Slots](/ai/platform-features/nlu/slots/overview)
* [Project Settings](/ai/administer/access/project-settings)


# Attachments
Source: https://docs.cognigy.com/ai/platform-features/nlu/attachments



*Attachments* allow you to connect [Flows](/ai/agents/develop/projects-and-flows/overview) and reuse [Lexicons](/ai/platform-features/nlu/slots/user-defined/lexicon) across multiple Flows within a Project.

## Key Features

* **Flow sharing**. Reuse Flows and their Intents across different Flows, eliminating the need to manually rebuild the NLU model in the attached Flows.
* **Lexicon sharing**. Reuse Lexicons across different Flows, ensuring consistent language understanding.
* **Centralized resource management**. Manage Flows and Lexicons in one location. This approach simplifies updates; a change in one place applies everywhere it's used.
* **Elimination of duplication**. You don't need to copy Flows or Lexicons. This approach saves space and keeps your Project more organized.

## Attached Flows

Attached Flows share their Intents with the Flow they're attached to. Attaching Flows is optional and
allows you to reuse existing Intents without duplicating them.
If no Flows are attached, the primary Flow relies only on its own Intents.

You don't need to manually train attached Flows after modifying their Intents, as long as they aren't used individually. For example, if you attach Flow B and Flow C to Flow A, you need to train Flow A. Flow B and Flow C will be trained automatically.
However, Flow B and Flow C won't be listed as trained on the **Flows** page because they're attached to Flow A.

To attach a Flow to another Flow, follow these steps:

<Tabs>
  <Tab title="GUI">
    1. Open your Flow and select **NLU** tab in the upper-right corner.
    2. Select the **Attached Flows** tab and activate the toggle next to the Flow that you want to attach to the current Flow.
  </Tab>

  <Tab title="API">
    * To a new Flow, use in the [Cognigy API POST /v2.0/flows](https://api-trial.cognigy.ai/openapi#post-/v2.0/flows) request.
    * To an existing Flow, use the [Cognigy API PATCH /v2.0/flows/](https://api-trial.cognigy.ai/openapi#patch-/v2.0/flows/-flowId-) request.

    When you attach Flows, the NLU model doesn't run automatically.
    Anytime a new Flow is attached, you need to manually rebuild the NLU model in the Flow to which you attached them.
    To run the NLU model manually, click **Build Model** in **NLU > Intents**.
  </Tab>
</Tabs>

## Attached Lexicons

Attached Lexicons share their Keyphrases with the Flow they're attached to.
Attaching Lexicons is important because it allows precise control over which Keyphrases are recognized in specific Flows.
By attaching different Lexicons to different Flows, you can create customized conversational experiences for different use cases.

To attach a Lexicon to a Flow, follow these steps:

<Tabs>
  <Tab title="GUI">
    1. Open your Flow and select **NLU** tab in the upper-right corner.
    2. Select the **Attached Lexicons** tab and activate the toggle next to the Lexicon that you want to attach to the current Flow.
  </Tab>

  <Tab title="API">
    * To a new Flow, use in the [Cognigy API POST /v2.0/flows](https://api-trial.cognigy.ai/openapi#post-/v2.0/flows) request.
    * To an existing Flow, use the [Cognigy API PATCH /v2.0/flows/](https://api-trial.cognigy.ai/openapi#patch-/v2.0/flows/-flowId-) request.

    When you attach Lexicons, the NLU model doesn't run automatically. Anytime a new Lexicon is attached, you need to manually rebuild the NLU model in the Flow to which you attached them. To run the NLU model manually, click **Build Model** in **NLU > Intents**.
  </Tab>
</Tabs>

## More Information

* [Flows](/ai/agents/develop/projects-and-flows/overview)
* [Lexicons](/ai/platform-features/nlu/slots/user-defined/lexicon)


# Intent Recognition with External Embedding Model
Source: https://docs.cognigy.com/ai/platform-features/nlu/external/external-nlu-intent-recognition



<a href="/release-notes/4.74"><Badge>Updated in 4.74</Badge></a>

An embedding is a sequence of numbers that represents the meaning of sentences and is used by the NLU for tasks such as Intent classification.

The *Intent Recognition with External Embedding Model* feature allows
using the `text-embedding-3-large` hosted by OpenAI and Azure OpenAI, replacing the default Cognigy embedding models.

<Note>
  The Azure OpenAI and OpenAI models are multilingual, and the NLU will work even if the Cognigy Flow is set in a different language than the user input. We recommend creating individual Flows for each localization.
</Note>

## Key Benefits

* The feature is designed for complex projects with a large number of Intents and example sentences.
* It enables you to set up the Intent model faster because you spend less time selecting and refining example sentences.
* There is a cost associated with using third-party models. It involves making external requests, which may incur expenses and latency.

## Prerequisites

* Add the [`text-embedding-3-large`](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) model provided by [Azure OpenAI](/ai/agents/develop/gen-ai-and-llms/providers/microsoft-azure-openai#add-models) or [OpenAI](/ai/agents/develop/gen-ai-and-llms/providers/openai#add-models). Note that if you use an external embedding model provider, Cognigy will send user inputs to the selected service. Ensure that your data processing policy reflects this practice.
* Consider that using the `text-embedding-3-large` external NLU model increases [Snapshot](/ai/agents/deploy/snapshots) size, leading to longer download and upload times.

## Set Up a Third-Party NLU model

To set up a third-party NLU model, follow these steps:

1. In the left-side menu of the Project, go to **Manage > Settings**.
2. Go to the **NLU Settings** section.
3. From the **Embedding Model Provider** list, select a model that you added as part of the [Prerequisites](#prerequisites).
4. Confirm changes.

## Train Flows

Changing the NLU embedding model requires all Flows in the Project to be retrained.
Otherwise, Intent recognition will fail because the new embeddings won't match the old ones used for training.

During Flow training, only enabled Intents are considered. Intents marked as disabled will be skipped.

You can train a single Flow or multiple Flows by navigating to **Build > Flows**.

## More Information

* [LLMs](/ai/agents/develop/gen-ai-and-llms/llms)
* [Intent Overview](/ai/platform-features/nlu/intents/overview)


# LLM-Based Intent Reranking
Source: https://docs.cognigy.com/ai/platform-features/nlu/external/intent-reranking



<a href="/release-notes/4.76"><Badge>Updated in 4.76</Badge></a>

*LLM-based Intent Reranking* is a process that uses a Large Language Model (LLM) to reorganize or reorder identified Intents based on data taken from the input text. This process involves using a more advanced language model to better understand the context and meaning of the user's input, which improves the accuracy of Intent classification.

Cognigy NLU analyzes user input and identifies up to 5 most probable Intents.
The result depends on various factors, including the total number of Intents, example sentences, rejected Intents, thresholds, and language specifics.
Sometimes the correct Intent is returned in the top-5 but not with the highest priority.
Adding an LLM request after the Cognigy NLU mapping process can improve accuracy.

Cognigy automatically includes descriptions of Intents to the results generated during the mapping process.
These descriptions become a part of the Input object and are stored in the `input.nlu.intentMapperResults.scores[x].description` object for each mapped Intent, where `x` represents the priority of the Intent.
The stored descriptions are sent to an external LLM via a prompt, which can be made via the LLM Prompt Node.
The external LLM analyzes the data and generates a result based on the retrieved descriptions. This result is recorded in the `intent.promptResult` object by default.

Once you have the correct Intent, you can continue the Flow based on this Intent. For example, if a user requested information about a product, you can provide product descriptions, answers to frequently asked questions, and options for placing an order.

## Prerequisites

* Add an [external LLM](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) that supports the LLM Prompt Node & LLM-powered Answer Extraction features.
* Create a Flow and Intents in it.

## How to Set up

To rerank Intents via an external LLM, follow these steps:

<AccordionGroup>
  <Accordion title="Add Descriptions to Intents">
    To add descriptions to Intents, follow these steps:

    1. In the left-side menu of the Project, go to **Build > Flows**.
    2. On the **Flows** page, select the Flow you want to change.
    3. In the upper-right corner, select **NLU**.
    4. On the **Intents** tab, locate the Intent to which you want to add a description.
    5. On the **Intent** page, in the upper-right corner, click <img alt="menu" /> **> Edit**.
    6. In the **Edit Intent** window, go to the **Description** field. Enter a clear and concise description that reflects the content of the example sentences listed in the Intent. Note that the **Description** field can contain no more than 200 characters.
    7. Click **Save**.
    8. Repeat steps 6–9 for other Intents within the Flow.
    9. In the upper-right corner, click **Build Model**.
  </Accordion>

  <Accordion title="Add an LLM Prompt Node">
    To add an LLM Prompt Node with the relevant prompt, follow these steps:

    1. In the Flow editor, add an LLM Prompt Node.
    2. In the LLM Prompt Node, go to the **Instruction (System Message/Prompt)** field and enter the following prompt for reranking:

    ```text theme={null}
    Your are an expert in intent recognition. Given a list of intents, their descriptions, and a query, you need to choose the right intent for the query.

    Given the following intents and descriptions: 

    {{input.nlu.intentMapperResults.scores.slice(0,5).map((item, index) => `Intent ${index +1}: ${item.name}\nDescription ${index + 1}: ${item.description}\n`).join("\n\n")}}

    And the following query: {{input.text}}

    Choose the correct intent for the query. Output the intent using the JSON format as follows: {"intent": intent_name}
    ```

    3. Navigate to the **Advanced** section and configure the recommended parameters as follows:<br />
       1. From the **Sampling Method** list, select **Temperature** and set the temperature to `0`.<br />
       2. From the **Response Format** list, select **JSON Object**.
    4. *(Optional)* If you want the model to return the similar result for the subsequent requests, specify a number in the **Seed** parameter. For example, `123`.
    5. Click **Save Node**.
  </Accordion>

  <Accordion title="Retrieve and Reuse Results">
    By default, the correct Intent will be stored in the `input.promptResult.intent` object.
    To retrieve this Intent for further use, follow these steps:

    1. Below the LLM Prompt Node, add a Code Node.
    2. In the **Code** field of the Code Node, specify `input.intent = input.promptResult.intent`.
    3. Click **Save Node**.
    4. Below the Code Node, add a Lookup Node.
    5. In the Lookup Node editor, configure the following fields:<br />
       1. From the **Type** list, select **Intent**.<br />
       2. Set the **Intent Level** field to **Default**, matching Intents at the most detailed hierarchy level.
    6. Click **Save Node**.
    7. Add child Case Nodes of the Lookup Node, specifying the Intent name in the **Value** field for each Node.

    Under each Case Node, you can place other Nodes, such as Say or Question, which will offer guides to the end user based on the selected Intent.
  </Accordion>
</AccordionGroup>

## Example

In this example, a user wants to order pizza and shares their preferences. The AI Agent recognizes if the user has ingredient restrictions and determines the correct Intent.

### Data

Assume that we have the following Intent data:

<AccordionGroup>
  <Accordion title="Intent Data">
    | Intent Name      | Intent Description                                                                                                                                                                                    | Sentences                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
    | ---------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Meat Pizza       | Intent for ordering a pizza with meat toppings such as pepperoni, sausage, bacon, ham, chicken, and ground beef. A popular choice for meat lovers. Rich in flavor and protein.                        | I'm craving a Meat Pizza loaded with ground beef, sausage, and ham. Could you make that for delivery?<br />For pickup, I'd like to order a Meat Pizza. Could you make it with plenty of bacon and extra cheese, please?<br />Can I get a Meat Pizza with ham and chicken as the toppings, cooked to a crispy perfection?<br />Could I please have a Meat Pizza with a mix of all your meat toppings? That's pepperoni, sausage, bacon, ham, and chicken<br />I'd like to order a large Meat Pizza with extra pepperoni, sausage, and bacon, please |
    | Vegetarian Pizza | Intent for ordering a meatless pizza. It contains milk-based products and vegetables. Often has cheese as a main ingredient, such as mozzarella, cheddar, feta, or goat cheese. Good for vegetarians. | May I please place an order for a vegetarian pizza?<br />I love the combination of fresh vegetables and creamy cheese on a vegetarian pizza<br />I'd like to order the vegetarian pizza, please. Can you ensure it has mozzarella, cheddar, or feta cheese with a variety of veggies?<br />I'm interested in the vegetarian pizza, please. Can you describe the cheese options available and the types of vegetables you use?<br />Could you prepare a vegetarian pizza for me? I prefer it with lots of milk-based cheese and assorted vegetables |
    | Veggie Pizza     | Intent for ordering a pizza with only vegetable toppings such as tomatoes, mushrooms, bell peppers, onions, olives, spinach, and broccoli. A light and refreshing option for health-conscious eaters. | I'll take a veggie pizza with a mix of tomatoes, olives, and broccoli<br />I'm in the mood for a veggie pizza loaded with bell peppers and onions<br />I'm looking forward to trying your veggie pizza with a variety of fresh vegetables<br />Could I please order a veggie pizza with extra mushrooms and spinach?<br />Do you have a special veggie pizza?                                                                                                                                                                                      |

    Consider the example of Intent reranking based on the following user input:
  </Accordion>

  <Accordion title="User Input">
    ```text theme={null}
    I'd like something low-calorie, but here's the catch — I don't eat any animal products, 
    and I'm allergic to milk-based ingredients.
    ```
  </Accordion>
</AccordionGroup>

### Result

<AccordionGroup>
  <Accordion title="Cognigy NLU">
    After applying Cognigy NLU Intent mapping, `Fallback Reject Intent`
    is assigned priority `1` with a score of `0.588120879379653`, instead of the correct Intent `Veggie Pizza`,
    which has priority `3` with a score of `0.026957729768803257`.

    ```json theme={null}
    {
      "nlu": {
        "intentMapperResults": {
          "scores": [
            {
              "id": null,
              "name": "null - Fallback Reject Intent",
              "score": 0.588120879379653,
              "negated": false,
              "confirmationSentence": null,
              "confirmationSentences": null,
              "disambiguationSentence": null,
              "flow": null,
              "description": ""
            },
            {
              "id": "8f44b7b1-d984-4da3-9558-6c128a190235",
              "name": "Vegetarian Pizza",
              "score": 0.2909909269629138,
              "negated": false,
              "confirmationSentence": null,
              "confirmationSentences": null,
              "disambiguationSentence": null,
              "flow": "f212ad29-e38e-4cfc-bc1c-47b12cfd9fc3",
              "description": "Intent for ordering a meatless pizza. It contains milk-based products and vegetables. Often has cheese as a main ingredient, such as mozzarella, cheddar, feta, or goat cheese. Good for vegetarians."
            },
            {
              "id": "a9418c1b-f956-402b-b317-cabe91950946",
              "name": "Veggie Pizza",
              "score": 0.026957729768803257,
              "negated": false,
              "confirmationSentence": null,
              "confirmationSentences": null,
              "disambiguationSentence": null,
              "flow": "f212ad29-e38e-4cfc-bc1c-47b12cfd9fc3",
              "description": "Intent for ordering a pizza with only vegetable toppings such as tomatoes, mushrooms, bell peppers, onions, olives, spinach, and broccoli. A light and refreshing option for health-conscious eaters."
            }
          ]
        }
      }
    }
    ```

    Alternatively, to view up to 5 probable results, you can output them via a [Say](/ai/agents/develop/node-reference/basic/say) Node by including the [prompt](#how-to-set-up) in the **Text** field.
  </Accordion>

  <Accordion title="Cognigy NLU + External LLM">
    After applying the Cognigy NLU Intent mapping and external LLM reranking, the correct Intent `Veggie Pizza` is identified.

    ```json theme={null}
    {
       "promptResult": {
          "intent": "Veggie Pizza"
       }
    }
    ```

    Once the AI Agent identifies the correct Intent, you can continue the Flow.
    For example, the AI Agent can offer pizza options from the identified category `Veggie Pizza`.
  </Accordion>
</AccordionGroup>

## More Information

* [Generate Intents](/ai/agents/develop/gen-ai-and-llms/generative-ai)


# Alexa
Source: https://docs.cognigy.com/ai/platform-features/nlu/external/nlu-connector-reference/alexa



<a href="/release-notes/2026.17"><Badge>Updated in 2026.17</Badge></a>

<Warning>
  The Alexa NLU connector has been deprecated. The removal is planned for January 2028. You can still create and use Alexa NLU connectors until the removal date, but no further feature development will occur. Before the removal date, migrate to an alternative NLU connector.
</Warning>

The Alexa NLU connector enables integration with [Amazon Alexa](https://www.alexa.com/) to use its language understanding results for powering AI Agents in Cognigy.

## Prerequisites

* Access to Amazon Alexa.

## Configuration on the NLU Provider Side

1. Log in to the [Amazon Developer Console](https://developer.amazon.com/alexa),
2. In the **Alexa Skills Kit** section, click **Create Skill**.
3. Select a skill type and click **Create Skill**.
4. Provide a name for your skill, for example, `Cognigy Skill`.
5. During the setup, define an Invocation Name, for example, `Cognigy Bot` that users will say to activate the skill on Alexa.
6. Select the default language and region where you want the skill to be available.
7. Configure the Intents and Slots according to the type of interaction your skill should handle. You can use built-in intents, for example, `AMAZON.HelpIntent`, or create custom intents.

## Configuration on the Cognigy Side

In **Cognigy**, go to **Build > NLU Connectors**, create an Alexa NLU connector, and configure the following settings.

| **Parameter**       | **Description**                                                                                                     |
| ------------------- | ------------------------------------------------------------------------------------------------------------------- |
| Invocation Name     | The phrase users will say to activate your Alexa skill. For example, `Alexa, ask [Skill Name] to [perform action]`. |
| Skills              | A list of skills that you configured.                                                                               |
| Reparse Alexa Slots | Allow Cognigy NLU to process the Slots provided by Alexa and extract structured data accordingly.                   |

## Test your NLU Connector

By default, the Cognigy NLU connector is used in the Interaction Panel for testing purposes.
To change this behavior, open the **Settings** tab in the Interaction Panel, then select the configured connector from the **NLU Connector** list.

## Add your NLU Connector to Endpoint Settings

By default, the Cognigy NLU connector is used in the Endpoint.
To change this behavior, follow these steps:

1. Go to **Deploy > Endpoints**.
2. Select the desired Endpoint and open the Endpoint Settings.
3. Go to **Natural Language Understanding > NLU Connector** and select the configured connector.


# NLU Connector Reference
Source: https://docs.cognigy.com/ai/platform-features/nlu/external/nlu-connector-reference/all-nlu-connectors



You can select [NLU connectors](/ai/platform-features/nlu/external/nlu-connectors/overview) of your choice, Cognigy native connectors, third-party connectors, or create a custom connector.

<CardGroup>
  <Card title="Cognigy NLU" href="/ai/platform-features/nlu/external/nlu-connector-reference/cognigy-nlu">
    The built-in NLU engine within Cognigy, providing automatic Intent and Slot recognition.
  </Card>

  <Card title="Alexa" href="/ai/platform-features/nlu/external/nlu-connector-reference/alexa">
    Integration with Amazon Alexa NLU service for voice-based input and processing.
  </Card>

  <Card title="Dialog Flow" href="/ai/platform-features/nlu/external/nlu-connector-reference/watson-assistant">
    Integration with Google Dialog Flow for advanced AI-driven conversations.
  </Card>

  <Card title="Lex" href="/ai/platform-features/nlu/external/nlu-connector-reference/lex">
    Integration with Amazon Lex, used for building conversational interfaces.
  </Card>

  <Card title="Luis" href="/ai/platform-features/nlu/external/nlu-connector-reference/luis">
    Integration with Microsoft LUIS for Intent recognition.
  </Card>

  <Card title="Watson Assistant" href="/ai/platform-features/nlu/external/nlu-connector-reference/watson-assistant">
    Integration with IBM Watson Assistant for advanced AI-driven conversations.
  </Card>

  <Card title="Code" href="/ai/platform-features/nlu/external/nlu-connector-reference/code">
    Allows custom integration with any third-party NLU provider.
  </Card>

  <Card title="No NLU" href="/ai/platform-features/nlu/external/nlu-connector-reference/no-nlu">
    Option for users who don't require NLU processing.
  </Card>
</CardGroup>


# Code
Source: https://docs.cognigy.com/ai/platform-features/nlu/external/nlu-connector-reference/code



The Code connector allows
you to integrate an NLU engine that isn't natively supported by Cognigy
or when you need a highly customized integration with a supported NLU engine
that goes beyond the default connector's capabilities.

To use this connector, follow these steps:

1. Go to **Build > NLU Connectors** and create a Code connector.
2. Add your JavaScript code to the **Transformer** section and activate the **Enable Code Transformer** toggle.

<Accordion title="Code Example with Description">
  ```js theme={null}
  createNluCodeConnector({

    /**
     * This function connects to an external NLU service, for example, Dialogflow, LUIS.
     * and returns the processed result.
     * @param text The message text from the user.
     * @param data Any additional data passed with the message.
     * @param language The language of the input message.
     * @returns nluResult Processed result from the external NLU service.
     */
    getNluResult: async ({ text, data, language }) => {

      // Example: Make an HTTP request to an external NLU API, for example, Dialogflow, LUIS.
      const response = await fetch('https://api.externalnlu.com/analyze', {
        method: 'POST',
        headers: {
          'Content-Type': 'application/json',
          'Authorization': 'Bearer YOUR_API_KEY'
        },
        body: JSON.stringify({
          query: text,
          language: language,
        }),
      });

      // Parse the response from the NLU API
      const nluResponse = await response.json();

      // Construct the NLU result based on the response from the NLU service
      let nluResult = {
        intent: nluResponse.intent || null,        // Extracted intent from the NLU response
        type: nluResponse.type || null,            // The type of the Intent (optional)
        slots: nluResponse.entities || {},         // Extracted entities as Slots (parameters)
        intentScore: nluResponse.intentConfidence || 0, // Confidence score of the Intent
        nlu: {
          intentId: nluResponse.intentId || null,  // Unique ID for the detected Intent
          intentFlow: nluResponse.intentFlow || null // Flow associated with the Intent (if available)
        }
      };

      // Return the transformed data and NLU result
      return {
        data,
        nluResult
      };
    }
  });
  ```

  |                                 | **Description**                                                                                                                                                                                                                                                                                                                                         |
  | ------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
  | **Function and its parameters** |                                                                                                                                                                                                                                                                                                                                                         |
  | `createNluCodeTransformer`      | This function creates a transformer that asynchronously sends the message (`text`) to an external NLU service for analysis. The function processes the response from the service, populates the `nluResult` object with values such as `intent`, `type`, `slots`, and `intentScore`, and returns the data object unchanged, unless explicitly modified. |
  | `text`                          | The message received from the endpoint, such as user input, that needs to be analyzed by the NLU service.                                                                                                                                                                                                                                               |
  | `data`                          | The data object associated with the message, containing additional metadata or context about the request. This object can be modified or passed through unchanged.                                                                                                                                                                                      |
  | `language`                      | The language code of the locale, representing the language in which the message is processed. For example, `en` for English, `de` for German.                                                                                                                                                                                                           |
  | **NLU Result Initialization**   | Initial values set for the NLU result object to ensure structure before any processing occurs.                                                                                                                                                                                                                                                          |
  | `intent`                        | Initially `null`, it will be populated with the Intent detected from the message, for example, `book_flight` or `greet_user`, after processing.                                                                                                                                                                                                         |
  | `type`                          | Initially `null`, it refers to the type of Intent, for example, `action`, `informational`. Will be set based on the NLU service's response.                                                                                                                                                                                                             |
  | `slots`                         | Initially `{}`, this object holds the extracted parameters (Slots) that are associated with the detected intent, for example, destination, date, and class.                                                                                                                                                                                             |
  | `intentScore`                   | Initially `0`, it represents the confidence score (ranging from 0 to 1) indicating how confident the NLU service is in the detected Intent.                                                                                                                                                                                                             |
  | `nlu`                           | An object containing metadata about the Intent: <br /> - `intentId` - initially `null`, the unique identifier for the detected Intent. <br /> - `intentFlow` - initially `null`, the Flow associated with the Intent.                                                                                                                                   |
  | **Return Object**               | The function returns the `data` and `nluResult`.                                                                                                                                                                                                                                                                                                        |
  | `data`                          | The unchanged data object passed in, which may contain context or additional information.                                                                                                                                                                                                                                                               |
  | `nluResult`                     | The object returned after processing, containing the final `intent`, `slots`, `intentScore`, and any other details extracted from the message using the NLU logic.                                                                                                                                                                                      |
</Accordion>

## More Information

* [All NLU Connectors](/ai/platform-features/nlu/external/nlu-connector-reference/all-nlu-connectors)


# Cognigy NLU
Source: https://docs.cognigy.com/ai/platform-features/nlu/external/nlu-connector-reference/cognigy-nlu



The Cognigy NLU is the native NLU connector integrated into the Cognigy platform.
This connector is selected by default in both the Endpoint Settings and the Interaction Panel.

To customize its behavior with [Transformers](/ai/platform-features/nlu/external/nlu-connectors/nlu-transformers), follow these steps:

1. Go to Build > NLU Connectors.
2. Create a new Cognigy NLU Connector.
3. Update the transformer code to suit your needs.

## More Information

* [All NLU Connectors](/ai/platform-features/nlu/external/nlu-connector-reference/all-nlu-connectors)


# Dialogflow
Source: https://docs.cognigy.com/ai/platform-features/nlu/external/nlu-connector-reference/dialog-flow



<a href="/release-notes/2026.17"><Badge>Updated in 2026.17</Badge></a>

<Warning>
  The Dialogflow NLU connector has been deprecated. The removal is planned for January 2028. You can still create and use Dialogflow NLU connectors until the removal date, but no further feature development will occur. Before the removal date, migrate to an alternative NLU connector.
</Warning>

The Dialogflow connector enables integration with [Google Dialogflow](https://cloud.google.com/dialogflow/docs) to use its language understanding results for powering AI Agents in Cognigy.

## Prerequisites

* Access to Dialogflow.

## Configuration on the NLU Provider Side

1. Go to [Dialogflow](https://dialogflow.cloud.google.com/) and create a new agent.
2. In the Google Cloud Console, go to **APIs & Services > Enable APIs and Services** and enable the **Dialogflow API**.
3. Define your intents and entities for the agent in Dialogflow.
4. In **Google Cloud Console**, go to **IAM & Admin > Service Accounts**, create a service account, and generate a **JSON Private Key**.

## Configuration on the Cognigy Side

In **Cognigy**, go to **Build > NLU Connectors**,
create a **Dialogflow NLU Connector**, then select the Dialogflow version.

<Tabs>
  <Tab title="v1">
    | **Parameter** | **Description**                      |
    | ------------- | ------------------------------------ |
    | Access Token  | A token used for API authentication. |
  </Tab>

  <Tab title="v2">
    | **Parameter**         | **Description**                                                          |
    | --------------------- | ------------------------------------------------------------------------ |
    | Dialogflow Project ID | The unique identifier for your Dialogflow project.                       |
    | Private Key           | The JSON credentials file containing the private key for authentication. |
  </Tab>
</Tabs>

## More Information

* [All NLU Connectors](/ai/platform-features/nlu/external/nlu-connector-reference/all-nlu-connectors)
* [Cognigy Help Center: Dialogflow](https://support.cognigy.com/hc/en-us/articles/360017466620-Dialogflow-External-NLU)


# Amazon Lex
Source: https://docs.cognigy.com/ai/platform-features/nlu/external/nlu-connector-reference/lex



The Amazon Lex NLU connector enables integration with [Amazon Lex](https://aws.amazon.com/lex/) to use its language understanding results for powering AI Agents in Cognigy.

## Prerequisites

* Access to Amazon Lex.

## Configuration on the NLU Provider Side

1. Log in to the **AWS Console**, then go to the **Lex Start page** to create a new bot.
2. Make sure to configure the languages that the bot will support, based on the Locale in Cognigy. This step is important for proper integration.
3. After configuring the bot, save your changes and build the bot. Ensure all languages are built, especially if you intend to use multiple languages in Cognigy.
4. In the **Lex Console**, navigate to **Bot versions** and create a new version for your bot.
5. Create a **Bot Alias** and associate it with the new version. Copy the **Bot ID** and **Bot Alias ID** from the URL for later use in Cognigy.
6. Create a new **IAM user** in the **AWS IAM Console**, ensuring the user has the appropriate permissions. Make sure the AWS user has the following permissions:

   <Accordion title="User Permissions">
     ```txt theme={null}
     iam:CreateUser
     iam:CreatePolicy
     iam:AttachUserPolicy
     lex:ListBotAliases
     lex:ListBotLocales
     lex:ListBots
     lex:ListBotVersions
     lex:ListBuiltInIntents
     lex:ListBuiltInSlotTypes
     lex:ListSlots
     lex:ListSlotTypes
     lex:DescribeBot
     lex:DescribeBotAlias
     lex:DescribeBotLocale
     lex:DescribeBotVersion
     lex:DescribeSlot
     lex:DescribeSlotType
     lex:CreateBot
     lex:CreateBotAlias
     lex:CreateBotLocale
     lex:CreateBotVersion
     lex:CreateIntent
     lex:CreateSlot
     lex:RecognizeText
     lex:StartConversation
     ```
   </Accordion>
7. Once the user is created, retrieve the **Access Key ID** and **Secret Access Key**.

## Configuration on the Cognigy Side

1. In **Cognigy**, go to **Build > NLU Connectors** and create an Amazon Lex NLU connector.
2. *(Optional)* Toggle **Reparse Lex Slots** if you want to use reparse Lex Slots.
3. Enter the retrieved credentials.

   | **Parameter**     | **Description & How to Find**                                                |
   | ----------------- | ---------------------------------------------------------------------------- |
   | Access Key        | The Access Key ID you retrieved when creating the IAM user.                  |
   | Secret Access Key | The Secret Access Key you retrieved when creating the IAM user.              |
   | AWS Region        | The AWS region in which your Lex bot was created (for example, `us-east-1`). |
   | Bot ID            | The Bot ID copied from the URL in the Lex Console.                           |
   | Bot Alias ID      | The Bot Alias ID copied from the URL in the Lex Console.                     |

## Test your NLU Connector

By default, the Cognigy NLU connector is used in the Interaction Panel for testing purposes.
To change this behavior, open the **Settings** tab in the Interaction Panel, then select the configured connector from the **NLU Connector** list.

## Add your NLU Connector to Endpoint Settings

By default, the Cognigy NLU connector is used in the Endpoint.
To change this behavior, follow these steps:

1. Go to **Deploy > Endpoints**.
2. Select the desired Endpoint and open the Endpoint Settings.
3. Go to **Natural Language Understanding > NLU Connector** and select the configured connector.

## More Information

* [All NLU Connectors](/ai/platform-features/nlu/external/nlu-connector-reference/all-nlu-connectors)
* [Cognigy Help Center: Amazon Lex](https://support.cognigy.com/hc/en-us/articles/4408910240530-Amazon-Lex-External-NLU)


# Microsoft LUIS
Source: https://docs.cognigy.com/ai/platform-features/nlu/external/nlu-connector-reference/luis



The Microsoft LUIS NLU connector enables integration with [Microsoft LUIS](https://www.luis.ai/) to use its language understanding results for powering AI Agents in Cognigy.

## Prerequisites

* Access to Microsoft LUIS.

## Configuration on the NLU Provider Side

1. Set up a new LUIS app in the **Microsoft LUIS portal**, define intents, add entities, and train the model.
2. Go to **Manage > Azure Resources** in your LUIS app to find the required parameters:
   * **Location** — the Azure region where the LUIS service is hosted. For example, `westus2`.
   * **Subscription Key** — the key (Primary or Secondary key) can be found in **Manage > Azure Resources**.
   * **App ID** — the App ID can be found in **Manage > Settings** of the LUIS app.
3. Ensure the app is published in **production mode** for external services, such as Cognigy, to access it.

## Configuration on the Cognigy Side

1. In Cognigy, go to **Build > NLU Connectors**, create a LUIS NLU connector.
2. Fill in the placeholders in the URL with the parameters you found in LUIS:

   ```txt theme={null}
   https://{location}.api.cognitive.microsoft.com/luis/v2.0/apps/{appId}?subscription-key={subscriptionKey}&q=
   ```
3. Paste this URL into the **Authentication URL** field.

## Test your NLU Connector

By default, the Cognigy NLU connector is used in the Interaction Panel for testing purposes.
To change this behavior, open the **Settings** tab in the Interaction Panel, then select the configured connector from the **NLU Connector** list.

## Add your NLU Connector to Endpoint Settings

By default, the Cognigy NLU connector is used in the Endpoint.
To change this behavior, follow these steps:

1. Go to **Deploy > Endpoints**.
2. Select the desired Endpoint and open the Endpoint Settings.
3. Go to **Natural Language Understanding > NLU Connector** and select the configured connector.

## More Information

* [All NLU Connectors](/ai/platform-features/nlu/external/nlu-connector-reference/all-nlu-connectors)
* [Cognigy Help Center: Microsoft LUIS](https://support.cognigy.com/hc/en-us/articles/360017517319-Microsoft-LUIS-External-NLU)


# No NLU
Source: https://docs.cognigy.com/ai/platform-features/nlu/external/nlu-connector-reference/no-nlu



The *No NLU* connector disables built-in Cognigy NLU for the selected Endpoint, meaning no Intents or Slots are detected.

The No NLU connector lets you handle the NLU process directly within your Flow logic for the following purposes:

* You need very fine-grained control over how NLU requests are made.
* You want to dynamically switch between different NLU services based on the conversation context.

## Use Flow Logic for NLU Execution

Within a Flow, you can control NLU execution by enabling the No NLU connector and using the following Nodes:

<AccordionGroup>
  <Accordion title="HTTP Request Node">
    Use an [HTTP Request](/ai/agents/develop/node-reference/service/http-request) Node
    to send the user input to your external NLU service
    and then process the response.
  </Accordion>

  <Accordion title="Extension Nodes">
    You can put the HTTP request and response logic in an [Extension Node](/ai/for-developers/extensions) to avoid repeating it in every Flow. This approach makes your Flows cleaner, easier to maintain, and avoids code duplication. So that the Extension Node becomes your custom NLU connector on the Flow level, even though you're not using the built-in [Code NLU](/ai/platform-features/nlu/external/nlu-connector-reference/code) connector.
  </Accordion>

  <Accordion title="Execute Cognigy NLU Node">
    By default, Cognigy NLU runs before the Flow executes. However, in some cases, you may need to process parts of the Flow before triggering NLU analysis to better control when an Intent is recognized.
    To achieve this, add an [Execute Cognigy NLU](/ai/agents/develop/node-reference/ai/execute-cognigy-nlu) Node to your Flow to trigger Cognigy NLU when needed.
  </Accordion>
</AccordionGroup>

## More Information

* [All NLU Connectors](/ai/platform-features/nlu/external/nlu-connector-reference/all-nlu-connectors)


# Watson Assistant
Source: https://docs.cognigy.com/ai/platform-features/nlu/external/nlu-connector-reference/watson-assistant



The Watson Assistant NLU connector enables integration with [IBM Watson Assistant](https://www.ibm.com/products/watsonx-assistant) to use its language understanding results for powering AI Agents in Cognigy.

## Prerequisites

* Access to IBM Watson Assistant.

## Configuration on the NLU Provider Side

1. Log in to [IBM Cloud](https://cloud.ibm.com), create a **Watson Assistant** service, and find the **Service URL** under **Credentials**.
2. In **Watson Assistant**, go to **Skills > Create Skill**, then select a skill type. Note that Action skills aren't supported for API connections.
3. *(Optional)* In the skill editor, go to **Entities > System Entities** and enable `@sys-date`, `@sys-number`, and `@sys-time`.
4. In **Skills Overview**, click the three dots next to your skill, select **View API Details**, and copy the **API Key** and **Skill ID**.

## Configuration on the Cognigy Side

In Cognigy, go to **Build > NLU Connectors**, create a Watson Assistant NLU connector, and enter the retrieved credentials.

| **Parameter**           | **Description**                                                                                                                                         |
| ----------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------- |
| API Key                 | An authentication key. Find this key in **Service Credentials** in IBM Cloud under Watson Assistant service.                                            |
| Skill ID (Workspace ID) | A unique skill identifier. Find this identifier by clicking the three dots next to the skill in **Skills Overview** and selecting **View API Details**. |
| Service URL             | An endpoint for accessing Watson Assistant API. Find this URL in **Service Credentials** under the Watson Assistant service.                            |

## Test your NLU Connector

By default, the Cognigy NLU connector is used in the Interaction Panel for testing purposes.
To change this behavior, open the **Settings** tab in the Interaction Panel, then select the configured connector from the **NLU Connector** list.

## Add your NLU Connector to Endpoint Settings

By default, the Cognigy NLU connector is used in the Endpoint.
To change this behavior, follow these steps:

1. Go to **Deploy > Endpoints**.
2. Select the desired Endpoint and open the Endpoint Settings.
3. Go to **Natural Language Understanding > NLU Connector** and select the configured connector.

## More Information

* [All NLU Connectors](/ai/platform-features/nlu/external/nlu-connector-reference/all-nlu-connectors)
* [Cognigy Help Center: Watson Assistant](https://support.cognigy.com/hc/en-us/articles/360017484979-Watson-Assistant-External-NLU#requirements-0-0)


# NLU Transformers
Source: https://docs.cognigy.com/ai/platform-features/nlu/external/nlu-connectors/nlu-transformers



*NLU Transformers* are used to customize the NLU Connectors pipeline.

With transformers, you can:

* Modify the incoming message from the Endpoint before it's sent to the NLU engine.
* Adjust the response from the NLU engine before it reaches the Flow execution.

NLU transformers are available for Cognigy NLU, as well as for Alexa, DialogFlow, Microsoft LUIS, and Watson Assistant connectors.

## Restrictions

* NLU transformers operate only on data exchanged with the NLU engine. They don't modify the prompt or response of an LLM Prompt Node or AI Agent Node.

## NLU Code transformer

With the NLU Code transformer, you can write a custom NLU pipeline in the Code Editor.

| **Function** | **Parameters**                                 | **Description**                                                                                                                                                                                                                       | **Return**                                                       |
| ------------ | ---------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------- |
| `preNlu`     | `text`, `data`, `language`                     | The `preNlu` transformer function processes data before sending it to the NLU engine. This function can perform tasks such as stop word removal, punctuation cleanup, noise reduction, text normalization, and spell checking.        | Returns modified `text` and `data` to be sent to the NLU engine. |
| `postNlu`    | `text`, `data`, `nluResult`, `connectorOutput` | The `postNlu` transformer function works with the raw NLU results. `connectorOutput` is the unfiltered output from the NLU engine, for example, DialogFlow, LUIS, or Watson, while `nluResult` contains the filtered and mapped data. | Returns modified `data` and `nluResult` to process the Flow.     |

## Transformer Settings

You can control the NLU Code transformer's behavior by enabling or disabling functions in the Transformer Settings.

| **Setting**                   | **Description**                                                                                                                                                                                                                                                                                     |
| ----------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Abort on Error in Transformer | Stops the execution of the transformer if an error occurs, preventing further processing. By default, transformer errors are logged on the Logs page, and execution continues normally. This means that if the transformer fails, the message is still processed as if the transformer didn't exist |
| Enable pre NLU Transformer    | Activates the `pre` function in the transformer before the NLU engine processes the input, allowing for pre-processing.                                                                                                                                                                             |
| Enable post NLU Transformer   | Activates the `post` function in the transformer after the NLU engine processes the input, allowing for post-processing.                                                                                                                                                                            |
| Debug Transformer Stack       | Displays the input and output data from each transformer in the Input object within the Interaction Panel, helping you track how data changes at each transformation step. This setting doesn't apply to Endpoints.                                                                                 |

<Accordion title="Example of the `transformerStack` object in the Interaction Panel">
  ```json theme={null}
  {
    "transformerStack": {
      "preNluInput": {
        "text": "hi!!!"
      },
      "preNluOutput": {
        "text": "hi",
        "data": {}
      },
      "postNluInput": {
        "text": "hi",
        "data": {},
        "nluResult": {
          "intent": "",
          "intentScore": 0,
          "slots": {},
          "type": "Greeting"
        },
        "connectorOutput": null
      },
      "postNluOutput": {
        "data": {},
        "nluResult": {
          "intent": "myIntent",
          "intentScore": 1,
          "slots": {},
          "type": "Greeting"
        }
      }
    }
  }
  ```
</Accordion>

## Available Modules

<AccordionGroup>
  <Accordion title="httpRequest">
    <Warning>
      * You can send one HTTP request per transformer execution.
      * You can't send HTTP requests to other Endpoints.
    </Warning>

    You can send HTTP requests to external services, for example, third-party NLU engines, from within a transformer.
    All transformer functions can access the `httpRequest` module for this purpose.
    The module is invoked as a function which takes a single configuration object as arguments.
    This object has the same format as the options object for the NPM module [Request](https://www.npmjs.com/package/request#requestoptions-callback).

    Example:

    ```js theme={null}
    handleInput: async () => {
      const result = await httpRequest({
        uri: "some-uri",
        method: "POST",
        body: {},
        json: true
      });

      console.log(result);
      return result;
    }
    ```
  </Accordion>

  <Accordion title="getCognigyNlu">
    You can execute Cognigy NLU from within a transformer. All NLU transformer functions can access the `getCognigyNlu` function. This function takes string as arguments.

    Example:

    ```js theme={null}
    const cognigyNluResult = await getCognigyNlu(text + "Lorem ipsum dolor sit amet, consectetur adipiscing elit.");
    ```
  </Accordion>

  <Accordion title="getSessionStorage">
    `SessionStorage` is a storage object available to all transformer functions and stays active throughout the session.
    This object can store variables for later use in other functions or track the conversation state.
    You can access it using the `getSessionStorage` function.
    This function returns a Promise that resolves to the `sessionStorage` object, which you can use as a JavaScript regular object.

    Modifying variables inside `StorageSession` can lead to several problems:

    * **Inconsistent State** You might change the object in a way that other parts of your application don't expect, which can cause errors or unexpected results.
    * **Asynchronous Issues**. Since JavaScript is asynchronous, other operations might try to access the session storage before your changes are applied, causing issues such as reading outdated data.

    Instead, make changes locally (in a temporary variable or copy) and then assign the updated value back to `SessionStorage`.

    Example:

    ```js theme={null}
    // Retrieve the sessionStorage object
    const sessionStorage = await getSessionStorage(); 

    // Make a copy of the array from sessionStorage (or initialize if it doesn't exist)
    let tempArr = sessionStorage.arr || []; // Default to an empty array if 'arr' doesn't exist

    // Modify the local copy (add new items to the array)
    tempArr.push(1);  // Add item 1
    tempArr.push(2);  // Add item 2

    // Save the modified array back to sessionStorage
    sessionStorage.arr = tempArr;  // Update sessionStorage with the new array
    ```
  </Accordion>

  <Accordion title="NPM Modules">
    The following NPM modules are available in the transformer:

    | **Module**  | **Description**                                                                     |
    | ----------- | ----------------------------------------------------------------------------------- |
    | uuid        | Generates unique identifiers (UUIDs) for items or sessions in applications.         |
    | crypto      | Provides cryptographic functions like hashing, encryption, and decryption.          |
    | \_ (Lodash) | A utility library offering functions for manipulating arrays, objects, and strings. |
    | moment      | A library for parsing, manipulating, and displaying dates and times in JavaScript.  |
  </Accordion>

  <Accordion title="Logging">
    You can use the exposed modules `console.log`, `console.info`, and `console.error` in the transformers to display logs on the [Logs page](/ai/agents/test/logs).
  </Accordion>
</AccordionGroup>

## More Information

* [Overview](/ai/platform-features/nlu/external/nlu-connector-reference/all-nlu-connectors)


# NLU Connectors
Source: https://docs.cognigy.com/ai/platform-features/nlu/external/nlu-connectors/overview



*NLU Connectors* allow Cognigy to connect with external Natural Language Understanding (NLU) providers. These connectors enable Cognigy to send user input to third-party NLU providers. The providers process the input to identify user intentions and extract relevant data.

The connectors ensure external NLU systems return results in a standardized format. This format allows easy switching between NLU providers without modifying Flows in Cognigy.

You can select the [NLU connectors of your choice](/ai/platform-features/nlu/external/nlu-connector-reference/all-nlu-connectors), including native connectors and third-party connectors, or create a custom connector.

## Key Benefits

* **Enhanced Data Privacy**. NLU connectors enable preprocessing of input text, allowing sensitive information to be masked or removed before it reaches the Intent Trainer. NLU connectors prevent sensitive data from being stored or used for training, protecting user privacy.
* **Improved Intent Training**. By preprocessing the input text and removing irrelevant or inappropriate data, NLU connectors ensure that the Intent Trainer receives cleaner, more focused data.
* **Flexible Slot Management**. NLU connectors provide a mechanism for customizing and adapting Slot mechanics, such as date handling. This mechanism allows you to fine-tune how Slots are extracted and used within the conversation, addressing specific requirements or edge cases.

## Restrictions

* When using external NLU connectors, API quota limits, additional terms, conditions, or charges may apply.

## Working with NLU Connectors

<Tabs>
  <Tab title="GUI">
    You can view, create, export, and delete NLU Connectors in **Build > NLU Connectors**.
  </Tab>

  <Tab title="API">
    You can view, create, update, and delete NLU Connectors using the [Cognigy API](https://api-trial.cognigy.ai/openapi#get-/v2.0/nluconnectors).
  </Tab>
</Tabs>

### NLU Transformers

Customize the pipeline in NLU Connectors by using the [NLU Transformers](/ai/platform-features/nlu/external/nlu-connectors/nlu-transformers).
With transformers, you can:

* Modify the incoming message from the Endpoint before it's sent to the NLU engine.
* Adjust the response from the NLU engine before it reaches the Flow execution.


# External NLU
Source: https://docs.cognigy.com/ai/platform-features/nlu/external/overview



With external NLU, you can connect external NLU providers to your AI Agents, as well as use Intent reranking and Intent recognition to improve accuracy and flexibility.

<CardGroup>
  <Card title="NLU Connectors" href="/ai/platform-features/nlu/external/nlu-connectors/overview">
    Connect third-party NLU providers to enhance Intent recognition capabilities.
  </Card>

  <Card title="NLU Connector Reference" href="/ai/platform-features/nlu/external/nlu-connector-reference/all-nlu-connectors">
    Explore configuration details and supported providers for external NLU integration.
  </Card>

  <Card title="Intent Reranking" href="/ai/platform-features/nlu/external/intent-reranking">
    Improve Intent accuracy by dynamically adjusting ranking based on confidence scores.
  </Card>

  <Card title="External NLU Intent Recognition" href="/ai/platform-features/nlu/external/external-nlu-intent-recognition">
    Use LLMs for intent recognition to enable faster setup in complex workflows and reduce manual refinement.
  </Card>
</CardGroup>


# Annotations
Source: https://docs.cognigy.com/ai/platform-features/nlu/intents/annotations



*Annotations* enable you to tag specific words within example sentences as placeholders.
These placeholders, linked to Slots, allow the NLU model to identify and extract key details from user inputs.
For example, in the sentence `I want a shirt`, you could annotate `shirt` as a placeholder tied to a `product` Slot,
helping the NLU recognize similar items, such as `pants` or `hats`.

## Key Benefits

* **Improved Detail Recognition**. Annotations help the NLU pinpoint and extract specific information, for example, names, dates, or products, from user inputs, making conversations more precise and context-aware.
* **Reduced Repetition**. By integrating Slots into example sentences, annotations eliminate the need to write repetitive variations.

## How to Use

To use Annotations, follow these steps:

1. In the Flow editor, go to **NLU > Intents** and select the desired Intent.
2. Highlight a word in an example sentence within the text field and click **+**. A row with two lists will appear.
3. Select **Slot Type** and **key** from the lists.

Multiple annotations can be added to a single sentence, each underlined in a unique color.

<Frame>
  <img alt="Intent example sentence with annotated Slot types highlighted" />
</Frame>

### Exact Matching

If two sentences are nearly identical but differ only in Keyphrases that share the same Slot, the NLU model detects them as highly similar.
In such cases, it automatically assigns a score of 0.95 to indicate a strong match.

#### Example

Example Sentence: `I want a burger`<br />
User Input: `I want a pizza`<br />
Result: If `pizza` and `burger` are both tagged with the same `product` Slot in the Lexicon, the match score will be `0.95`.

### Recommendations

* **Annotate whole words or phrases**. When you annotate words or phrases by using a Lexicon Slot in an Example Sentence, annotate them as a whole to ensure correct matching. For example:

  * In the sentence `I want a monthly subscription`
    * If you assign only `month` to a Lexicon Slot, the NLU engine doesn't recognize `monthly`.
    * If you assign `monthly` to the Lexicon Slot, the NLU engine recognizes the Intent correctly.
  * In the sentence `Ich möchte ein monatliches Abonnement abschließen`
    * If you assign only `monatlich` to a Lexicon Slot, the NLU engine doesn't recognize `monatliches`.
    * If you assign `monatliches` to the Lexicon Slot, the NLU engine recognizes the Intent correctly.

* **Avoid selecting bordering spaces**. When highlighting a word for annotation, make sure you select only the word itself without leading or trailing spaces. For example, select `monthly` instead of ` monthly` or `monthly `. Including extra spaces in the selection can cause the NLU processing to fail.

* **Annotate only 50% of your example sentences with Slots for optimal Intent matching**. This balanced approach instructs the NLU model that the Slot is relevant to the Intent, while also indicating that identifying the Slot isn't mandatory for a successful Intent match.

* **Use common words instead of rare ones**. Replace hard-to-understand terms with simpler ones to help the NLU model understand the context better. For example, instead of `PremiumPayID`, you can use `credit card`. Alternatively, you can include `credit card` in the example sentence and annotate it with a [Lexicon Slot](/ai/platform-features/nlu/slots/user-defined/lexicon) that maps to `PremiumPayID`.

## More Information

* [Example Sentences](/ai/platform-features/nlu/intents/example-sentences)
* [Intents](/ai/platform-features/nlu/intents/overview)


# Intent Conditions
Source: https://docs.cognigy.com/ai/platform-features/nlu/intents/conditions



*Intent Conditions* allow you to dynamically enable or disable a specific Intent based on real-time data. You can define them using [CognigyScript](/ai/platform-features/cognigyscript) and extend them with JavaScript.

By default, AI Agents evaluate all available Intents. However, in some cases, certain Intents should only be considered under specific conditions. For example:

* The `Cancel Credit Card` Intent should be recognized if the user has a credit card.
* The `Schedule Appointment` Intent should be recognized during business hours.

With Intent conditions, you can control when an Intent is active, improving the accuracy and relevance of AI Agent responses.

## Key Benefits

* **Granular control**. Compared to States, which are designed for managing a set of Intents, Intent conditions offer more fine-grained control over individual Intents. They allow you to enable or disable specific Intents based on multiple factors, providing greater flexibility in handling dynamic scenarios.
* **Improved AI Agent Precision**. Prevents the AI from triggering irrelevant Intents, reducing false positives and improving the user experience.

## How to Use

To set a condition for a specific Intent, navigate to **NLU > Intents** in the Flow editor.
Open the Intent, then go to **Advanced > Condition** to configure the condition.

If no condition is set, the Intent is always available.

### Syntax

You can define Intent conditions using CognigyScript:

<AccordionGroup>
  <Accordion title="Single Condition">
    A single Intent condition is an expression that evaluates to `true` or `false`. The AI Agent will detect the Intent only if the condition is `true`.

    The `!!` operator ensures the value is evaluated as a boolean.

    **Examples:**

    Check if the Keyphrase of the first item in the `city` Slot is equal to `Düsseldorf`:

    ```js theme={null}
    !!(input.slots.city[0].keyphrase === "Düsseldorf")
    ```

    Check if the `topics` array in the Context object doesn't include the term `Cognigy`:

    ```js theme={null}
    !!(!context.topics.includes("Cognigy"))
    ```
  </Accordion>

  <Accordion title="Multiple Conditions">
    To apply multiple Intent conditions, combine them using logical operators:

    * **AND (`&&`)** – requires all conditions to be true.
    * **OR (`||`)** – requires at least one condition to be true.

    **Examples:**

    Enable Intent if the user asks about Düsseldorf and the topic is Cognigy:

    ```js theme={null}
    input.slots.city[0].keyphrase === "Düsseldorf" && context.topics.includes("Cognigy")
    ```

    Enable Intent if the user asks about Düsseldorf or Cognigy:

    ```js theme={null}
    input.slots.city[0].keyphrase === "Düsseldorf" || context.topics.includes("Cognigy")
    ```
  </Accordion>
</AccordionGroup>

## Use Cases

<AccordionGroup>
  <Accordion title="Restricting an Intent Based on User Profile">
    A bank AI Agent should allow users to cancel their credit cards **only if they have one**.

    **Condition:**

    ```js theme={null}
    !!(profile.has_credit_card)
    ```

    **Result:** the Intent is only active for users with a credit card.
  </Accordion>

  <Accordion title="Checking Multiple Profile Attributes">
    The AI Agent should allow credit card cancellation only if the user owns a card that is activated.

    **Condition:**

    ```js theme={null}
    profile.has_credit_card === true && profile.credit_card_activated === true
    ```

    **Result:** the Intent is enabled only for customers with an active credit card.
  </Accordion>
</AccordionGroup>

## More Information

* [Intent Rules](/ai/platform-features/nlu/intents/rules)


# Confirmation Words
Source: https://docs.cognigy.com/ai/platform-features/nlu/intents/confirmation-words



*Confirmation Words* let you define a custom list of words and phrases that the AI Agent recognizes as `yes` or `no` responses.

This feature is useful for straightforward use cases where you need exact control over how the system recognizes affirmations
(`yes`, `yep`) or negations (`no`, `nah`).

Unlike machine learning-based Intent matching, confirmation words rely on explicit, user-defined terms.
The AI Agent matches user input directly against your list, ensuring predictable behavior without requiring extensive training.

## Key Benefits

* **Simple and precise**. Confirmation words work well for small, predictable sets of responses.
* **Fully customizable**. You decide which words trigger `yes` or `no`.

## Restrictions

* The AI Agent recognizes only exact matches or predefined synonyms from the Confirmation Words list.
* The AI Agent doesn't adapt to new confirmation words variations unless you add them yourself.

## How to Use

You can apply confirmation words globally across all Flows in a Project or customize them for a specific Flow.

<Tabs>
  <Tab title="Project-Level Settings">
    To use confirmation words for all Flows, do the following:

    1. In the left-side menu of the Project, go to **Manage > Settings**.
    2. On the **Settings** page, click **NLU Settings**.
    3. In the **General Flow Logic** section, select **Confirmation Words**.
    4. Enter the `positive` or `negative` confirmation words, one per line. Save changes.
  </Tab>

  <Tab title="Flow-Level Settings">
    By default, all Flows inherit the **Confirmation Words** and **Threshold** settings from the Project-level configuration. You can override these settings for a specific Flow.
    To use confirmation words for a specific Flow, do the following:

    1. Open the existing Flow.
    2. In the upper-right corner of the **Flow Editor** page, select **Settings > Configuration**.
    3. In the **General Flow Logic** section, select **Confirmation Words**.
    4. Enter the `positive` or `negative` confirmation words, one per line. Save changes.
  </Tab>
</Tabs>

## How to Test

Check if confirmation words work as expected via the [Interaction Panel](/ai/agents/test/interaction-panel/overview):

1. In the existing Flow, open the Interaction Panel by clicking <img alt="interaction-panel" /> **Chat with your Agent** in the upper-right corner of the page.
2. In the Flow editor, add a Question Node and select the **Yes/No** input type.
3. Run the Flow by typing a test response including a word from your confirmation list, for example, `yes` or `nope`.
4. On the **Info** tab, select **Input** to view the Input object in JSON format. In the `input.type` object, you will see the answer depending on the user input.

   Positive Confirmation:

   ```json theme={null}
   "type": "pAnswer",
   ```

   Negative Confirmation:

   ```json theme={null}
   "type": "nAnswer",
   ```

## More Information

[Yes/No Intents](/ai/platform-features/nlu/intents/yes-no-intents)


# Default Reply
Source: https://docs.cognigy.com/ai/platform-features/nlu/intents/default-reply



The *Default Reply* feature provides a configurable,
Intent-specific response triggered when the NLU model detects a matching Intent from the user input.

## Key Benefits

* **Instant Response Delivery**. A predefined reply, for example, `Thank you for your order`, is sent as soon as the Intent is recognized, improving the user experience with prompt feedback.
* **Flexible Flow Control**. You can decide whether the Flow proceeds or pauses after the default reply, offering adaptability to different use cases.
* **Enhanced Understanding of User Queries**. With the default reply providing additional context, the AI Agent can better interpret diverse user inputs, leading to improved recognition accuracy and more effective responses.

## Prerequisites

Before using the default reply, ensure the following:

1. A Flow is created in Cognigy with at least one defined Intent.
2. Access to the NLU settings within the Flow to configure the default reply.

## Restriction

* The default reply is tied to specific Intents and requires manual configuration. It won't activate if it is overridden by other settings like attached Flows or Question Node validations.

## How to Use

Add the default reply to a specific Intent.
If you've configured the default reply for multiple Intents, adjust the settings at the Flow level to ensure consistent default reply behavior.

### Add the Default Reply

1. In the Flow editor, go to **NLU > Intents** and select an Intent.
2. Go to the **Default Reply** section and configure the following settings:

| **Setting**                           | **Description**                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
| ------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| Include Default Reply in NLU Training | Defines whether the default reply should be used in the NLU model training. You can select one of the following options:<br />- **Use Flow Settings** — the NLU training settings will follow the configuration in the **Include Default Reply in NLU Training** setting defined in **Settings > Configuration**. <br />- **On** — the default reply will be included in the NLU model training. <br /> - **Off** — the default reply won't be used in the NLU model training. |
| Channel                               | Specifies the channel to which the response applies. The default channel is **AI**.                                                                                                                                                                                                                                                                                                                                                                                            |
| Output Type                           | Specifies the type of response, supporting the same output types as the [Say Node](/ai/agents/develop/node-reference/basic/say). The default output is **Text**.                                                                                                                                                                                                                                                                                                               |
| Reset                                 | Resets the **Output Type** setting to its default value.                                                                                                                                                                                                                                                                                                                                                                                                                       |

### Configure the Default Reply for All Intents in the Flow

1. In the Flow editor, go to **Settings > Configuration** and select an Intent.
2. Go to the **General Flow Logic** section and configure the following settings:

| **Setting**                           | **Description**                                                                                                                                                                                                 |
| ------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Pass Default Reply into Flow          | Passes the default reply into the `nlu.intentMapperResults.finalIntentDefaultReply` Input object for use in Flow logic. The default reply message won't appear in the chat. The setting is disabled by default. |
| Continue Flow after Default Reply     | Continues executing the Flow logic immediately after the default reply has been delivered, without waiting for the user input. The setting is disabled by default.                                              |
| Include Default Reply in NLU Training | Includes the default replies in the NLU model training for matching Intents. This setting is enabled by default and can be customized for each Intent.                                                          |

### How to Test

1. In the Interaction Panel, enter a phrase matching the Intent. For example, `I want to order a pizza`.
2. Verify that the default reply is sent immediately.

The Flow continues or pauses depending on how you configured the **Continue Flow after Default Reply** setting.

## Example

Imagine a pizza ordering AI Agent.

Intent: `Order Pizza`<br />
Default reply: `Great choice! Your order is being processed.`<br />
User Input: `I'd like a pizza`

The NLU detects the Intent, the AI Agent sends the default reply,
and the Flow either continues to ask for toppings or pauses.

## More Information

* [Intents](/ai/platform-features/nlu/intents/overview)
* [Example Sentences](/ai/platform-features/nlu/intents/example-sentences)


# Intent Example Sentences
Source: https://docs.cognigy.com/ai/platform-features/nlu/intents/example-sentences



*Intent Example Sentences* are phrases that show how users might express their intentions to the AI Agent. A diverse set of example sentences enables the AI Agent to recognize a larger variety of user inputs. By associating these examples with specific Intents, the AI Agent can better understand and respond to user needs. These sentences help ensure the NLU model is well-trained and ready for human-AI interactions.

## Limitations

* You can add up to 2000 example sentences per Intent.
* The minimum number of example sentences is 5. If you attempt to train an Intent with fewer than 5 example sentences, a [red indicator](/ai/platform-features/nlu/intents/intent-analyzer#explore-intent-analyzer) will appear once the build completes.

## Managing Intent Sentences

<AccordionGroup>
  <Accordion title="Create Intent Sentences">
    <Tabs>
      <Tab title="Manually">
        To create an Intent sentence, go to the Intent in **NLU > Intents** in your Flow.
        Enter the example sentence and press <kbd>Enter</kbd> or <kbd>Tab</kbd>.
      </Tab>

      <Tab title="Automatically (Generative AI)">
        To use Generative AI for creating Intent example sentences, read the [Generative AI](/ai/agents/develop/gen-ai-and-llms/generative-ai) article.
      </Tab>
    </Tabs>
  </Accordion>

  <Accordion title="Delete Intent Sentences">
    If you need to delete or remove example sentences, you can hover over the sentence and click <img alt="delete sentence" />.
  </Accordion>
</AccordionGroup>

### Recommended Number of Example Sentences

For optimal performance, use 5-100 example sentences.
With this range, you can provide the AI Agent with enough data to recognize the Intent and avoid overcomplicating the model with unnecessary examples.

If you don't meet the minimum of 5 example sentences or exceed the maximum of 100, Cognigy shows a warning icon prompting you to adjust the number of sentences.

### Tips for Writing Effective Example Sentences

* **Avoid punctuation marks**. Keep the sentences simple and without punctuation. Using punctuation marks in example sentences, such as commas and question marks, can reduce recognition accuracy.
* **Avoid repetition**. Don't write similar sentences like `I want to order Pizza` and `I want to order Pasta`. Instead, use [Lexicons](/ai/platform-features/nlu/slots/user-defined/lexicon) to group similar terms and attach them to your Flow. This way, you can create a more general sentence like `I want to order food`, which will automatically recognize `Pizza` and `Pasta` as food-related items.
* **Use natural language instead of placeholders**. Avoid placeholder variables like `LEX_SLOT_123` and limit uppercase usage, for example, don't write `FOOD` in all caps unless it's an abbreviation.
* **Ensure uniqueness**. Make sure your example sentences are different from each other. If they're too similar, the AI Agent may struggle to differentiate Intents, which could cause incorrect results or trigger confirmation sentences often.

## More Information

* [Overview](/ai/platform-features/nlu/overview)
* [Annotations](/ai/platform-features/nlu/intents/annotations)


# Intent Analyzer
Source: https://docs.cognigy.com/ai/platform-features/nlu/intents/intent-analyzer



<a href="/release-notes/4.79"><Badge>Updated in 4.79</Badge></a>

The *Intent Analyzer* provides immediate feedback on the quality of example sentences and Intents within the Cognigy NLU model, enabling you to optimize AI Agents for better understanding and performance.

## Key Benefits

* **Rapid Quality Assessment**. You can evaluate the NLU model's effectiveness at a glance without extensive manual testing.
* **Targeted Improvements**. The Intent analyzer includes detailed scoring that highlights specific areas for refinement, thereby improving Intent recognition.

## Restrictions

* After editing example sentences, the NLU model must be retrained to update scores.
* Strong overall Intent model scores don't guarantee perfect individual Intents; red or yellow Intents may still need work.
* An Intent requires at least five example sentences. If trained with fewer, a negative feedback indicator will appear after the build, signaling that while the Intent still functions, its performance will be significantly reduced.

## How to Set up

1. In the Flow editor, go to **NLU > Intents**.
2. Add a variety of example sentences into each Intent.
3. In the upper-right corner, click **Build Model** to process the sentences and generate the initial model.
4. View scores for the overall model, individual Intents, and example sentences.

### Explore Intent Analyzer

The Intent Analyzer assesses three main components:

1. **Overall Intent Model** — shows a total score for the NLU model's quality.
2. **Individual Intent** — shows scores for each Intent, which reflect how well the NLU model recognizes them and if they overlap with other Intents. You will see the NLU score along with names of any overlapping Intents.
3. **Individual Example Sentence** — each example sentence receives a score based on how effective it is within the NLU model. You can see the score and any overlapping Intents by hovering over the traffic light in the interface.

<Frame>
  <img alt="Intent Analyzer with traffic light scores for model quality" />
</Frame>

The Intent analyzer uses a traffic light color system to provide users with three-status scoring feedback.

The color coding is explained in the table.

| **Color** | **Score Range**   | **Description**                                                                                                                                                  | **Recommendations**                                                                                                                                                                                                                                          |
| --------- | ----------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| Green     | Score > 0.8       | A green score indicates high confidence in Intent recognition. The NLU model is performing well and is ready for further testing by users.                       | Note that a perfect NLU model — where all sentences score green — isn't necessary, the traffic light system helps quickly identify areas that may need attention and refinement.                                                                             |
| Yellow    | 0.8 > Score ≥ 0.6 | A yellow score indicates moderate confidence. The NLU model is performing well but shows that the Intent recognition could be more consistent.                   | Review example sentences with yellow scores. Add more varied examples or adjust poorly scoring examples to cover additional variations. Monitor Intents over time for any further inconsistencies.                                                           |
| Red       | Score \< 0.6      | A red score indicates significant issues with Intent recognition. The NLU model may fail to recognize the intent accurately, and urgent improvements are needed. | Review the training data for this intent. Focus on revising or expanding example sentences to better represent the variety of ways users might express the query. Consider re-training the NLU model with additional or revised data to improve recognition. |

## How to Test

You can test NLU scores using the following interfaces:

<Tabs>
  <Tab title="GUI">
    In the Interaction Panel, activate the [debug mode](/ai/agents/test/interaction-panel/chat#debug-mode) feature and explore the detailed results in the `input.nlu.intentMapperResults.scores` object.
  </Tab>

  <Tab title="API">
    For more complex testing, you can use the API [POST /v2.0/projects//nlu/scores](https://api-trial.cognigy.ai/openapi#post-/v2.0/projects/-projectId-/nlu/scores) request to test NLU scores without the load from Flow execution.

    For example, assume the user asks the AI Agent: `Could you help me with the issue?` Add this question as the value for the parameter `sentence`. Specify the [Project ID](/ai/agents/develop/projects-and-flows/projects), [Flow Reference ID](/ai/agents/develop/projects-and-flows/overview), and [Locale Reference ID](/ai/platform-features/translation-and-localization/localization) of the AI Agent you want to test.

    **Request**

    ```json theme={null}
    Post /v2.0/projects/{projectId}/nlu/scores
    Content-Type: application/json

    {
      "flowReferenceId": "<flow-reference-id>",
      "localeReferenceId": "<locale-reference-id>",
      "sentence": "Can you help me with this issue?"
    }
    ```

    The response will return a collection of objects that includes matching NLU Intents and scores for the given sentence.

    **Response**

    ```json theme={null}
    [
      {
        "id": "129fdbd3-54d4-4704-a019-b579a7901641",
        "name": "Flight Booking",
        "score": 0.9333265866488035,
        "negated": false,
        "confirmationSentence": null,
        "confirmationSentences": null,
        "disambiguationSentence": null,
        "flow": "0d59e1c7-17e4-4737-aafa-4b27b48e6885",
        "description": "Handles queries related to booking flights, including availability, pricing, and reservations"
      },
      {
        "id": "0c22d2d8-edda-4e12-8d8f-6f5177771fc5",
        "name": "Payments",
        "score": 0.5039746888040336,
        "negated": false,
        "confirmationSentence": null,
        "confirmationSentences": null,
        "disambiguationSentence": null,
        "flow": "0d59e1c7-17e4-4737-aafa-4b27b48e6885",
        "description": "Manages questions and actions related to payment processes, methods, and issues"
      },
      {
        "id": "11a33c75-d074-4408-b437-5c00071d6c46",
        "name": "FAQ",
        "score": 0.0669071045766796,
        "negated": false,
        "confirmationSentence": null,
        "confirmationSentences": null,
        "disambiguationSentence": null,
        "flow": "0d59e1c7-17e4-4737-aafa-4b27b48e6885",
        "description": "Provides answers to frequently asked questions regarding various topics or services"
      }
    ]
    ```
  </Tab>
</Tabs>

## More Information

* [Intents](/ai/platform-features/nlu/intents/overview)


# Inheritance Mechanism
Source: https://docs.cognigy.com/ai/platform-features/nlu/intents/intent-hierarchy/inherit-child-example-sentences



The *Inheritance Mechanism* feature lets parent Intents use example sentences from their child Intents.
This mechanism helps the AI Agent better understand user inputs and determine the intended action.

### Key Benefits

* **Wider Scope for Parent Intents**. Parent Intents get a better trained NLU model by including examples from child Intents.
* **Clear Hierarchy**. You can see how Intents relate to each other and how they're scored.

## Prerequisites

Before using the inheritance mechanism, make sure you have a multilevel Intent structure.
For example, Level 1 parent Intents with Level 2 or Level 3 child Intents.

## How to Use

To activate the inheritance mechanism, follow these steps:

1. In the Flow editor, go to **NLU > Intents** and select the parent Intent you want to set up.
2. Activate the **Inherit Example Sentences from Child Intents** setting.
3. Save your changes and rebuild the model to apply them.

### Intent Hierarchy Matching

The parent Intent includes examples from its child Intents in its training data.
When a user sends an input, during Intent matching, both parent and child Intents can score highly.

The child Intent's details are saved in the `input.intent` Input object,
while the parent Intent's details appear in `input.nlu.intentMapperResults.scores`.
For a deeper understanding of the resolution process, you can trace the Intent path using `input.intentLevel` and `input.nlu.intentMapperResults.intentPath`.

<Accordion title="Input Object">
  ```json theme={null}
  {
    "input": {
      "intent": "Child Intent",  
      "nlu": {
        "intentMapperResults": {
          "scores": [
            {
              "intent": "Child Intent",
              "score": 0.6833273843878324
            },
            {
              "intent": "Parent Intent",
              "score": 0.5099832885597476
            }
          ],
          "intentPath": [
            "Parent Intent",
            "Child Intent"
          ]
        }
      },
      "intentLevel": {
        "level1": "Parent Intent",
        "level2": "Child Intent",
        "level3": null
      }
    }
  }
  ```
</Accordion>

### Example

Consider the inheritance mechanism in a customer support AI Agent.
The parent Intent (`Support`) inherits example sentences from its child Intents
(`Technical Support` and `Billing Support`).
This structure enables the AI Agent to handle general inquiries while accurately identifying specific Intents.

<AccordionGroup>
  <Accordion title="Intent Data">
    | **Level** | **Name** | **Intent Name**     | **Example Sentences**                                                                                        | **Description**                                                               |
    | --------- | -------- | ------------------- | ------------------------------------------------------------------------------------------------------------ | ----------------------------------------------------------------------------- |
    | Level 1   | Parent   | `Support`           | `I need assistance` <br /> `Can you help me?` <br /> `I have an issue`                                       | The high-level parent Intent recognizing broad customer queries.              |
    | Level 2   | Child    | `Technical Support` | `My internet is down` <br /> `I can't connect to the Wi-Fi` <br /> `My computer is broken`                   | The child Intent under the `Support` parent for issues related to technology. |
    | Level 2   | Child    | `Billing Support`   | `I need help with my invoice` <br /> `I have a question about my bill` <br /> `Can you explain this charge?` | The child Intent under the `Support` parent for issues related to billing.    |
  </Accordion>

  <Accordion title="Result">
    | **User Input**                | **Recognized Intent**              |
    | ----------------------------- | ---------------------------------- |
    | `I need assistance`           | `Support` (Parent Intent)          |
    | `My internet is down`         | `Technical Support` (Child Intent) |
    | `I need help with my invoice` | `Billing Support` (Child Intent)   |
  </Accordion>
</AccordionGroup>

## More Information

* [Overview](/ai/platform-features/nlu/intents/overview)
* [Intent Mapping Priority](/ai/platform-features/nlu/intents/intent-hierarchy/intent-mapping-priority)


# Intent Mapping Priority
Source: https://docs.cognigy.com/ai/platform-features/nlu/intents/intent-hierarchy/intent-mapping-priority



<a href="/release-notes/4.43"><Badge>Updated in 4.43</Badge></a>

The *Intent Mapping Priority* feature helps avoid ambiguity between parent and child Intents, which can occur when user input matches both.

With the Intent mapping priority, you can choose whether to prioritize the parent Intent or its child Intents. This approach allows you to control which Intent is more likely to be selected.

For better accuracy, use full sentences when creating example phrases for Intents.

## Key Benefits

* **Enhanced Accuracy in Intent detection**. By setting priorities, you can guide the system to focus on either parent or child Intents as desired.
* **Flexible customization**. Adjust mapping behavior to suit specific use cases, ensuring the right Intent is triggered.

## Prerequisites

Before using this feature, ensure the following:

* A Flow is created or available in your Project.
* Parent and child Intents are defined in **NLU > Intents**.

## Restrictions

* This feature works best with full-sentence Intent examples. Single words or short phrases may lead to inaccurate mapping results.

## How to Use

To set a priority, follow these steps:

1. Open an Intent in **NLU > Intents**.
2. In the **Intent Mapping Priority** field, select one of the following options:
   * **Prioritise This Intent** — set a priority for the current parent Intent. By default, this option is set up.
   * **Prioritise Child Intents** — set a priority for child Intents.
3. Save changes and click **Build model**. Build a model every time you want to apply your changes to the NLU model.

## Examples

Assume that you have the following Intent data:

| Intent  | Intent name        | Sentences                                                                                                                                                                                                                                                                   |
| ------- | ------------------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Parent  | Order Pizza        | I'd like to order a pizza. <br /> Can I have a pizza delivered? <br /> Can I get a pizza please? <br /> I need to purchase a pizza. <br /> I want to buy a pizza.                                                                                                           |
| Child 1 | Order Ham Pizza    | I want to get a ham pizza. <br /> Can I have a ham pizza please? <br /> I need a ham pizza. <br /> I would like to order a ham pizza. <br /> I am in the mood of for a ham pizza. <br /> Can I order a ham pizza? <br /> I'd like to have a ham pizza.                      |
| Child 2 | Order Cheese Pizza | I want to get a cheese pizza. <br /> Can I have a cheese pizza please? <br /> I need a cheese pizza. <br /> I would like to order a cheese pizza. <br /> I am in the mood of for a cheese pizza. <br /> Can I order a cheese pizza? <br /> I'd like to have a cheese pizza. |

Consider various examples of how each option of the **Intent Mapping Priority** setting works.

<AccordionGroup>
  <Accordion title="Prioritise This Intent">
    Set up **Prioritise This Intent** for the parent Intent.

    User input: `I want to order a bacon pizza`

    Result: `intent: Order Pizza, intent score: 0.33`

    JSON example of object stored in the `nlu.intentMapperResults.scores` array:

    ```json theme={null}
    "scores": [
            {
              "id": "6262e9e7-9dc5-4146-9153-96ef0693c18f",
              "name": "Order Pizza",
              "score": 0.3363160895380178,
              "negated": false,
              "confirmationSentence": null,
              "confirmationSentences": null,
              "disambiguationSentence": null,
              "flow": "898601c7-c5db-44f3-bfa0-bce265af02ad"
            }
          ]
    ```
  </Accordion>

  <Accordion title="Prioritise Child Intents">
    Set up **Prioritise Child Intents** for the parent Intent.

    User input: `I want to order a bacon pizza`

    Result: `intent: Order Ham Pizza, intent score: 0.63`

    JSON example of object stored in the `nlu.intentMapperResults.scores` array:

    ```json theme={null}
    "scores": [
            {
              "id": "8bba4b5a-1166-4459-bd1c-9e8b5391727e",
              "name": "Order Ham Pizza",
              "score": 0.6494886591790712,
              "negated": false,
              "confirmationSentence": null,
              "confirmationSentences": null,
              "disambiguationSentence": null,
              "flow": "898601c7-c5db-44f3-bfa0-bce265af02ad"
            },
            {
              "id": "c16f7afc-2875-49dc-8d7a-e06629d7fef9",
              "name": "Order Cheese Pizza",
              "score": 0.24775283391091624,
              "negated": false,
              "confirmationSentence": null,
              "confirmationSentences": null,
              "disambiguationSentence": null,
              "flow": "898601c7-c5db-44f3-bfa0-bce265af02ad"
            }
          ]
    ```
  </Accordion>
</AccordionGroup>

## More Information

* [Overview](/ai/platform-features/nlu/intents/overview)
* [Inheritance Mechanism](/ai/platform-features/nlu/intents/intent-hierarchy/inherit-child-example-sentences)
* [Intents](/ai/platform-features/nlu/intents/overview)


# Intent Hierarchy
Source: https://docs.cognigy.com/ai/platform-features/nlu/intents/intent-hierarchy/overview



*Intent Hierarchy* allows you to structure Intents into layered groups, connecting parent Intents with their children and grandchildren.
This organization of Intents helps AI Agents interpret user inputs by inheriting example sentences across levels. It increases flexibility and boosts Intent detection accuracy for more natural conversations.

## Restrictions

* Intent hierarchies are limited to three levels (Level 1, Level 2, and Level 3), preventing deeper nesting.
* By default, parent Intents don't use example sentences from their child Intents. To change this behavior, use the [Inheritance Mechanism](/ai/platform-features/nlu/intents/intent-hierarchy/inherit-child-example-sentences) feature.
* When user input matches both parent and child Intents, this overlap may cause ambiguity. Use the [Intent Mapping Priority](/ai/platform-features/nlu/intents/intent-hierarchy/inherit-child-example-sentences) feature to ensure the AI Agent selects the most specific Intent.

## Intent Hierarchy Structure

An Intent hierarchy follows this structure:

* **Parent Intents** – top-level Intents that represent a broader category or topic. They can inherit example sentences from their child Intents for broader recognition.
* **Child Intents** – more specific Intents nested under parent Intents. Each child Intent focuses on a narrower aspect of the parent's category and can share example sentences with their parent Intents.

<AccordionGroup>
  <Accordion title="Example: Intent Hierarchy Structure">
    ```txt theme={null}
    Flow
    ├── Intent A
    │   └── Intent A.1
    │       └── Intent A.2
    └── Intent B
    ```

    * **Intent A** (Level 1) – the parent Intent, representing a broad Intent category.
      * **Intent A.1** (Level 2) – a child Intent, providing a more specific Intent under `Intent A`.
        * **Intent A.1.1** (Level 3) – a grandchild Intent, the most detailed Intent under `Intent A.1`.
    * **Intent B** (Level 1) – another parent Intent, which stands alone at this level.
  </Accordion>

  <Accordion title="Example: Intent Hierarchy Structure in Customer Support Flow">
    ```txt theme={null}
    Customer Support Flow
    ├── Order Support
    │   └── Order Status
    │       └── Tracking Information
    └── Account Support
    ```

    * **Order Support** (Level 1) – the parent Intent, representing general questions about orders.
      * **Order Status** (Level 2) – a child Intent under `Order Support` that handles questions about order tracking.
        * **Tracking Information** (Level 3) – a grandchild Intent under `Order Status` that deals with requests for tracking numbers or delivery updates.
    * **Account Support** (Level 1) – a separate parent Intent, focusing on account-related questions.
  </Accordion>
</AccordionGroup>

## Working with Intent Hierarchy

In **NLU > Intents**, drag Intents from the list onto a parent Intent to build hierarchies.

## More Information

* [Intents](/ai/platform-features/nlu/intents/overview)


# Intent Trainer
Source: https://docs.cognigy.com/ai/platform-features/nlu/intents/intent-trainer



<a href="/release-notes/2026.7"><Badge>Updated in 2026.7</Badge></a>

One of the key strengths of AI Agents is their ability to improve over time.
Cognigy offers the *Intent Trainer* feature that allows you to refine AI Agents based on existing dialogs.
This tool improves the AI Agent's understanding by letting you review and add user inputs to [Intents](/ai/platform-features/nlu/intents/overview).

## Key Features

* **Analysis of Collected User Inputs**. Review and analyze Intent Trainer records to identify areas for improvement.
* **Adding Inputs to Intents**. Choose which Intent Trainer records will enhance specific Intents and add them to the Intent Sentences list.
* **Instant Training**. After refining Intent Trainer records, quickly update the NLU model on the Intent Trainer page. This action eliminates the need to rebuild the model on the Flow page.
* **Automatic Scoring**. The Train capability automatically scores Intents, and you can visually track progress with color-coded icons and scoring data in the list.

## Limitations

* By default, the Intent Trainer records Time-to-Live (TTL) is set to 10 days (14400 minutes). If you have on-premises installations, you can change this value using the following variable: `TRAINERRECORD_TTL_IN_MINUTES`. For example, to set the TTL to 30 days, configure the variable as follows: `TRAINERRECORD_TTL_IN_MINUTES="43200"`.
* The maximum file size for uploading Intent Trainer records is 150 MB.

## Working with Intent Trainer

In **Tweak > Intent Trainer**, you can:

1. [Find Intent Trainer records by using filters](#filter-intent-trainer-records)
2. [Refine Intent Trainer records: add to Intents, skip, or ignore](#manage-intent-trainer-records)
3. [Train the NLU model](#train-the-nlu-model)
4. [*(Optional)* Import and Export Intent Trainer records](#import-and-export-intent-trainer-records)

### Filter Intent Trainer Records

Review the collected input records from users and search for them using filters.

<AccordionGroup>
  <Accordion title="Predefined Filters">
    The **Filter Preset** allows you to filter records based on predefined categories for more efficient navigation and analysis.

    | **Filter Preset Option** | **Description**                                                |
    | ------------------------ | -------------------------------------------------------------- |
    | Show All                 | View all available records without any filters applied.        |
    | Found Intents            | View records with Intents identified by the system.            |
    | Found Lexicon Slots      | View records where Lexicon Slots were detected.                |
    | Poor Intent Score        | View records with Intents that have a low confidence score.    |
    | Fair Intent Score        | View records with Intents that have a medium confidence score. |
  </Accordion>

  <Accordion title="Custom Filters">
    When you select **Custom** in the Filter Preset, you can apply any filter. If you modify a filter while a preset option is selected, the Filter Preset automatically changes to **Custom**.

    | **Filter**                                               | **Description**                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
    | -------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Snapshot                                                 | Select a specific Snapshot to view Intents and related records. By default, **No Snapshot** is selected.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |
    | Locale                                                   | Filter records by language or locale when multiple locales are used. By default, **Any Locale** is selected.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
    | Intent                                                   | Filter by specific Intents to refine the results. By default, **Any Intent** is selected.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | Review Status                                            | Filter records by their review status to track progress:<br /> <ul><li>**Not Reviewed** — a record was not reviewed.</li><li>**Reviewed** — a record was added to the Intent and reviewed.</li><li>**Ignored** — a record, along with all future inputs containing the same text, is ignored.</li><li>**Skipped** — a record is skipped for now and can be revisited later, with the option to review it again.</li></ul>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | Found Intent                                             | Filter records based on whether an Intent was identified or not. By default, **No Intent Found** is selected. You can select one of the following options: <br /><ul><li>**Found Intent** — shows records where an Intent was successfully matched.</li><li>**No Intent Found** — shows records where no Intent was identified.</li></ul>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | [Found a Slot](/ai/platform-features/nlu/slots/overview) | View all Slots found based on the selected filter option. By default, **Not Selected** is chosen. You can select one of the following options: <br /><ul><li>**Found Slots** — shows records where a Slot was identified.</li><li>**No Slot Found** — shows records where no Slot was identified.</li></ul>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | Intent Score                                             | Filter records by confidence score, shown as color-coded icons. By default, **Not Selected** is chosen. You can select one of the following options: <ul><li>**Poor** (0–0.3) — low confidence score, meaning the Intent was not accurately identified.</li><li>**Fair** (0.3–0.7) — medium confidence score, showing a moderately accurate identification.</li><li>**Good** (0.7–1.0) — high confidence score, indicating a very accurate Intent identification.</li></ul>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
    | Input Types                                              | Filter records by user input type. You can select one of the following input types: <br /> <ul><li>**Positive Answer** — user responded affirmatively. For example, `yes`, `sure`.</li><li>**Negative Answer** — user responded negatively. For example, `no`, `never`.</li><li>**Greeting** — user initiated a greeting. For example, `hi`, `hello`.</li><li>**Goodbye Message** — user expressed farewell. For example, `goodbye`, `see you later`.</li><li>**Statement** — user made a declarative statement. For example, `I like this product`.</li><li>**Command** — user issued a command. For example, `turn on the light`.</li><li>**Why Question** — user asked a `why` question. For example, `Why is it included in the package?`.</li><li>**How Question** — user asked a `how` question. For example, `How does this work?`.</li><li>**Yes or No Question** — user asked a `yes/no` question. For example, `Do you have this item in stock?`.</li></ul> |
    | [Slot](/ai/platform-features/nlu/slots/overview)         | Filter records by selected Slot type. You can select one of the following options: <br /> <ul><li>**None** — no Slots identified.</li><li>**Lexicon Slots** — identified lexicon-based Slots, which are user-defined categories.</li><li>**System Slots** — Slots that are predefined in the system, such as dates or numbers.</li></ul>                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |
  </Accordion>
</AccordionGroup>

### Manage Intent Trainer Records

Based on the analysis, decide which user inputs will improve a particular Intent, add those inputs to the corresponding Intent.

<AccordionGroup>
  <Accordion title="Add Records to Intents">
    <Tabs>
      <Tab title="GUI">
        1. In the Intent Trainer, select a record from the list and click **Add to Intent**.
        2. In the **Edit Record** window, select an Intent to which you want to add a record.\
           Change the text of the Intent in the **Text** field.\
           Additionally, you can create a new Keyphrase or Synonym.\
           Save changes and apply them. The Intent record will be added as a sentence to the selected Intent.
      </Tab>

      <Tab title="API">
        You can add a user input record to a specific Intent using the [Cognigy API POST /v2.0/trainer/batch](https://api-trial.cognigy.ai/openapi#post-/v2.0/trainer/batch) request, specify `"action": "addToIntent"` in the `operations` object.
      </Tab>
    </Tabs>
  </Accordion>

  <Accordion title="Skip Records">
    <Tabs>
      <Tab title="GUI">
        1. In the Intent Trainer, select a record from the list and click **Skip**.
        2. Apply changes. This action moves the input to **Skipped** records, but it will reappear in **Not reviewed** if the same input is entered again.
      </Tab>

      <Tab title="API">
        You can skip Intent Trainer records using he [Cognigy API POST /v2.0/trainer/batch](https://api-trial.cognigy.ai/openapi#post-/v2.0/trainer/batch) request, specify `"action": "skip"` in the `operations` object.
      </Tab>
    </Tabs>
  </Accordion>

  <Accordion title="Ignore Records">
    <Tabs>
      <Tab title="GUI">
        1. In the Intent Trainer, select an Intent record from the list and click **Ignore**.
        2. Apply changes. This action moves the user input to the **Ignored** records, and if a user enters the same input, it will also be ignored.
      </Tab>

      <Tab title="API">
        You can skip Intent Trainer records using he [Cognigy API POST /v2.0/trainer/batch](https://api-trial.cognigy.ai/openapi#post-/v2.0/trainer/batch) request, specify `"action": "ignore"` in the `operations` object.
      </Tab>
    </Tabs>
  </Accordion>
</AccordionGroup>

### Train the NLU Model

After adding a record to the Intent, click **Train** in the top-right corner of the **Intent Records** page.\
You don't need to run **Build Model** in the Flow — the Intent Trainer has already scored the Intent, shown in the scoring data and color-coded icons.

### Import and Export Intent Trainer Records

You can import and export records between production and development environments.\
The Intent Trainer [temporarily stores records](#limitations).
To ensure data safety, export the records for local storage.

<AccordionGroup>
  <Accordion title="Import Records">
    <Tabs>
      <Tab title="GUI">
        1. Go to **Tweak > Intent Trainer**.
        2. In the upper-right corner, click <img alt="menu" /> **> Import Trainer Records**.
        3. Select a file in the `CTRAIN` format from your computer and click **Open**.

        Once the file is uploaded, you will receive a system success message.
      </Tab>

      <Tab title="API">
        You can import Intent Trainer records using the [Cognigy API POST /v2.0/trainer/upload](https://api-trial.cognigy.ai/openapi#post-/v2.0/trainer/upload) request.
      </Tab>
    </Tabs>
  </Accordion>

  <Accordion title="Export Records">
    <Tabs>
      <Tab title="GUI">
        1. Go to **Tweak > Intent Trainer**.
        2. In the upper-right corner, click <img alt="menu" /> **> Export Trainer Records**.
        3. Select a date range by clicking the date and selecting the desired date in the calendar.
        4. To include reviewed records in the file, activate **Include reviewed**.
        5. Click **Confirm**, then **Download Trainer Records**.

        The file will be downloaded in the `CTRAIN` format.
      </Tab>

      <Tab title="API">
        You can export Intent Trainer records using the Cognigy API. First, create a file via the [POST /v2.0/trainer/package](https://api-trial.cognigy.ai/openapi#post-/v2.0/trainer/package) request, and then download it via the [POST /v2.0/trainer/downloadlink](https://api-trial.cognigy.ai/openapi#post-/v2.0/trainer/downloadlink).
      </Tab>
    </Tabs>
  </Accordion>
</AccordionGroup>

## More Information

* [Intents](/ai/platform-features/nlu/intents/overview)


# Intents
Source: https://docs.cognigy.com/ai/platform-features/nlu/intents/overview



<a href="/release-notes/2026.3"><Badge>Updated in 2026.3</Badge></a>

In Cognigy, *Intents* identify user's goals based on their input.
Each Intent is defined by example sentences that show how users might express a specific request.
For example,
an Intent named `OrderFood` might include sentences like `Can you get me some food?` or `I'd like to order a pizza`.
The NLU engine analyzes and scores these Intents to match user inputs.

Each Intent includes:

* [Example Sentences](#add-example-sentences)
* [Additional Configuration](#apply-additional-settings)

## Working with Intents

<Tabs>
  <Tab title="GUI">
    You can create, edit, and manage Intents as well as download and upload them in **NLU > Intents** of the Flow editor.
  </Tab>

  <Tab title="API">
    You can create, edit, and manage Intents as well as download and upload them using the [Cognigy API](https://api-trial.cognigy.ai/openapi#get-/v2.0/flows/-flowId-/Intents).
  </Tab>
</Tabs>

### Download and Upload Intents

Intents can be uploaded to and downloaded from Cognigy as CSV or JSON files. Use these CSV or JSON files to import Intents, edit using your favorite spreadsheet or developer tools, and move them between Flows.

<AccordionGroup>
  <Accordion title="CSV">
    Your CSV should be formatted as follows:

    * No header.
    * Cognigy automatically detects various CSV formats. If you encounter any issues, use UTF-8 encoding, comma (`,`) as the delimiter, double quotes (`"`) as your quote character, and newline (`\n`) as your line terminator.
    * You are required to fill in the following columns:

    | Name                                                                                                                | Property type                                                                                                                                                                                                                                                                                                                                                                                | Value                                                                                                        |
    | ------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------ |
    | The name of the Intent used for identification and processing in Cognigy. The Intent name is a `string` type value. | The property type corresponds to the following Intent components and settings:<br /> `exampleSentence` — a sample user input.<br />`defaultReply` — the standard response when the Intent is triggered.<br />`confirmationSentence` — a response that asks the user for confirmation.<br />`rules` - the conditions to handle the Intent and guide the conversation based on the user input. | The value is a `string` type value and corresponds to an example sentence utterance, such as `I want pizza`. |

    CSV format example:

    ```txt theme={null}
    |--------------|----------------------|------------------------------------------------|
    | rejectIntent | exampleSentence      | This request cannot be fulfilled at this time. |
    | Intent1      | exampleSentence      | Could you please assist me with this issue?    |
    | Intent1      | defaultReply         | Your request has been successfully processed.  |
    | Intent1      | defaultReply         | Thank you for your inquiry.                    |
    | Intent1      | confirmationSentence | Are you sure you want to proceed?              |
    | Intent2      | exampleSentence      | I'd like to make a reservation, please.        |
    | Intent2      | exampleSentence      | I need assistance with my account.             |
    ```

    <Warning>
      Keep regional settings consistent on your operating system (OS), as they affect the CSV delimiter. Mismatched settings can break the file and cause upload failure.
    </Warning>
  </Accordion>

  <Accordion title="JSON">
    JSON format serves as a comprehensive specification for defining Intents within Cognigy.

    The provided JSON example illustrates the structure and content of an Intent, including Intent name (`Pizza`), example sentences, rules, tags, disambiguation sentence, default reply, and other Intent-related settings.

    JSON format example:

    ```json theme={null}
    [
        {
            "name": "Pizza",
            "exampleSentences": [
                "I'd like to order many pizzas",
                "Do you have any meat pizza options?",
                "I'm interested in trying a vegan pizza",
                "Could I have a cheese pizza, please?",
                "I'm craving pizza right now"
            ],
            "rules": [],
            "tags": [],
            "condition": "",
            "disambiguationSentence": "",
            "childFeatures": false,
            "biasTowardsParentOrChildIntents": "parents",
            "parent": "",
            "defaultReply": {},
            "entryPoint": "",
            "isRejectIntent": false,
            "isDisabled": false,
            "overrideIntentDefaultRepliesAsExamples": ""
        }
    ]
    ```
  </Accordion>
</AccordionGroup>

### Add Example Sentences

When you create an Intent, add [example sentences](/ai/platform-features/nlu/intents/example-sentences) that help the NLU model recognize user queries.

### Build the NLU Model

Whenever you make changes to the Intent or its related settings, you need to rebuild the NLU model.

If the NLU model needs rebuilding, the [red indicator](/ai/platform-features/nlu/intents/intent-analyzer#explore-intent-analyzer) will be displayed:

* next to the Flow name in the Flows list.
* next to the **Build Model** button in **NLU > Intents** of the Flow editor.

If your Flow includes States and Intents with [Intent Conditions](/ai/platform-features/nlu/intents/conditions),
you can exclude them from training using **Quick Build**.
This approach speeds up the NLU model building process,
especially for large Flows that contain many States and Intent Conditions.
If you don't use these features,
both **Build Model** and **Quick Build** will deliver the same model quality and training speed.

You can build the NLU model for a single Flow or for all Flows in your Project:

<AccordionGroup>
  <Accordion title="Single Flow">
    1. In the left-side menu of the Project, go to **Build > Flows**.
    2. On the **Flows** page, select the Flow that you want to train.
    3. In the upper-right corner, select **NLU**.
    4. On the **Intents** tab, click **Build Model** or <img alt="expand-locale-list" /> **> Quick Build**.

    If you have attached Flows to the current Flow, the model will automatically be built for the attached Flows. In the UI, trained attached Flows will appear as untrained (with a red indicator), which you can ignore.
  </Accordion>

  <Accordion title="All Flows">
    1. In the left-side menu of the Project, go to **Build > Flows**.
    2. On the **Flows** page, click **Train all Flows**. If **Train all Flows** isn't available, for version 2026.2 or earlier, set `FEATURE_TRAIN_ALL_PROJECT_FLOWS: "true"` in `values.yaml` for on-premises installations, or upgrade to version 2026.3 for SaaS.

    Note that during training, Flows are retrieved from the database sequentially, following their default storage order. Training doesn't run in parallel — each Flow is trained one after another, regardless of the number of service replicas or deployed locales.

    If the NLU model is built successfully, you will receive a green success notification.
    Otherwise, an error notification will appear with details; for example, you may need to add more example sentences.
  </Accordion>
</AccordionGroup>

### Apply Additional Settings

<AccordionGroup>
  <Accordion title="Basic Additional Settings">
    | **Setting**                                  | **Description**                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
    | -------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
    | Inherit Example Sentences from Child Intents | Allows example sentences from child Intents to be inherited by the parent Intent. For more information, read [Intent Hierarchy](/ai/platform-features/nlu/intents/intent-hierarchy/overview).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
    | Intent Mapping Priority                      | Defines the priority of the Intent when mapped against other Intents during the recognition process. For more information, read [Intent Mapping Priority](/ai/platform-features/nlu/intents/intent-hierarchy/intent-mapping-priority).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             |
    | Rules                                        | Allows configuring rules to restrict Intent execution based on specific conditions. For more information, read [Rules](/ai/platform-features/nlu/intents/rules).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
    | Default Reply                                | Defines the fallback response if no other reply is configured or applicable. For more information, read [Default Reply](/ai/platform-features/nlu/intents/default-reply).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
    | Intent Entrypoint                            | Determines where the Flow execution starts when the Intent is triggered. You can select one of the following options: <br />- **Execute from START in this Flow** — this option ensures that the Flow begins from the **Start** Node when the Intent is triggered. The option is selected by default. <br />- **Execute normally** — the Flow execution continues from the Node where the Intent was recognized, without restarting the Flow. This option is useful when you want to maintain the conversation's context and avoid restarting the Flow. <br />- **Execute from a Node in this Flow** — the Flow execution starts from a specific Node when the Intent is triggered. This option provides flexibility to define custom Entrypoints within the Flow. |
  </Accordion>

  <Accordion title="Advanced Additional Settings">
    | **Setting**             | **Description**                                                                                                                                                                                                                                                                                                                                                                                                                                            |
    | ----------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Tags                    | Allows assigning tags to the Intent for easier organization and filtering.                                                                                                                                                                                                                                                                                                                                                                                 |
    | Condition               | Defines the conditions that must be met for the Intent to be executed. For more information, read [Conditions](/ai/platform-features/nlu/intents/conditions).                                                                                                                                                                                                                                                                                              |
    | Confirmation Sentence   | Specifies the sentence used to confirm the Intent with the end user. For more information, read [Confirmations](/ai/platform-features/nlu/intents/confirmation-words).                                                                                                                                                                                                                                                                                     |
    | Disambiguation Sentence | Provides a sentence to clarify the Intent if there is ambiguity during recognition. The disambiguation sentence will be exposed in the Input object: <br />- `input.nlu.intentMapperResults.finalIntentDisambiguationSentence` — the disambiguation sentence for the best-matching Intent that was selected.<br />- `input.nlu.intentMapperResults.scores[0].disambiguationSentence` — the disambiguation sentence for a specific Intent that was matched. |
    | Analytics Step          | Marks the Intent for analytics tracking, allowing it to be included in analytical reports in Insights and OData.                                                                                                                                                                                                                                                                                                                                           |
  </Accordion>
</AccordionGroup>

## More Information

* [Overview](/ai/platform-features/nlu/overview)


# Reject Intent
Source: https://docs.cognigy.com/ai/platform-features/nlu/intents/reject-intent



The *Reject Intent* feature allows you to block the NLU model from recognizing user inputs that aren't relevant to your AI Agent.
This approach helps AI Agents focus on in-context questions, ignoring irrelevant ones.

## Key Benefits

* **Enhanced focus on core capabilities**. By filtering out irrelevant inputs, your AI Agent can prioritize the tasks it's designed to handle.
* **Improved user experience**. By avoiding responses to out-of-scope queries, you can prevent confusion and keep conversations on track.

## Restrictions

* This feature may struggle with complex or varied off-topic inputs unless paired with [standard Intents](/ai/platform-features/nlu/intents/overview) or [Intent rules](/ai/platform-features/nlu/intents/rules).

## Limitations

* You can create only one reject Intent per Flow.

## How to Use

To use this feature, create a reject Intent and add several example sentences.

<AccordionGroup>
  <Accordion title="Create a Reject Intent">
    <Tabs>
      <Tab title="GUI">
        To create a reject Intent, go to **NLU > Intents** your Flow.
        Under the **Create Intent** button, click ![menu](https://docs.cognigy.com/_assets/icons/black-and-white/vertical-ellipsis.svg) and select **Create 'Reject Intent'**.
      </Tab>

      <Tab title="API">
        You can create a reject Intent using the [Cognigy API /v2.0/flows//intents](https://api-trial.cognigy.ai/openapi#post-/v2.0/flows/-flowId-/intents) request and specifying the `"isRejectIntent": true` parameter.
      </Tab>
    </Tabs>
  </Accordion>

  <Accordion title="Add Example Sentences">
    On the **Reject Intent** page, add a list of example sentences representing the user inputs you want the NLU model to ignore.
    Save changes and click **Build Model** to apply them. You can also generate these sentences [using Generative AI](/ai/agents/develop/gen-ai-and-llms/generative-ai).
    Similar to standard Intents, you can also add [annotations](/ai/platform-features/nlu/intents/annotations) to each example sentence.
  </Accordion>
</AccordionGroup>

## How to Test

Test your feature in the Interaction Panel:

1. Open the Interaction Panel and activate **Debug Mode**.
2. Enter examples matching your reject Intent sentences to verify if they're ignored by the NLU model. If the NLU model catches them, the Input object will have the following format:

   ```json theme={null}
    "nlu": {
       "intentMapperResults": {
         ...
         "intentPath": [
           "Reject Intent"
         ],
         "scores": [
           {
             "id": "f0db738b-e1c3-4841-a6d9-490e082b8ec5",
             "name": "Reject Intent",
             "score": 1,
            ... 
           }
         ]
       },
   ```

## Example

Imagine your AI Agent is designed to assist drivers with parking-related queries.
A user asks `Where can I find a park nearby?` intending a green space rather than a parking lot.
By including similar reject Intent sentences, you can train the AI Agent to filter out these requests, ensuring it stays focused on parking assistance.

<Accordion title="More Reject Example Sentences">
  | **Reject Intent Name** | **Example Sentences**                                                                                                                                                                                                                                                                                               |
  | ---------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
  | `OutOfScope_Parks`     | - Where can I find a park nearby?<br />- Are there any green spaces around here? <br />- Is there a park I can visit close by? <br />- Where's the nearest nature spot? <br />- Can you point me to a local park? <br />- Are there any nice parks in this area?  <br />- Where can I take a walk in a park nearby? |
</Accordion>

## More Information

* [Intents](/ai/platform-features/nlu/intents/overview)
* [Intent Rules](/ai/platform-features/nlu/intents/rules)


# Intent Rules
Source: https://docs.cognigy.com/ai/platform-features/nlu/intents/rules



*Intent Rules* are conditions you define to help your AI Agents recognize user inputs and trigger a specific Intent based on precise criteria.
These rules provide more control and flexibility than standard Intents, which work by recognizing patterns in example sentences.

## Key Benefits

* **Precision**. Intent rules are especially valuable when dealing with technical payloads (structured data) or specific conditions that Intents don't easily handle.
* **Context-Sensitive Triggers**. You can enforce certain conditions, such as checking the user's login status or ensuring required data is present before triggering the Intent.
* **Simple Conditions**. They work well when you have clear, straightforward conditions for triggering the Intent. For example, specific keywords or patterns in the text.

## Restrictions

* Intent rules always have a score of `1` when they match, thresholds don't apply.
* Intent rules take priority by forcing an Intent match when their conditions are met, ignoring both Intent Sentences and Intent conditions. In contrast, Intent conditions determine whether Intent Sentences should be used.

## How to use

To set a rule for a specific Intent, navigate to **NLU > Intents** in the Flow editor.
Open the desired Intent, then go to **Rules** to configure the rule.

You can add multiple rules to an Intent.
You write these rules using CognigyScript, which allows you to define the conditions that must be met to trigger the Intent.
These conditions usually rely on input, context, or other dynamic data.

### Syntax

<AccordionGroup>
  <Accordion title="Basic Text Matching">
    This rule triggers the Intent if the user input contains the word `pizza`.

    ```js theme={null}
    ci.text.includes("pizza")
    ```
  </Accordion>

  <Accordion title="Comparing Input Text">
    The rule matches if the user exactly says `I want to order a pizza`.
    This rule only matches this exact phrase, and minor variations (like punctuation changes) would prevent the Intent from being triggered, making it less robust than an ML-based approach.

    ```js theme={null}
    ci.text == "I want to order a pizza"
    ```
  </Accordion>

  <Accordion title="Using Context Data">
    The rule checks if the user has selected `pizza` as the only item in their `food` Slot.

    ```js theme={null}
    input.slots.food[0].keyphrase == "pizza" && input.slots.food.length == 1
    ```
  </Accordion>

  <Accordion title="Combining Multiple Conditions">
    The rule triggers the Intent if the user mentions `order` and has selected one item in their `food` Slot.

    ```js theme={null}
    ci.text.includes("order") && input.slots.food.length === 1
    ```
  </Accordion>

  <Accordion title="Using Regular Expressions">
    This rule uses a regular expression to match the phrase `order` followed by any characters and then `pizza`.

    ```js theme={null}
    ci.text.match(/order\s.*\spizza/)
    ```
  </Accordion>

  <Accordion title="Dynamic Input Matching">
    This rule checks if the user input matches a dynamic sentence structure based on the user's first name, such as `John wants pizza`.

    ```js theme={null}
    ci.text === `${input.slots.first_name[0].keyphrase} wants pizza`
    ```
  </Accordion>
</AccordionGroup>

### Rules and Intent Sentences: Prioritization

Intent rules always have higher priority than Intent sentences.
This means that if a user input triggers both a rule and an Intent sentence, the rule will be triggered first.

### Resolving Conflicts Between Multiple Rules

If multiple rules apply simultaneously, the order of rules in the Flow hierarchy determines which one is triggered.

There are three main principles:

```text theme={null}
Main Flow
├── Intent A
│   └── Intent A.1
│       └── Intent A.2
└── Intent B

Attached Flow 1
└── Intent C
    └── Intent C.1
        └── Intent C.2

Attached Flow 2
└── Intent D
```

| Principle                                 | Description                                                                                                                             | Example                                                                                                                                                                          |
| ----------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Intent rules lower in the list come first | If multiple Intent rules match, the one that appears last in the list is triggered.                                                     | If `Intent A.2` and `Intent B` match the same input, `Intent B` will be triggered because it is lower in the list.                                                               |
| Children come first                       | If both parent and child Intents match, the child Intent takes precedence.                                                              | If `Intent A.1` and `Intent A.2` match the same input, `Intent A.2` will be triggered.                                                                                           |
| Attached Flows come third                 | Intents in attached Flows are processed after those in the main Flow. The order of attached Flows determines which Intent is triggered. | If `Intent C` in Attached Flow 1 and `Intent D` in Attached Flow 2 both match the same input, `Intent C` will be triggered because it comes first in the list of attached Flows. |

## Use Cases

<AccordionGroup>
  <Accordion title="Processing an Order Request">
    If you build a system that handles order requests from a user via an API
    (coming from a Webchat or frontend interface), the request might include structured data, such as JSON:

    ```json theme={null}
    {
      "action": "order",
      "product": "pizza",
      "quantity": 2,
      "userId": "user123"
    }
    ```

    You might want the system to trigger an `Order` Intent when it receives this specific structure.

    **Rule**:

    ```js theme={null}
    input.data.action === "order" && input.data.product === "pizza" && input.data.quantity >= 1
    ```

    This rule ensures that the `order` Intent is only triggered if the data contains an `action` of `order`, the `product` is `pizza`, and the `quantity` is greater than or equal to 1.
  </Accordion>

  <Accordion title="Checking for Specific Parameters in a Payload">
    Imagine you have an API that receives user's login data:

    ```json theme={null}
    {
      "action": "login",
      "username": "john_doe",
      "password": "secret"
    }
    ```

    You can create a rule to trigger an `Authenticate` Intent based on the specific parameters in the payload:

    **Rule**:

    ```js theme={null}
    input.data.action === "login" && input.data.username !== "" && input.data.password !== ""
    ```

    This rule ensures the Intent is triggered only when the payload contains both `username` and `password` fields.
  </Accordion>

  <Accordion title="Validating and Triggering Action Based on a Webhook Payload">
    You might have a webhook integration where you receive data from an external service, such as a payment gateway.
    The payload might look something like this:

    ```json theme={null}
    {
      "payment_status": "success",
      "transaction_id": "txn123456",
      "amount": 100.0
    }
    ```

    To trigger an Intent when a `successful payment` is detected, you can use the following rule:

    **Rule**:

    ```js theme={null}
    input.data.payment_status === "success" && input.data.transaction_id !== undefined && input.data.amount > 0
    ```
  </Accordion>
</AccordionGroup>

## More Information

* [Overview](/ai/platform-features/nlu/overview)
* [Intent Conditions](/ai/platform-features/nlu/intents/conditions)


# Thresholds
Source: https://docs.cognigy.com/ai/platform-features/nlu/intents/thresholds



The *Thresholds* feature provides configurable score boundaries
that determine how the NLU model interprets and classifies user inputs based on their Intent.
This feature enables precise control over Intent recognition,
ensuring the NLU can confidently identify Intents, prompt for reconfirmation when needed, or reject unclear inputs.

## Key Benefits

* **Improved Accuracy**. Adjustable thresholds enhance the reliability of Intent mapping by setting clear confidence levels.
* **Flexible Control**. You can fine-tune the system's sensitivity to match specific use cases or performance needs.

## Prerequisites

Before using thresholds, ensure the following:

* An active Flow is set up within the Project.
* Intents are defined to allow the NLU to classify user inputs.

## How to Use

Explore threshold types, then set them up.

### Explore Threshold Types

| **Type**                 | **Description**                                                                                                               |
| ------------------------ | ----------------------------------------------------------------------------------------------------------------------------- |
| Confidence Threshold     | Intents with a score higher than this threshold are considered as confirmed.                                                  |
| Reconfirmation Threshold | Intents with a score higher than the reconfirmation threshold but lower than the confidence threshold require reconfirmation. |
| Not Found Threshold      | Intents with a score lower than the reconfirmation threshold aren't considered.                                               |

#### Reconfirmation Threshold Specifics

The reconfirmation threshold behaves differently depending on whether a [confirmation sentence](/ai/platform-features/nlu/intents/overview#apply-additional-settings) exists in your Intent:

* With a confirmation sentence, the reconfirmation threshold triggers the reconfirmation sentence.
* Without a confirmation sentence, the reconfirmation threshold confirms the user input.

### Set up Thresholds

Thresholds can be configured at either the Flow or Project level. To set them up:

<Tabs>
  <Tab title="Flow level">
    1. In the Flow editor, go to **Settings > Configuration**.
    2. Activate the **Thresholds** option. As soon as activated, this option overrides the **Thresholds** values set up at the Project level.
    3. Adjust the threshold sliders:
       * **Reconfirmation Threshold** (default: 0.2):
         * Move the slider left to increase reconfirmation prompts.
         * Move the slider right to reduce reconfirmation prompts.
       * **Confidence Threshold** (default: 0.4):
         * Move the slider left to allow lower-confidence matches.
         * Move the slider right to require higher confidence for matches.
  </Tab>

  <Tab title="Project level">
    1. In the left-side menu of your Project, go to **Manage > Settings**.
    2. Go to **NLU > Thresholds**.
    3. Adjust the threshold sliders:
       * **Reconfirmation Threshold** (default: 0.2):
         * Move the slider left to increase reconfirmation prompts.
         * Move the slider right to reduce reconfirmation prompts.
       * **Confidence Threshold** (default: 0.4):
         * Move the slider left to allow lower-confidence matches.
         * Move the slider right to require higher confidence for matches.

    Fine-tune the sliders: move them closer for decisive responses or farther apart for cautious reconfirmation.
  </Tab>
</Tabs>

## Example

Imagine a user asks `Can you book a flight?` in the chat with the AI Agent. The NLU model assigns a score to the `BookFlight` Intent:

| **Score** | **Threshold Range**                                                   | **Outcome**                                            | **AI Agent Response**            |
| --------- | --------------------------------------------------------------------- | ------------------------------------------------------ | -------------------------------- |
| 0.5       | Above confidence threshold (0.4)                                      | The Intent is confirmed and the conversation proceeds. | —                                |
| 0.3       | Between reconfirmation threshold (0.2) and confidence threshold (0.4) | AI Agent seeks confirmation.                           | `Did you mean to book a flight?` |
| 0.1       | Below reconfirmation threshold (0.2)                                  | The Intent not recognized.                             | `I'm not sure what you mean.`    |

## More Information

* [Intents](/ai/platform-features/nlu/intents/overview)


# Trigger Intent
Source: https://docs.cognigy.com/ai/platform-features/nlu/intents/trigger-intent



The *Trigger Intent* feature lets you
manually activate a specific Intent in your Flow using the syntax `cIntent:` followed by the Intent name.
This feature overrides the default Intent mapping process, giving you precise control over the Flow's next steps based on user input, such as a message or button click.

Use the Trigger Intent feature when you need to:

* Skip NLU processing and direct the Flow to a specific Intent.
* Ensure consistent navigation in response to button clicks or system actions.
* Improve the user experience by minimizing errors from misinterpreted user input.

## Key Benefits

* **Precise Flow Control**. Ensures the correct Intent drives the Flow forward.
* **Postback Support**. Triggers a specific Intent smoothly when a user clicks a button with a Postback value.

## Prerequisites

Before using this feature, ensure the following:

1. You set up your Flow with specific Intents for the scenarios you want to trigger.
2. You have a Say, Question, or Optional Question Node in the Flow editor.

## Restrictions

* This feature bypasses full NLU processing, preventing default replies from triggering and automatically setting the Intent score to 1. Don't use this feature for general NLU analysis.

## How to Use

1. In the Flow editor, identify the Intent needed to continue your Flow. For example, `Continue Booking`.
2. In a Say, Question, or Optional Question Node, use the syntax `cIntent:Continue Booking` as a Postback value in [Output Types](/ai/agents/develop/node-reference/basic/say) that support Postbacks.
   Follow these syntax rules:
   * No space should follow the colon, though spaces within the Intent name are allowed. For example, `cIntent:Continue Booking`.
   * *(Optional)* Add a text for Slot parsing after a pipe symbol (`|`). For example, `cIntent:continueBooking|book tomorrow`.

## How to Test

Test the feature in the Interaction Panel:

1. Open the Interaction Panel and enable **Debug Mode**.
2. Test your Flow with the `cIntent:continueBooking` Postback value. The system debug message shows:
   ```text theme={null}
   Intent Match
   Flow: Slots
   Intent: Continue Booking
   Intent score: 1
   ```
3. Test your Flow with the `cIntent:continueBooking|book tomorrow` to verify Slot extraction, for example, `tomorrow`. The system debug message shows:
   ```text theme={null}
   System Slot Match
   Start: February 26, 2025 (2025-02-26T00:00:00+01:00)
   Text: tomorrow
   Slot: DATE
   ```

## More Information

* [Intents](/ai/platform-features/nlu/intents/overview)
* [Slots](/ai/platform-features/nlu/slots/overview)


# Yes/No Intents
Source: https://docs.cognigy.com/ai/platform-features/nlu/intents/yes-no-intents



<a href="/release-notes/4.98"><Badge>Updated in 4.98</Badge></a>

The *Yes/No Intents* provide a machine learning-powered feature to detect `yes` and `no` responses similarly to traditional Intent matching, enabling a more natural and fluent conversational experience. Training a model with a small set of example sentences recognizes various `yes` and `no` phrases, enhancing an AI Agent's ability to understand user confirmations or denials.

The [Question](/ai/agents/develop/node-reference/basic/question), [Optional Question](/ai/agents/develop/node-reference/basic/optional-question) and [Code](/ai/for-developers/code/overview) Nodes support the Yes/No Intents feature.

## Key Benefits

* **Improved Natural Language Understanding**. Unlike the earlier [Confirmation Words](/ai/platform-features/nlu/intents/confirmation-words) feature, which required manual synonym definitions, Yes/No Intents use a preinstalled vocabulary that can be expanded. This approach makes it easier to detect diverse expressions.
* **Enhanced User Experience**. Yes/No Intents enable AI Agents to respond accurately to a broader range of affirmative or negative responses, reducing misunderstandings and improving conversations.

## Restrictions

* Yes/No Intents don't support the Intent Trainer.
* Yes/No Intents don't provide training traffic lights feedback, which may limit visibility into model performance.

## How to Use

To set up this feature, do the following:

<AccordionGroup>
  <Accordion title="Select a Locale">
    Each Locale has its own Yes/No Intents model, requiring separate training. The Universal Locale includes English examples, but additional examples should be added for better accuracy in other languages.

    To configure a locale, do the following:

    <Tabs>
      <Tab title="GUI">
        1. To see the Project's configured locales, click **Manage > Localization**.
        2. Select the locale and click <img alt="menu" />, then **Edit Yes/No Intents**.
        3. In the **Edit Yes/No Intents** pane, you can configure Intents for the current locale:
           * **Yes Intent** — by default, the setting is toggled on. When the setting is toggled off, the Intent is excluded from the build.
           * **No Intent** — by default, the setting is toggled on. When the setting is toggled off, the Intent is excluded from the build.
           * **Reject Intent** — by default, the setting is toggled on. You need to add examples manually. When the setting is toggled off, the Intent is excluded from the build.
        4. Additionally, for Yes and No Intents, you can add [rules](/ai/platform-features/nlu/intents/rules).
        5. Click **Save & Build** to build a model. The Yes/No model is always built when you edit and save settings.
      </Tab>

      <Tab title="API">
        To create or update sentences for a specific Locale, use the following Cognigy API requests:

        * **Create**: [POST /v2.0/locales//yesnointents//sentences](https://api-trial.cognigy.ai/openapi#post-/v2.0/locales/-localeId-/yesnointents/-intentId-/sentences)
        * **Update**: [PATCH /v2.0/locales//yesnointents//sentences/](https://api-trial.cognigy.ai/openapi#patch-/v2.0/locales/-localeId-/yesnointents/-intentId-/sentences/-sentenceId-)
      </Tab>
    </Tabs>
  </Accordion>

  <Accordion title="Use Yes/No Intents for Specific or All Flows">
    Select how you want to apply the Yes/No Intents feature: to the entire Project or to the selected Flow.

    <Tabs>
      <Tab title="Project-Level Settings">
        To use Yes/No Intents for all Flows, do the following:

        1. In the left-side menu of the Project, go to **Manage > Settings**.
        2. On the **Settings** page, click **NLU Settings**.
        3. In the **General Flow Logic** section, select one of the following options from the **Yes/No Logic** list:
           * **Yes/No Intents with extended rules** — evaluates the standard Confirmation words logic first. If the standard Confirmation Words logic fails to determine the input type, the Yes/No Intents model is evaluated.
           * **Yes/No Intents** — evaluates the Yes/No Intents logic for every user input. The standard Confirmation Words logic is only used as a fallback when no Yes, No, or Reject Intents from the Yes/No Intents model are triggered.
        4. If you selected **Yes/No Intents – with extended rules** or **Yes/No Intents**, configure **Yes/No threshold**. The confidence threshold slider sets the minimum score required for Yes/No Intents to trigger. A Yes, No, or Reject Intent will only be triggered if the Intent score is greater than or equal to this threshold. This value applies to Yes/No Intents models for all locales. You can change this value for a specific Flow.
        5. Click **Save** to apply settings and build a model. The Yes/No model is always built when you edit and save settings.

        If you have set settings at the project level but want to change them for certain Flows, you should go to the Flow-level settings.
      </Tab>

      <Tab title="Flow-Level Settings">
        By default, all Flows inherit the **Yes/No Logic** and the **Yes/No threshold** settings from the project-level configuration. You can override these settings for a specific Flow.
        To use Yes/No Intents for a specific Flow, do the following:

        1. Open the existing Flow.
        2. In the upper-right corner of the **Flow Editor** page, select **Settings**.
        3. On the **Configuration** tab, in **General Flow Logic** activate checkbox near the **Yes/No Logic** setting.
        4. From the **Yes/No Logic** list, select one of the following options:
           * **Yes/No Intents with extended rules** — evaluates the standard Confirmation Words logic first. If the standard Confirmation Words logic fails to determine the input type, the Yes/No Intents model is evaluated.
           * **Yes/No Intents** — evaluates the Yes/No Intents logic for every user input. The standard Confirmation Words logic is only used as a fallback when no Yes, No, or Reject Intents from the Yes/No Intents model are triggered.
        5. If you selected **Yes/No Intents – with extended rules** or **Yes/No Intents**, configure **Yes/No threshold**. The confidence threshold slider sets the minimum score required for Yes/No Intents to trigger. A Yes, No, or Reject Intent will only be triggered if the Intent score is greater than or equal to this threshold. This value applies to Yes/No Intents models for all locales in the current Flow.
        6. Click **Save** to apply settings and build a model. The Yes/No model is always built when you edit and save settings.
      </Tab>
    </Tabs>
  </Accordion>
</AccordionGroup>

## How to Test

Check if the Yes/No Intent feature works as expected via the [Interaction Panel](/ai/agents/test/interaction-panel/overview):

1. In the existing Flow, open the Interaction Panel by clicking <img alt="interaction-panel" /> **Chat with your Agent** in the upper-right corner of the page.
2. Run a Flow.
3. On the **INFO** tab, select **Input** to view JSON. In the `yesNoIntentResults` object, you will see the result of Yes/No Intents model evaluation.

<AccordionGroup>
  <Accordion title="Examples of possible results">
    <Tabs>
      <Tab title="Yes Intent">
        ```json theme={null}
        "nlu": {
          "yesNoIntentResults": {
              "finalIntentName": "yesIntent",
              "finalIntentScore": 0.8887347172756052,
              "scores": [
                    {
                    "id": "a396391d-3cab-43ab-8b8a-cefd0135041b",
                    "name": "yesIntent",
                    "score": 0.8887347172756052
                    },
                    {
                    "id": "fa2ab1d3-b5cd-4021-91c6-8b642ec51dd8",
                    "name": "noIntent",
                    "score": 0.32411091251643426
                    }
              ]
            }
        },
        "mode": "TextOnly",
        "type": "pAnswer"
        ```
      </Tab>

      <Tab title="No Intent">
        ```json theme={null}
        "nlu": {
          "yesNoIntentResults": {
              "finalIntentName": "noIntent",
              "finalIntentScore": 0.9659235010108965,
              "scores": [
                    {
                    "id": "fa2ab1d3-b5cd-4021-91c6-8b642ec51dd8",
                    "name": "noIntent",
                    "score": 0.9659235010108965
                    },
                    {
                    "id": "a396391d-3cab-43ab-8b8a-cefd0135041b",
                    "name": "yesIntent",
                    "score": 0.021069213045105353
                    },
                    {
                    "id": "c646dd2b-f7d2-45d8-8451-cae95142c81f",
                    "name": "rejectIntent",
                    "score": 0.0034825030367928644
                    }
              ]
          }
        },
        "mode": "TextOnly",
        "type": "nAnswer"
        ```
      </Tab>

      <Tab title="Reject Intent">
        ```json theme={null}
        "nlu": {
          "yesNoIntentResults": {
              "finalIntentName": "rejectIntent",
              "finalIntentScore": 0.7307951831435443,
              "scores": [
                    {
                    "id": "c646dd2b-f7d2-45d8-8451-cae95142c81f",
                    "name": "rejectIntent",
                    "score": 0.7307951831435443
                    },
                    {
                    "id": "a396391d-3cab-43ab-8b8a-cefd0135041b",
                    "name": "yesIntent",
                    "score": 0.03191703050840985
                    },
                    {
                    "id": "fa2ab1d3-b5cd-4021-91c6-8b642ec51dd8",
                    "name": "noIntent",
                    "score": 0.007946727310282035
                    }
                  ]
              }
        },
        "mode": "TextOnly",
        "type": "Statement"
        ```
      </Tab>

      <Tab title="Fallback Reject Intent">
        ```json theme={null}
        "nlu": {
          "yesNoIntentResults": {
              "finalIntentName": null,
              "finalIntentScore": null,
              "scores": [
                    {
                    "id": null,
                    "name": "null - Fallback Reject Intent",
                    "score": 0.7277484634707633
                    },
                    {
                    "id": "a396391d-3cab-43ab-8b8a-cefd0135041b",
                    "name": "yesIntent",
                    "score": 0.207322792467384
                    },
                    {
                    "id": "c646dd2b-f7d2-45d8-8451-cae95142c81f",
                    "name": "rejectIntent",
                    "score": 0.018095447571065546
                    }
              ]
            }
        },
        "mode": "TextOnly",
        "type": "Statement"
        ```

        Depending on the **Yes/No Logic** setting, the Yes/No Intents model may not be evaluated. In that case, the `yesNoIntentResults` has the default values:

        ```json theme={null}
        "yesNoIntentResults": {
          "finalIntentName": null,
          "finalIntentScore": null,
          "scores": []
        }
        ```

        Yes/No Intents are evaluated independently of regular Intents
        and don't overwrite the existing fields for Intents in the Input object.
        Evaluation of Yes/No Intents only affects the `type` and `yesNoIntentResults`.
      </Tab>
    </Tabs>
  </Accordion>

  <Accordion title="FAQ">
    **Q1**: What happens if the **Yes/No** logic is set to **Yes/No Intents**, but both `yes` and `no` Intents have been individually disabled in the Locale settings?

    **A1**: The standard **Confirmation Words** logic is used.

    **Q2**: Are Yes/No Intents independent of standard intent scoring?

    **A2**: Yes, the Yes/No Intents are completely independent of standard intent scoring (`input.intentScore` or `input.nlu.intentMapperResults`). They specifically find the types `nAnswer` and `yAnswer` within the Flow. These types are used for the **Yes/No** type in [Question Nodes](/ai/agents/develop/node-reference/basic/question) and for [confirming Intents](/ai/platform-features/nlu/intents/overview), similar to [confirmation words](/ai/administer/access/project-settings#nlu-settings).

    **Q3**: How are Yes/No Intents evaluated when the Execute Flow node is used with and without the **Parse Intents** toggled on?

    **A3**: Yes/No Intent scoring behavior differs only if the Yes/No Intent settings are different in the two Flows.

    For example:

    Prerequisites:

    * Flow A has `Yes/No Intents` and only one node, which is an Execute Flow node that executes Flow B.
    * Flow B has `Yes/No Intents - with extended rules` and one Say node.

    Scenario 1:

    1. The Parse Intents setting is disabled in the Execute Flow node of Flow A.
    2. When the user sends `yes` in Flow A, Yes Intent is triggered in Flow B.
    3. Flow B inherits the Yes/No Logic from Flow A.

    Scenario 2:

    1. The Parse Intents setting is enabled in the Execute Flow node of Flow A.
    2. When the user sends `yes` in Flow A, Yes Intent is not triggered in Flow B.
    3. Flow B uses its own Yes/No Logic.
  </Accordion>
</AccordionGroup>

## Example

Imagine a user interacting with an AI Agent in a Star Wars quiz:

**Without Yes/No Intents**:

```txt theme={null}
User: Hello! 
AI Agent: Do you want to play the Star Wars quiz?
User: Why not
AI Agent: Not sure I understood this correctly. Try again.
```

**With Yes/No Intents**:

```txt theme={null}
User: Hello! 
AI Agent: Do you want to play the Star Wars quiz?
User: Why not
AI Agent: Let's start!
```

## More Information

* [Confirmation Words](/ai/platform-features/nlu/intents/confirmation-words)


# NLU Languages
Source: https://docs.cognigy.com/ai/platform-features/nlu/language-support



*NLU Languages* refer to the languages supported by Cognigy NLU. It supports over 100 languages,
enabling the processing of user inputs to identify [Intents](/ai/platform-features/nlu/intents/overview) and extract Slots.
Cognigy NLU can accurately detect user intentions and capture specific information,
adapting to the distinct characteristics and nuances of each supported language.

For instance, Cognigy NLU automatically adjusts to language-specific conventions,
such as varying date formats or currency symbols, based on the user language.
This flexibility eliminates manual adjustments, making it a powerful tool for creating multilingual AI Agents.

## Language Support in Cognigy Installations

The process for installing languages depends on your installation type:

* **Shared SaaS**. On `trial`, `app`, and `au-02`, you already have access to a preinstalled list of supported languages. However, if you're on `app-us` and `app-uk` and want to add a language other than English or German, contact [Cognigy technical support](/help/get-help).
* **Dedicated SaaS**. To extend the initial list of supported languages, contact [Cognigy technical support](/help/get-help).
* **On-Premises**. To add languages to your environment, refer to the [Which Languages to Deploy](/ai/administer/installation/migration/from-nlu-v1-to-v2-migration#choose-which-languages-to-deploy) section.

## List of Supported Languages

In your Project, under **Manage > Localization**,
you can select from 28 popular NLU languages and the **Universal** Locale,
which is designed for multilingual support.

<Accordion title="List of Supported Languages in the UI">
  * Arabic
  * Bengali
  * Chinese (simplified)
  * Danish
  * Dutch
  * English (Australia)
  * English (Canada)
  * English (India)
  * English (UK)
  * English (US)
  * German
  * Hindi
  * Japanese
  * Korean
  * Finnish
  * French
  * Italian
  * Norwegian
  * Polish
  * Portuguese (Brazil)
  * Portuguese
  * Russian
  * Spanish
  * Swedish
  * Tamil
  * Thai
  * Turkish
  * Vietnamese
</Accordion>

### Universal Locale

The *Universal Locale* supports key features such as Intent detection, Slot extraction, and language-specific processing.

This Locale enables Intent mapping in any language.
Combined with Slot detection and [Intent rules](/ai/platform-features/nlu/intents/rules), the Universal Locale helps you create Flows that work across multiple languages.

<Accordion title="Universal Locale">
  * Afrikaans
  * Albanian
  * Arabic
  * Azerbaijani
  * Bahasa Indonesia
  * Bahasa Melayu
  * Belarusian
  * Bengali
  * Bosnian
  * Bulgarian
  * Catalan
  * Chinese (Traditional / Cantonese)
  * Croatian
  * Czech
  * Danish
  * Dutch
  * English
  * Estonian
  * Finnish
  * French
  * Georgian
  * German
  * Greek
  * Gujarati
  * Haitian
  * Hebrew
  * Hindi
  * Hungarian
  * Icelandic
  * Indonesian
  * Irish
  * Italian
  * Japanese
  * Kazakh
  * Korean
  * Latin
  * Latvian
  * Lithuanian
  * Macedonian
  * Malay
  * Norwegian
  * Persian (Farsi)
  * Polish
  * Portuguese
  * Punjabi
  * Romanian
  * Russian
  * Serbian
  * Sicilian
  * Slovak
  * Slovenian
  * Spanish
  * Sundanese
  * Swedish
  * Tamil
  * Thai
  * Turkish
  * Ukrainian
  * Urdu
  * Uzbek
  * Vietnamese
  * Laotian
  * Sinhalese
  * and more
</Accordion>

#### Restrictions

The Universal Locale enables Intent mapping across languages but works best with space-tokenized languages like English,
where spaces separate words.
For example, Keyphrase detection uses spaces `(" ")` to identify words or phrases.
This distinction is important because languages like Chinese, Japanese, and Thai don't use spaces, which can affect the accuracy of Keyphrase detection.

## Yes/No Intents and Confirmation Words

To ensure Question Nodes function correctly, Cognigy NLU must recognize words that indicate positive or negative confirmations.
This recognition is predefined for some languages. For example, with the English Locale, the Question Node recognizes `yes` as a positive confirmation and `no` as a negative confirmation.
For the Universal Locale feature, you must define these confirmations using [Yes/No](/ai/platform-features/nlu/intents/yes-no-intents) Intents or confirmation words.
Otherwise, the AI Agent won't recognize confirmations in other languages.

## Rules

[Rules](/ai/platform-features/nlu/intents/rules) allow you to define custom conditions using [CognigyScript](/ai/platform-features/cognigyscript) to determine whether an Intent is triggered.
For example, you can create a rule `input.slots.food[0].keyphrase === "pizza"`, meaning that the Intent will be triggered whenever the user mentions the word `pizza`.

## More Information

* [Intents](/ai/platform-features/nlu/intents/overview)


# NLU Overview
Source: https://docs.cognigy.com/ai/platform-features/nlu/overview



*Cognigy Natural Language Understanding (Cognigy NLU)* is the core engine of Cognigy that allows AI Agents to interpret human language. Cognigy NLU processes text and speech to determine meaning, identify user queries, and understand context, enabling more humanlike conversations.

<CardGroup>
  <Card title="Intents" href="/ai/platform-features/nlu/intents/overview">
    Learn the fundamentals of Intents and how they're used in Cognigy to enhance conversational AI capabilities
  </Card>

  <Card title="NLU Performance Tuning" href="/ai/platform-features/nlu/intents/intent-analyzer">
    Fine-tune Intent recognition by analyzing accuracy and confidence. Adjust thresholds to optimize detection sensitivity
  </Card>

  <Card title="Rules and Conditions" href="/ai/platform-features/nlu/intents/rules">
    Define rules and conditions to refine Intent handling, ensuring accurate user input recognition and response
  </Card>

  <Card title="Slots" href="/ai/platform-features/nlu/slots/overview">
    Use Slots to extract data from user input, enabling dynamic interactions within conversations
  </Card>

  <Card title="Affirmative Words" href="/ai/platform-features/nlu/intents/confirmation-words">
    Configure words and phrases for affirmative recognition, improving AI Agents understanding capabilities
  </Card>

  <Card title="Intent Hierarchy" href="/ai/platform-features/nlu/intents/intent-hierarchy/overview">
    Structure and prioritize Intents to improve recognition and mapping accuracy
  </Card>

  <Card title="Attachments" href="/ai/platform-features/nlu/attachments">
    Attach additional resources like Flows and Lexicons to enhance Intent processing and response generation
  </Card>

  <Card title="External NLU" href="/ai/platform-features/nlu/external/overview">
    Integrate third-party NLU services for extended Intent recognition and language processing capabilities
  </Card>

  <Card title="Miscellaneous" href="/ai/platform-features/nlu/text-cleaner">
    Explore additional NLU tools
  </Card>
</CardGroup>


# Slot Fillers
Source: https://docs.cognigy.com/ai/platform-features/nlu/slot-fillers



*Slot Fillers* let you monitor every user input for specific [Slot](/ai/platform-features/nlu/slots/overview) matches
and store them in the [Context](/ai/agents/develop/ai-agent-memory/context) object.
This way, if an answer is already provided, the AI Agent skips the corresponding question in the Question Node.
For example, if a user says
`I'm flying from San Francisco on United` and the AI Agent asks
`What city are you departing from?` — this question would be redundant and create a poor user experience.
Instead, Slot Fillers ensure that only missing information is requested.

## Key Benefits

* **Collecting and Remembering Information**. With Slot Fillers, all required data is gathered and retained throughout the conversation, reducing errors and making conversations with the AI Agent smoother and more natural.
* **Handling Different Input Types**. Slot Fillers accept various formats, for example, dates and numbers, making it flexible for user inputs.

To learn more about the benefits of using Slot Fillers and how they work, watch this video:

<iframe title="Cognigy Sessions: Cognigy Slot Fillers" />

## Restrictions

* Slot Fillers fail when multiple values share the same Slot type, making it impossible to distinguish between them. For example, `I'm flying from SFO to ORD on United`. The AI Agent can't distinguish between departure airport and arrival airport because both match `airport_code`.
  To avoid this restriction, you can:
  * Create separate Slot Fillers for each value:
    * `Departure Airport` — use a Context Key, for example, `departure_city`.
    * `Arrival Airport` — use a Context Key, for example, `arrival_city`.
  * Modify the Flow to ask distinct questions for each Slot instead of extracting both from a single user input.

## Working with Slot Fillers

<Tabs>
  <Tab title="GUI">
    You can create, edit, and delete Slot Fillers in **NLU > Slot Fillers** in the Flow editor.
  </Tab>

  <Tab title="API">
    You can create, edit, and delete Slot Fillers by using [Cognigy API](https://api-trial.cognigy.ai/openapi#patch-/v2.0/flow/-flowId-/slotfillers).

    To capture and store Slot matches, ensure you specify the Slot Filler in your Question Node by selecting the **Slot** question type, then activate the **Skip if answer exists in Context** option in the Node.
  </Tab>
</Tabs>

### Slot Filler Settings

<AccordionGroup>
  <Accordion title="Required Fields">
    | Field              | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        | Example                          |
    | ------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------- |
    | Name               | The display name of the Slot Filler, used for identification.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      | `Departure Airport Slot`         |
    | Type               | Defines the type of input expected. You can select from the following Slot types: <br />- **Slot** – a [Lexicon Slot](/ai/platform-features/nlu/slots/user-defined/lexicon). Make sure a Lexicon is attached and that this Slot is included in it.<br />- **Regular Expression** – a [Regex Slot](/ai/platform-features/nlu/slots/user-defined/regex).<br />- **Intent** – extracts values based on recognized Intents. Useful when Slot filling is tied to Intent classification.<br />- **System Slot** – [predefined Slot types](/ai/platform-features/nlu/slots/system-defined) that automatically extract common entity types such as dates, numbers, temperatures, and more. | -                                |
    | Slot Name          | This parameter appears if the **Slot** type is selected. Specifies which Slot this Filler is associated with.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      | `airport_code`, `date`, `name`   |
    | Regular Expression | This parameter appears if the **Regular Expression** type is selected. Enter the regular expression to extract specific patterns from user input. Regex expressions must start and end with `/`. Flags like `g` for global matching aren't supported. Example: `/^\d{3}-\d{3}-\d{4}$/`.                                                                                                                                                                                                                                                                                                                                                                                            |                                  |
    | Context Key        | The variable name used to store the extracted value in the Context object.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         | `departure_city`, `arrival_date` |
  </Accordion>

  <Accordion title="Optional Fields">
    | Field                                    | Description                                                                                                                                                                                                                                                                                                         | Example                                               |
    | ---------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------- |
    | Additional Validation                    | Allows extra validation rules to ensure correct formatting or values.                                                                                                                                                                                                                                               | Ensure date format is `YYYY-MM-DD`.                   |
    | Result Location                          | Determines where to extract the Slot value from. By default, the Slot value is extracted from the user input. You can set a CognigyScript expression in this parameter to override the detected Slot value with a value from the [Input, Context, or Profile objects](/ai/agents/develop/ai-agent-memory/overview). | `{{input.bookingReference}}`                          |
    | Store result in Context                  | Stores the extracted value in the Context object.                                                                                                                                                                                                                                                                   | `"departure_city": "SFO"`.                            |
    | Store result in Contact Profile          | Stores the extracted value in the Contact Profile for future interactions.                                                                                                                                                                                                                                          | Store **"preferred\_airline"** for returning users.   |
    | Store Detailed Results                   | Stores metadata about the extracted Slot, for example, confidence scores.                                                                                                                                                                                                                                           | Confidence score of `0.92` for `"ORD"`.               |
    | Use Positive Keyphrases Only             | Limits Slot extraction to predefined keyphrases.                                                                                                                                                                                                                                                                    | Only accept `"San Francisco"` for `departure_city`.   |
    | Skip if the result is already in Context | Prevents asking for the Slot again if a valid value already exists in the Context object.                                                                                                                                                                                                                           | If `"departure_city": "LAX"` is set, don't ask again. |
  </Accordion>
</AccordionGroup>

## Example

<AccordionGroup>
  <Accordion title="Step 1. Add Slots">
    1. Create a new Flow called `Slot Fillers`.
    2. Navigate to the **NLU** tab and attach relevant Lexicons, for example, `airports` and `airlines`.
    3. On the **Slot Fillers** tab, follow these steps:
       1. Create a new Slot Filler named `Airline`.
       2. Select **Slot** from the **Type** list.
    4. Enter `airline` in both the **Slot name** and **Context Key** fields.
    5. Repeat the steps 3.1 and 3.2 to create another Slot Filler named `Airport`, enter `airport_code` in the **Slot name** and **Context Key** fields.
    6. Click **Build Model**.
  </Accordion>

  <Accordion title="Step 2. Configure Question Nodes">
    1. Create a Question Node, enter `What is your departure airport?` in the **Text** field and select **Slot** from the **Question Type** list.
    2. Enter `airport_code` in **Slot Name** and **Context Key to use**.
    3. Enable **Skip if answer exists in Context**.
    4. Repeat the process for `What is your airline?`, enter `airline` in the **Slot name** and **Context Key** fields.
  </Accordion>

  <Accordion title="Step 3. Test Slot Fillers">
    Provide all information at once: <br />
    **User Input**: `I'm leaving San Francisco on United.`<br />
    **Result**: `No additional questions are asked.`<br /><br />
    Provide partial information: <br />
    **User Input**: `I'm flying on United.`<br />
    **AI Agent**: `What is your departure airport?`<br />
    **User Input**: `SFO`<br />
    **Result**: The AI Agent skips asking for the airline.
  </Accordion>
</AccordionGroup>

## More Information

* [Overview](/ai/platform-features/nlu/slot-fillers)


# Slots
Source: https://docs.cognigy.com/ai/platform-features/nlu/slots/overview



*Slots* are predefined data types or custom phrases detected by the Cognigy NLU
and stored in the [Input](/ai/agents/develop/ai-agent-memory/input) object.
They allow AI Agents to extract structured information from user input and use it dynamically in conversations.

There are two types of Slots:

* [System-Defined](/ai/platform-features/nlu/slots/system-defined) — built-in Slots that Cognigy can use to automatically recognize common data types such as dates, times, numbers, and emails.
* [User-Defined](/ai/platform-features/nlu/slots/user-defined/lexicon) — custom Slots based on your vocabularies or regular expressions, allowing detection of specific phrases, entities, or domain-specific terms.

## Working with Slots

Working with Slots depends on the type of Slot you choose. While some Slots automatically extract values from the user input, others require manual configuration to map detected values to specific variables.

### Accessing Slots via CognigyScript

By default, Slots are stored in the Input object, and you can access them via [CognigyScript](/ai/platform-features/cognigyscript).

The `{{input.slots}}` object contains all system- and user-defined Slots found in a user input.

### Tokenize your Slots

Create [Tokens](/ai/platform-features/tokens) for your Slots to speed up the process of adding them to conversation logic and output fields.

## More Information

* [Slot Fillers](/ai/platform-features/nlu/slot-fillers)


# System-Defined Slots
Source: https://docs.cognigy.com/ai/platform-features/nlu/slots/system-defined



*System-Defined Slots* refer to predefined Slot types that the system automatically detects in user inputs.
These Slots are used to categorize and recognize specific types of data, such as dates, times, durations, and more.

## Key Benefits

* **Automatic Extraction**. Cognigy automatically detects and categorizes user input into predefined Slots, reducing manual configuration.
* **Data Consistency**. By using system-defined Slots, Cognigy ensures all extracted data is consistently formatted and accurate.
* **Simplified Integration**. System-defined Slots simplify the handling of complex data types, enabling smooth integration with backend systems.

## All System-Defined Slots

In the JSON object, all system-defined Slot names are capitalized for easy identification.

<Accordion title="List of System-Defined Slots">
  <div>
    <table>
      <colgroup>
        <col />

        <col />

        <col />
      </colgroup>

      <thead>
        <tr>
          <th>Slot Name</th>
          <th>User Input</th>
          <th>Slot JSON</th>
        </tr>
      </thead>

      <tbody>
        <tr>
          <td>DATE</td>
          <td>tomorrow, August 1st at 5pm</td>

          <td>
            ```json theme={null}
            "DATE": [
              {
                "start": {
                  "day": 2,
                  "month": 6,
                  "year": 2018,
                  "hour": 12,
                  "minute": 0,
                  "second": 0,
                  "millisecond": 0,
                  "text": "tomorrow",
                  "weekday": 6,
                  "dayOfWeek": "Saturday",
                  "ISODate": "2018-06-02T12:00:00Z"
                },
                "end": null
              }
            ]
            ```
          </td>
        </tr>

        <tr>
          <td />

          <td>next week</td>

          <td>
            ```json theme={null}
            "DATE": [
              {
                "start": {
                  "year": 2021,
                  "month": 12,
                  "day": 13,
                  "hour": 0,
                  "minute": 0,
                  "second": 0,
                  "milliseconds": 0,
                  "weekday": 1,
                  "dayOfWeek": "Monday",
                  "ISODate": "2021-12-13T00:00:00+01:00",
                  "plain": "December 13, 2021",
                  "grain": "week"
                },
                "end": null,
                "text": "next week"
              }
            ]
            ```
          </td>
        </tr>

        <tr>
          <td>NUMBER</td>
          <td>one hundred, 143</td>

          <td>
            ```json theme={null}
            "NUMBER": [
              100,
              143
            ]
            ```
          </td>
        </tr>

        <tr>
          <td>DURATION</td>
          <td>10 minutes 20 seconds</td>

          <td>
            ```json theme={null}
            "DURATION": {
              "hour": null,
              "minute": 10,
              "second": 20
            }
            ```
          </td>
        </tr>

        <tr>
          <td>TEMPERATURE</td>
          <td>50 degrees</td>

          <td>
            ```json theme={null}
            "TEMPERATURE": [
              50
            ]
            ```
          </td>
        </tr>

        <tr>
          <td>AGE</td>
          <td>32 years old</td>

          <td>
            ```json theme={null}
            "AGE": [
              32
            ]
            ```
          </td>
        </tr>

        <tr>
          <td>EMAIL</td>
          <td>[info@cognigy.com](mailto:info@cognigy.com)</td>

          <td>
            ```json theme={null}
            "EMAIL": [
              "info@cognigy.com"
            ]
            ```
          </td>
        </tr>

        <tr>
          <td>PERCENTAGE</td>
          <td>5 percent</td>

          <td>
            ```json theme={null}
            "PERCENTAGE": [
              5
            ]
            ```
          </td>
        </tr>

        <tr>
          <td>URL</td>

          <td>
            <a href="https://cognigy.com">
              [https://cognigy.com](https://cognigy.com)
            </a>
          </td>

          <td>
            ```json theme={null}
            "URL": [
              "https://cognigy.com"
            ]
            ```
          </td>
        </tr>

        <tr>
          <td>MONEY</td>
          <td>500 dollars</td>

          <td>
            ```json theme={null}
            "MONEY": [
              {
                "value": 500,
                "unit": "$"
              }
            ]
            ```
          </td>
        </tr>

        <tr>
          <td>DISTANCE</td>
          <td>12 kilometers</td>

          <td>
            ```json theme={null}
            "DISTANCE": [
              {
                "value": 12,
                "unit": "kilometre"
              }
            ]
            ```
          </td>
        </tr>
      </tbody>
    </table>
  </div>
</Accordion>

## DATE Slots

DATE Slot detection depends on the time zone you set in the [Project](/ai/administer/access/project-settings#nlu-settings) or through [API functions](/ai/for-developers/code/api-functions).

<Accordion title="Supported Input Patterns and Formats">
  | Languages       | Supported input pattern examples                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |
  | --------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
  | English         | Now, Today, Tomorrow, Yesterday, Last Friday, Last night<br />17 August 2013 - 19 August 2013<br />This Friday from 13:00 - 16.00<br />5 days ago<br />2 weeks from now<br />Sat Aug 17 2013 18:40:39 GMT+0900 (JST)<br />2014-11-30T08:15:30-05:30<br />Day after tomorrow<br />Days before yesterday<br />Next Tuesday                                                                                                                                                                                                                                                      |
  | German          | Jetzt, Heute, Morgen, Gestern, Vorgestern<br />Letzte Nacht, Heute Nacht, Diese Nacht, Morgen Vormittags, Gestern Abend<br />Nach einem Tag, In einer Woche<br />15/01/2021, 15.01.2021, 15. Jan 2021, Freitag den 15. Januar 2021<br />17.08.2013 - 19.08.2013, 17.08.2013 bis 19.08.2013<br />Kommenden Montag                                                                                                                                                                                                                                                              |
  | Dutch           | Nu, Deze Nacht, Deze Vannacht, Gisteren, Morgen, Gisteravond, Eergisteren, Overmorgen<br />Morgen Middag, Deze Ochtend, Gisteren Voormiddag, Deze Namiddag, Morgen Avond, Overmorgen Nacht<br />Binnen Een Dag, Na een halve Maand<br />15/01/2021, 15.01.2021, 15. Jan 2021, Vrijdag de 15de Januari 2021<br />17.08.2013 - 19.08.2013, 17.08.2013 tot 19.08.2013<br />Laatste/vorige/volgende/komende Woensdag                                                                                                                                                              |
  | Japanese        | 今日, 当日, 昨日, 明日, 今夜, 今夕, 今晩, 今朝<br />２０１６年２月１７日です<br />Sat Aug 17 2013<br />01/30/2018                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         |
  | French          | 15.04.2023 12:00<br />15 aout 2015<br />15 aout 2015 a 12 octobre 2016<br />aujourd'hui, demain, hier<br />hier soir, demain après-midi <br />avant-hier<br />vendredi prochain                                                                                                                                                                                                                                                                                                                                                                                               |
  | Other languages | For languages like Finnish, Cognigy supports standard date formats but has limited support for casual expressions. For example: <br />ylihuomenna (the day after tomorrow) — the system correctly matches it to a DATE Slot, but inconsistencies may arise due to limitations with casual expressions.<br />kolme viikkoa sitten (three weeks ago) — the system identifies it as a DURATION Slot instead of a DATE Slot.<br />kuukauden päästä (in one month) — the system doesn't recognize this input, highlighting a gap in handling relative time expressions in Finnish. |
</Accordion>

### DATE Format Handling

DATE Slot parsing follows the NLU language set in the Flow.
Different regions use different date formats for slash-separated dates.
The table outlines the format handling:

| Languages                                         | Slash date format ordering                                                                                                                                                                       |
| ------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| Default (English UK, German, Universal, and more) | Dates follow day/month/year (`dd/mm/YYYY`) order.                                                                                                                                                |
| English US, Japanese, Chinese                     | Dates follow month/day/year (`mm/dd/YYYY`) order. However, if the system detects an invalid date, for example, `13/01/2023` in `mm/dd/YYYY` format, it automatically falls back to `dd/mm/YYYY`. |

For a full list of country-specific date formats, refer to the [Date\_format\_by\_country](https://en.wikipedia.org/wiki/Date_format_by_country) page on Wikipedia.

### Invalid Date Handling

Cognigy automatically checks all dates from user inputs.
If a date is invalid, it won't be assigned to the DATE Slot.

The system ensures:

* Leap year validation: `February 29, 2024` is valid, but `February 29, 2023` isn't.
* Correct month-day mapping: `April 31` is invalid.
* Fallback correction for ambiguous formats: `13/01/2023` defaults to `dd/mm/YYYY` if `mm/dd/YYYY` is invalid.

### Future and Past Date Mapping

DATE Slot detection applies the current year by default unless specified otherwise. The system follows these rules:

| **Scenario**                                           | **User Input**     | **Today's Date** | **Processing Logic**                                                                                                                                         | DATE Slot Result                      |
| ------------------------------------------------------ | ------------------ | ---------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------- |
| Date is in the current month, no year provided         | `"July 21"`        | July 26, 2023    | Since the input date falls within the current month and no year is provided, the system assumes the current year.                                            | `"year": 2023, "month": 7, "day": 21` |
| Date is in a past month, no year provided              | `"March 30"`       | July 26, 2023    | The input date has already passed in the current year. Since no year is specified, the system assumes it refers to the next occurrence in the upcoming year. | `"year": 2024, "month": 3, "day": 30` |
| Date is in a past month, with a specific year provided | `"March 30, 2021"` | July 26, 2023    | The input includes a specific year, so the system directly assigns that year without modification.                                                           | `"year": 2021, "month": 3, "day": 30` |

## More Information

* [Overview](/ai/platform-features/nlu/slots/system-defined)
* [User-Defined Slots (Lexicons)](/ai/platform-features/nlu/slots/user-defined/lexicon)


# Lexicon Slots
Source: https://docs.cognigy.com/ai/platform-features/nlu/slots/user-defined/lexicon



<a href="/release-notes/4.52"><Badge>Updated in 4.52</Badge></a>

*Lexicons* are structured vocabularies that help your AI Agent recognize specific words or phrases, known as Keyphrases, in user inputs. Within a Lexicon, you define Keyphrases along with their categories and equivalent words. Keyphrases can include examples like product names, car brands, or zip codes.

*Lexicon Slots* are user-defined categories within a Lexicon. You assign a Lexicon Slot to a Keyphrase, to classify the type of information it represents. This approach provides contextual meaning for the system to understand the Keyphrase's role in the conversation.

## Key Benefits

* **Improved Recognition of User Inputs**. Lexicon Slots allow the AI Agent to recognize variations of Keyphrases (Synonyms), enabling flexible user input handling without requiring exact matches.
* **Consistency Across Conversations**. Lexicon Slots ensure uniform interpretation of terms across different Flows, improving reliability.
* **Easier Maintenance**. Centralized Lexicon management makes it easy to update Lexicon Slots, reducing errors.

## Limitations

* Maximum of 250 Lexicon Slots per Lexicon.
* Maximum of 30 characters per Lexicon Slot name.
* Maximum of 1 Keyphrase and 199 Synonyms per Lexicon entry (row).
* Maximum of 100K entries per Lexicon.

## Key Components of a Lexicon

<Frame>
  <img alt="Lexicon editor with keyphrases, synonyms, slots, and data columns" />
</Frame>

| **Component**              | **Description**                                                                                                                                                                                                                                                                                                                                                                                  | **Example**                                                                                                 |
| -------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ----------------------------------------------------------------------------------------------------------- |
| Keyphrase                  | A word or phrase that the NLU should recognize. A Keyphrase acts as the main entry in the Lexicon.                                                                                                                                                                                                                                                                                               | `"Apple iPhone"`                                                                                            |
| (Lexicon) Slots            | Categorizes a Keyphrase to provide context in a conversation. They help define what type of entity the Keyphrase represents. <br /> You can access detected Slots in the Flow using the Input object with the following syntax: `{{input.slots.SlotName}}`, where `SlotName` is the name of the Slot. For example, if the Slot name is device, you can reference it as `{{input.slots.device}}`. | `"device": true, "mobile": true` > `"Apple iPhone"` is categorized as both `device` and `mobile`.           |
| Synonyms                   | Allows the AI Agent to recognize variations of a Keyphrase. If a user types a Synonym instead of the Keyphrase, the system still understands it as the same Keyphrase.                                                                                                                                                                                                                           | `"Cool phone"` > Recognized as `"Apple iPhone"`                                                             |
| Data (Metadata/Attributes) | Stores additional structured data related to a Keyphrase, such as product IDs, brands, or custom attributes.                                                                                                                                                                                                                                                                                     | `{ "id": "12345", "brand": "Apple", "releaseYear": 2023 }` > `"Apple iPhone"` is linked to product details. |

## Working with Lexicon Slots

Lexicon Slots are managed within Lexicons, which you can create and configure in Cognigy to support conversational Flows.

<Tabs>
  <Tab title="GUI">
    You can create, download, and upload Lexicons in **Build > Lexicons**.
    Also, you can generate Lexicons by using [Generative AI](/ai/agents/develop/gen-ai-and-llms/generative-ai).
  </Tab>

  <Tab title="API">
    You can add, download, and upload Lexicons using the [Cognigy API](https://api-trial.cognigy.ai/openapi#get-/v2.0/lexicons).
  </Tab>

  <Tab title="CLI">
    You can clone, restore, push, pull, and diff Lexicons using the [Cognigy CLI](https://github.com/Cognigy/Cognigy-CLI/blob/main/README.md).
  </Tab>
</Tabs>

### Attach Lexicons for Slot Usage

Lexicon Slots aren't automatically recognized by the NLU model.
To use them in a Flow, you must [manually attach a Lexicon to the Flow](/ai/platform-features/nlu/attachments).
Once the Lexicon is attached, Slots are populated when Keyphrases or their synonyms are detected in user input.

### Upload and Download Lexicons

For easy Slot management, you can upload and download Lexicons in CSV format.

<AccordionGroup>
  <Accordion title="CSV File Requirements">
    To upload the file, make sure your CSV file is UTF-8 encoded.

    The CSV file should have the following format:

    ```txt theme={null}
    KEYPHRASE,"TAGS","SYNONYMS","{DATA}"
    ```

    Example:

    ```txt theme={null}
    Keyphrase1,"tag","synonym1","{""key1"":""value""}"
    Keyphrase2,"tag,tag2","synonym2,synonym2_2","{""key2"":""value""}"
    Keyphrase3,"tag,tag2,tag3","synonym3,synonym3_2,synonym3_3","{""key3"":""value""}"
    ```

    Fields beginning with the characters `@`, `+`, `-`, or `=` are not allowed.
  </Accordion>

  <Accordion title="Upload Lexicons">
    To upload Lexicons, follow these steps:

    1. In the upper-right corner of the Lexicon editor, click <img alt="upload" />.
    2. In the **Import Lexicon CSV file** window, select one of the following options:
       * **Skip importing duplicate Keyphrases** — the option skips importing any Synonyms from the CSV file that already exist in the current Lexicon. If there are duplicate Synonyms in the CSV file, they will be ignored during the import process. Only new and unique Synonyms will be imported.
       * **Overwrite the current Keyphrases** — the option replaces all the current Keyphrases in the Lexicon with the Keyphrases from the CSV file. If there are any existing keyphrases in the Lexicon, they will be completely overwritten, and only the Keyphrases from the CSV file will be retained.
       * **Merge Tags & Synonyms from the current and imported Keyphrases** — the option allows merging the Synonyms of Keyphrases from both the current Lexicon and the CSV file. Only unique Synonyms will be merged. If a Synonym exists in both the current Lexicon and the CSV file, this Synonym will not be merged.
    3. Click **Choose CSV file**.

    To track the uploading process, you can view the **Import Lexicon** task by clicking <img alt="task-menu" /> in the upper-right corner.
  </Accordion>

  <Accordion title="Download Lexicons">
    To download Lexicons, follow these steps:

    1. In the upper-right corner of the Lexicon editor, click <img alt="download" />. Wait for the CSV file to be generated for downloading. To track this process, you can view the **Export Lexicon** task by clicking <img alt="task-menu" /> in the upper-right corner.
    2. When the task is completed, click <img alt="cloud" /> in the upper-right corner of the Lexicon editor.

    The file will be downloaded.
  </Accordion>
</AccordionGroup>

## More Information

* [Regex](/ai/platform-features/nlu/slots/user-defined/regex)
* [System-Defined Slots](/ai/platform-features/nlu/slots/system-defined)
* [Overview](/ai/platform-features/nlu/slots/overview)


# User-Defined Slots
Source: https://docs.cognigy.com/ai/platform-features/nlu/slots/user-defined/overview



*User-Defined Slots* are custom Slots based on your vocabularies or regular expressions, allowing detection of specific phrases, entities, or domain-specific terms.

In Cognigy, you can create the following user-defined Slots:

* [Lexicon Slots](/ai/platform-features/nlu/slots/user-defined/lexicon) — specific words or phrases (and their variations) from a custom word list.
* [Regex Slots](/ai/platform-features/nlu/slots/user-defined/regex) — structured patterns, like phone numbers or emails, using regular expressions.


# Regex Slots
Source: https://docs.cognigy.com/ai/platform-features/nlu/slots/user-defined/regex



*Regex Slots* are user-defined Slots, which apply [regular expressions](https://regex101.com/) (regex) to match and extract specific patterns from user inputs.

Unlike Lexicon Slots, which rely on Keyphrases, Regex Slots are ideal for structured data that follows a predictable format, such as:

* Phone numbers (`(123) 456-7890`)
* Email addresses (`user@example.com`)
* Product codes (`ABC-123-XYZ`)

You can use this type of Regex Slots only as [Slot Fillers](/ai/platform-features/nlu/slot-fillers).

## Recommendations

* Regex Slots require careful crafting to avoid overly broad or restrictive patterns.
* Complex regular expressions can be error-prone; test them thoroughly to avoid missing matches or capturing incorrect data.
* Avoid conflicts with Lexicon Slots and System Slots by ensuring unique Slot names.

## Working with Regex Slots

To start working with Regex Slots, you can use one of the following options:

<Tabs>
  <Tab title="Apply to an Entire Flow">
    To define a regex pattern that runs on all user inputs in the Flow, follow these steps:

    1. In the Flow editor, go to **NLU > Slot Fillers**.
    2. In the upper-left corner, click **+ New Slot Filler**.
    3. In the **New Slot Filler** window, fill in the following fields:
       * **Name** – enter a unique name for the Slot Filler. For example, `phone_number`.
       * **Type** – select **Regular Expression**.
       * **Regular Expression** – enter a regex pattern. Patterns must start and end with `/`. Flags like `g` for global matching aren't supported.
       * **Context Key** – enter a Slot name to store the matched data in the Context object. For example, `phone_number`.
  </Tab>

  <Tab title="Apply After a Specific Point in a Flow">
    To define a regex pattern that runs after a specific user input in the Flow, follow these steps:

    1. In the Flow editor, add a [Regex Slot Filler](/ai/agents/develop/node-reference/ai/regex-slot-filler) Node where you want to process the input, for example, after a Question Node.
    2. In the Node editor, configure the following fields:
       * **Regular Expression** - enter a regex pattern without leading and trailing slashes, for example, `^\d{3}-\d{3}-\d{4}$`.
       * **Flags** – enter flags you want to apply to the regular expression, for example, `i` for case-insensitive, `m` for multiline. Note that the `g` flag should always be specified.
       * **Slot Name** – enter the Slot name to store the matched data in the `slots.<slot-name>` Input object. For example, `phone_number`.

    Once a Regex Slot is filled, the extracted data can be used in the conversation logic, for example, for the following purposes:

    * Validating user input, for example, ensuring a valid email format.
    * Triggering specific actions, for example, saving a phone number to the Contact Profile.
    * Combining with Question Nodes to prompt users to re-enter missing or incorrectly formatted information.
  </Tab>
</Tabs>

### Example

These examples show how regex patterns fill Slots in the Context object.

<AccordionGroup>
  <Accordion title="Single Match">
    **User Input**: `My phone number is 555-123-4567`<br />
    **Regex Pattern**: `/\d{3}-\d{3}-\d{4}/`<br />
    **Slot Name**: `phone_number`<br />
    **Result**: The `context.phone_number` Slot is filled with `555-123-4567`. The AI Agent responds `Thanks for providing your phone number: {{context.slots.phone_number}}!`
  </Accordion>

  <Accordion title="Multiple Matches">
    When the user input contains multiple values that match the pattern, only the first match is stored.

    **User Input**: `The booking codes are 123456 and 789123`<br />
    **Regex Pattern**: `/\b\d{6}\b/`<br />
    **Slot Name**: `booking_codes`<br />
    **Result**: The `context.booking_codes` Slot is filled with `123456`.
  </Accordion>
</AccordionGroup>

## Use Cases

* Extracting a booking code: `/^[A-Z]{3}-\d{4}$/` to capture `XYZ-1234`.
* Capturing email addresses with the `@company` domain: `/^[a-zA-Z0-9._%+-]+@company\.com$/`.
* Identifying dates in specific formats: `/^\d{2}\/\d{2}\/\d{4}$/` for `04/22/2025`.

## More Information

* [Lexicon Slots](/ai/platform-features/nlu/slots/user-defined/lexicon)
* [Overview](/ai/platform-features/nlu/slots/user-defined/overview)
* [System-Defined](/ai/platform-features/nlu/slots/system-defined)


# Text Cleaner
Source: https://docs.cognigy.com/ai/platform-features/nlu/text-cleaner



<a href="/release-notes/4.58"><Badge>Updated in 4.58</Badge></a>

Cognigy helps you clean up user input (text or voice) with built-in functions. For example, if a user says `alpha tango hotel double seven three nine minus L for London and C for Canada`, it can automatically be converted to `ath 7739 - lc`.

These Text Cleaner functions support the English (`en`) and German (`de`) locales and are available in:

* [Clean Text Node](/ai/agents/develop/node-reference/ai/clean-text)
* [Question Nodes](/ai/agents/develop/node-reference/basic/question)
* [Code Node](/ai/for-developers/code/modules)
* [Transformers](/ai/for-developers/transformers/overview#text-cleaner-class)

To learn more about the benefits of using Text Cleaner functions and how they work, watch this video:

<iframe title="Cognigy Sessions: Cognigy Slot Fillers" />

## Functions

| Function Name                      | Description                                                                                                                                                                                                 |
| ---------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| cleanDisallowedSymbols             | Removes all symbols that are not explicitly allowed. All letters and numbers are allowed by default, and additional allowed symbols can be set.                                                             |
| resolveSpelledOutNumbers           | Replaces all number words with their numerical representation. For example, "five and three hundred nineteen" > "5 319".                                                                                    |
| resolvePhoneticAlphabet            | Detects and replaces all words that are part of the phonetic alphabet. For example, "alpha tango lima" > "a t l".                                                                                           |
| replaceSpecialWords *(Deprecated)* | Replaces specifically set words with their replacements. For example, "lufthansa" > "lh". <br /><br /> As of Cognigy 4.58.0, this function is deprecated. Use the `replaceSpecialPhrases` function instead. |
| replaceSpecialPhrases              | Replaces specific words or phrases with replacements. For example, "Conversational AI" is changed to "CAI", "New York" to "NY", and "lufthansa" to "lh".                                                    |
| resolveSpelledOutAlphabet          | Resolves phrases like "a for anton b as in bertram" to "a b".                                                                                                                                               |
| resolvePhoneticCounters            | Resolves strings like "3 times 2" to "222" or "double 4" to "44". Learn more about special rules in the [Rules for resolvePhoneticCounters](#rules-for-resolvephoneticcounters) section.                    |
| contractSingleCharacters           | Joins all single characters standing alone into a full string. For example, "my name is c o g n i g y" > "my name is cognigy".                                                                              |
| contractNumberGroups               | Joins all numbers standing next to each other. For example, "his number is 333 43 22 44" > "his number is 333432244".                                                                                       |
| trimResult                         | Trims the start and end of the string and replaces all double (or more) spaces with single spaces.                                                                                                          |

### Rules for resolvePhoneticCounters

This function is designed to convert phrases like "3 times 2" into their numeric representation, such as "3 times 2" to "222" or "double 4" to "44." The function handles sentences that include both a multiplier (for example, "3 times" or "double") and a multiplicand (for example, "2" or "4"). The multiplicand can be either a number or a character.

Some examples can be ambiguous. For instance, consider the phrase "double ap 3 4", which could be interpreted as either "aap34" or "apap34."

To avoid this ambiguity, Cognigy defined the following rules for multiplicand types:

| Multiplicand type                                                        | Rule                                                                                                                                                                                      | Example                                                                                        |
| ------------------------------------------------------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------- |
| **Number**                                                               |                                                                                                                                                                                           | 2 times **3**, 2 times **16**                                                                  |
| Multiplicand is a number less than 13                                    | The multiplicand is repeated itself.                                                                                                                                                      | "2 times 3" will return "33" <br /> "2 times 11" will return "1111"                            |
| Multiplicand is a number greater than or equal to 13                     | Only the first digit is repeated. <br /><br /> The first digit is given priority due to the more common usage of phrases like "3 times 4" or "3 times 12," compared to "3 times 25".      | "2 times 16" will return "116"                                                                 |
| **Character**                                                            |                                                                                                                                                                                           | 2 times **alpha**, 2 times **ox**                                                              |
| Multiplicand is a phonetic character                                     | It is resolved, and the character is repeated.                                                                                                                                            | "2 times alpha" will return "aa"                                                               |
| Multiplicand has more than one character and is not a phonetic character | Only the first character is repeated. <br /><br /> Cognigy assumes that the speech-to-text (STT) system transcribed the user input incorrectly and that the user intended to say another. | "2 times ox" will return "oox" <br /> (we assume that the user intended to say "2 times o, x") |


# Platform Features
Source: https://docs.cognigy.com/ai/platform-features/overview

Explore Cognigy platform features including NLU, CognigyScript, translation, localization, and Packages.

The Cognigy platform includes features to better refine and manage your AI Agents, as well as to help you customize AI Agent conversations with end users.

<CardGroup>
  <Card title="Natural Language Understanding" href="/ai/platform-features/nlu/overview">
    <br />Cognigy supports natural language understanding (NLU) engines to interpret human language and help steer AI Agent interactions.
  </Card>

  <Card title="CognigyScript" href="/ai/platform-features/cognigyscript">
    <br />CognigyScript is a superset of JavaScript that provides access to the AI Agent memory through JSON objects.
  </Card>

  <Card title="Translation and Localization" href="/ai/platform-features/translation-and-localization/localization">
    <br />Cognigy offers robust translation and localization features so that your AI Agents can interact with any end user regardless of the language they speak.
  </Card>

  <Card title="Connections" href="/ai/platform-features/connections">
    <br />Connections are configurations that guarantee your AI Agents securely interact with external services, systems, or APIs.
  </Card>

  <Card title="Packages" href="/ai/platform-features/packages">
    <br />Packages are Cognigy's exporting and importing tool for sharing your AI Agents' resources, such as Flows and connections.
  </Card>
</CardGroup>


# Packages
Source: https://docs.cognigy.com/ai/platform-features/packages



<a href="/release-notes/2026.16"><Badge>Updated in 2026.16</Badge></a>

*Packages* are a useful tool for exporting and importing Project resources, such as Flows or Lexicons, to share with other teams. With Packages, you can transfer resources from one Project to another. When you create a Package, it automatically includes its direct dependencies. Once a Package has been downloaded, you can share it as a ZIP file.

Compared to [Snapshots](/ai/agents/deploy/snapshots), working with Packages provides more flexibility. With Packages, you have control over which resources you import and export.

Packages can include the following Project resources:

* [AI Agents](/ai/agents/develop/manage-ai-agents)
* [Connections](/ai/platform-features/connections)
* [Endpoints](/ai/agents/deploy/endpoints/overview)
* [Flows](/ai/agents/develop/projects-and-flows/overview) (including Nodes and Intents)
* [Goals](/ai/agents/analyze/goals-and-tasks/goals)
* [Handover Providers](/ai/escalate/handovers)
* [Knowledge Stores](/ai/agents/develop/knowledge-ai/knowledge-store)<sup><a href="#footnote1">1</a></sup>
* [Large Language Models](/ai/agents/develop/gen-ai-and-llms/llms)
* [Lexicons](/ai/platform-features/nlu/slots/user-defined/lexicon)
* [NLU Connectors](/ai/platform-features/nlu/external/nlu-connectors/overview)
* [Playbooks](/ai/agents/test/playbooks)

## Limitations

* The maximum Package size depends on how you create or upload it:
  * **Cognigy UI**: You can create and upload Packages of up to 8 GB. For on-premises installations, you can increase this limit by setting the `RESUMABLE_UPLOAD_MAX_FILE_SIZE` environment variable in `values.yaml`. The same limit also applies to [Snapshots](/ai/agents/deploy/snapshots).
  * **Cognigy API**: You can create and upload Packages of up to 2 GB. Packages that exceed this limit can't be created or uploaded. For on-premises installations, you can increase this limit by setting the `PACKAGE_MAX_FILE_SIZE` environment variable in `values.yaml`.
* [Endpoints](/ai/agents/deploy/endpoints/overview) in Packages must have a name with up to 180 characters. Endpoints with longer names cause the Package export to fail.

## Restrictions

* Importing Packages from a newer version of Cognigy into an older version may cause unexpected errors.
* Cognigy skips unsupported resources while importing a Package. For example, deprecated Endpoints or LLMs with outdated configuration schemas.
* If you export a Package from a Project with [Extensions](/ai/for-developers/extensions), you must install the respective Extensions in the target Project as well before importing the Package.

## Working with Packages

<Tabs>
  <Tab title="GUI">
    You can export and import Packages in **Manage > Packaging**. You can also export a resource as a Package from the respective resource page by selecting <img alt="menu" /> **> Export as Package**.
  </Tab>

  <Tab title="API">
    You can view, create, and delete Packages using the [Cognigy API](https://api-trial.cognigy.ai/openapi#tag--Packages) as well as download and upload them.
  </Tab>
</Tabs>

<Accordion title="Package Import">
  * During the import, you may need to [resolve conflicts](#resolve-import-conflicts) if the Package contains a duplicate of a Project resource.
  * Don't import Packages for deploying Knowledge.AI resources in a production Project. AI Agents' knowledge may be temporarily unavailable during the import.
  * Node IDs remain unchanged during the import to ensure that hardcoded links between Flows and Nodes don't break. If you need to import a Flow multiple times into a Project, delete and re-create the Nodes before importing the Package to prevent Node ID conflicts.
</Accordion>

### Map Locales Before Importing Flows

Locales aren't imported with the Flow. However, you can include the localized content in the import process by mapping the Locales.

<Accordion title="Map Locales">
  Before importing the Flow, create the necessary Locales in the target Project and then import the Package. You can map the Locales during the import. The localized content from the Package overwrites the Project content if you replace the Flow.

  For version 4.67 and later, you can map any Locale from the Package to the target Project's primary Locale.

  If you map a non-primary Locale from the Package with incomplete localized content to the Project's primary Locale, Cognigy copies the missing content from the Package's primary Locale.
</Accordion>

### Resolve Import Conflicts

You can resolve conflicts in the **Conflict Management** section that appears during the import.

<Accordion title="FAQ">
  **Q1**: What should I do if I encounter errors during Knowledge Store Import?

  **A1**: Try to upload the Package again. During the reupload, select only the Knowledge Stores that were not fully imported.

  **Q2:** How can I avoid browser request timeouts during Knowledge Store Package upload?

  **A2:** To avoid browser request timeouts, keep the Package size low. Separate your resources into smaller Packages.
</Accordion>

## More Information

* [Localization](/ai/platform-features/translation-and-localization/localization)

***

<span>
  <sup>1</sup>
  When you import a Knowledge Store [Package](/ai/platform-features/packages), the source and target Projects must use the same embedding model. If they use different embedding models, recreate the [Knowledge Sources](/ai/agents/develop/knowledge-ai/knowledge-source/overview) to generate new embeddings with the target Project's model.
</span>


# Tokens
Source: https://docs.cognigy.com/ai/platform-features/tokens



<a href="/release-notes/2025.13"><Badge>Updated in 2025.13</Badge></a>

*Tokens* are visual representations of [CognigyScript](/ai/platform-features/cognigyscript) expressions that you store and reuse in Cognigy. Cognigy has a selection of default Tokens that are associated with Input, Context, and Profile objects.

<Note>
  Tokens automatically include curly brackets in the script. For this reason, you don't need to wrap the CognigyScript in the **Script** field or the Tokens in text fields with curly brackets.
</Note>

## Restrictions

* Only Node text fields with the <img alt="token" /> icon support Tokens.

## Working with Tokens

In **Manage > Tokens**, you can view, create, and delete Tokens as well as copy their Reference ID.

Tokens include the following information:

* **Name** — the Token name displayed in Nodes and the Token Management Menu.
* **Script** — the CognigyScript expression used to access the data. You don't need to wrap the CognigyScript expression in `{{ }}`.
* **Type** — the category of the Token in the search list:
  * **Input** — access to the data in the [Input](/ai/agents/develop/ai-agent-memory/input) object.
  * **Context** — access to the data in the [Context](/ai/agents/develop/ai-agent-memory/context) object.
  * **Profile** — access to the data in the [Profile](/ai/agents/develop/ai-agent-memory/profile) object.
  * **Custom** — specific CognigyScript expression with access to the data in one or more [Cognigy objects](/ai/agents/develop/ai-agent-memory/overview#lifespan-of-cognigy-objects). For example, `"Should your order " + context.customerOrderAmount + " be sent to " + profile.address + "?"`.

You can also create and use Tokens in the Node editor or create Tokens from [JSON editors](#create-and-use-tokens-in-the-node-editor).

<Note>
  Deleting a Token doesn't remove the CognigyScript from the Nodes that used the deleted Token.
</Note>

### Default Tokens

Cognigy provides a selection of default tokens that you can readily use in your Flows.

<Accordion title="Default Tokens">
  | Token                        | Type    | Description                                                                                                                                                                                                                                                                                                                                                          | Target path                                                       |
  | ---------------------------- | ------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------- |
  | AI Agent Description         | Input   | The description of the AI Agent, used to give context about the AI Agent's role.                                                                                                                                                                                                                                                                                     | `context.ai.description`                                          |
  | AI Agent Image URL           | Input   | The URL of the AI Agent's avatar.                                                                                                                                                                                                                                                                                                                                    | `context.ai.imageUrl`                                             |
  | AI Agent Instructions        | Input   | System-level instructions used to guide the AI Agent's behavior during conversations.                                                                                                                                                                                                                                                                                | `context.ai.instructions`                                         |
  | AI Agent Knowledge Store     | Input   | The Knowledge Store assigned to the AI Agent, used for answering questions based on external documents.                                                                                                                                                                                                                                                              | `context.ai.knowledgeStore`                                       |
  | AI Agent Name                | Input   | The display name of the AI Agent shown to users during conversations.                                                                                                                                                                                                                                                                                                | `context.ai.name`                                                 |
  | AI Agent Output              | Input   | The response sent by the AI Agent after processing the prompt and context.                                                                                                                                                                                                                                                                                           | `input.aiAgentOutput`                                             |
  | AI Agent Reference ID        | Input   | The unique ID of the AI Agent used internally to identify and reference the specific AI Agent configuration.                                                                                                                                                                                                                                                         | `context.ai.referenceId`                                          |
  | Answer                       | Input   | Displays the extracted Input. For example, if the user is asked for a number and the answer was `I like the number 7` the extracted answer is `7`.                                                                                                                                                                                                                   | `input.result`                                                    |
  | Available Agents             | Input   | Displays the current available agents in your system.                                                                                                                                                                                                                                                                                                                | `input.availableAgents`                                           |
  | Channel                      | Input   | Displays the current used channel. For example, `adminconsole`.                                                                                                                                                                                                                                                                                                      | `input.channel`                                                   |
  | Current Flow Name            | Input   | Displays the current used Flow name.                                                                                                                                                                                                                                                                                                                                 | `input.flowName`                                                  |
  | Current Time                 | Input   | Displays the current time. For example, `2024-01-08T15:14:59+01:00`                                                                                                                                                                                                                                                                                                  | `input.currentTime.ISODate`                                       |
  | Data                         | Input   | Displays the data payload of the message.                                                                                                                                                                                                                                                                                                                            | `input.data`                                                      |
  | Execution                    | Input   | Represents the number of processed Inputs for this conversation.<br />The first user message in a conversation has an execution value of 1, the second one has an execution value of 2, then 3 and so on.                                                                                                                                                            | `input.execution`                                                 |
  | First Attachment URL         | Input   | Attachments are an array pointing to the URL of the first element.                                                                                                                                                                                                                                                                                                   | `input.attachments[0].url`                                        |
  | First Currency Slot          | Input   | Array pointing to `input.slots`.                                                                                                                                                                                                                                                                                                                                     | `input.slots.MONEY[0].value`                                      |
  | First Date Slot              | Input   | Array pointing to `input.slots`.                                                                                                                                                                                                                                                                                                                                     | `input.slots.DATE[0].start.ISODate`                               |
  | First Email Slot             | Input   | Array pointing to `input.slots`.                                                                                                                                                                                                                                                                                                                                     | `input.slots.EMAIL[0]`                                            |
  | First Number Slot            | Input   | Array pointing to `input.slots`.                                                                                                                                                                                                                                                                                                                                     | `input.slots.NUMBER[0]`                                           |
  | First Percentage Slot        | Input   | Array pointing to `input.slots`.                                                                                                                                                                                                                                                                                                                                     | `input.slots.PERCENTAGE[0]`                                       |
  | Handover Status              | Input   | No data is displayed when handover is not defined. <br /> Indicates status:<ul><li>`completed` - the handover was finished by the agent.</li><li>`cancelled` - the user has cancelled the Handover request.</li><li>`error` - an error occurred when requesting the handover.</li></ul>                                                                              | `input.handover.status`                                           |
  | Input ID                     | Input   | Displays the input ID of the last answer.                                                                                                                                                                                                                                                                                                                            | `input.inputId`                                                   |
  | Intent                       | Input   | Displays a recognized Intent when intents are defined.                                                                                                                                                                                                                                                                                                               | `input.intent`                                                    |
  | Intent Score                 | Input   | Displays the score rate between `0` and `1` of a recognized intent.                                                                                                                                                                                                                                                                                                  | `input.intentScore`                                               |
  | KSearch:TopK Texts           | Input   | Displays the output of the Knowledge Search Node to the user.                                                                                                                                                                                                                                                                                                        | `input.knowledgeSearch?.topK?.map(result=>result?.text).join('')` |
  | Language                     | Input   | Displays the language code for the active NLU language. For example, `en-US`.                                                                                                                                                                                                                                                                                        | `input.language`                                                  |
  | Last Question Result         | Input   | Displays the extracted Input as the answer to an AI Agent question. For example, `I want to order a pizza`.                                                                                                                                                                                                                                                          | `input.result`                                                    |
  | LLM Prompt Result            | Input   | The response sent from the LLM, for example in the Search Extract Output or LLM Prompt Nodes.                                                                                                                                                                                                                                                                        | `input.promptResult`                                              |
  | Microsoft:SSO Request Status | Input   | Used for Teams Single Sign-On (SSO). Indicates the status of the Azure Bot Services connection configured in the [Endpoint settings](/ai/agents/deploy/endpoint-reference/azure-bot-services#specific-endpoint-settings):<ul><li>**pending** - token just requested.</li><li>**declined** - token invalidated.</li><li>**completed** - SSO token received.</li></ul> | `input.data.microsoftSsoPermissionRequest`                        |
  | Microsoft:SSO Token          | Input   | The Microsoft Graph API SSO token for the Azure Bot Services connection configured in the [Endpoint settings](/ai/agents/deploy/endpoint-reference/azure-bot-services#specific-endpoint-settings).                                                                                                                                                                   | `input.data.microsoftSsoToken`                                    |
  | Mode                         | Input   | Displays the type of the message. For example, `Text only`.                                                                                                                                                                                                                                                                                                          | `input.mode`                                                      |
  | Session ID                   | Input   | Displays the unique session ID of your conversation.                                                                                                                                                                                                                                                                                                                 | `input.sessionId`                                                 |
  | State                        | Input   | Displays the state. For example, `default`.                                                                                                                                                                                                                                                                                                                          | `input.state`                                                     |
  | Text                         | Input   | Displays the text you edited in the text field of a Node.                                                                                                                                                                                                                                                                                                            | `input.text`                                                      |
  | Type                         | Input   | Displays the type. For example, `Statement`.                                                                                                                                                                                                                                                                                                                         | `input.type`                                                      |
  | Understood                   | Input   | Indicates `true`/`false`.                                                                                                                                                                                                                                                                                                                                            | `input.understood`                                                |
  | URL Token                    | Input   | The URL token of the Endpoint. The identifier after the last `/` in the Endpoint URL.                                                                                                                                                                                                                                                                                | `input.URLToken`                                                  |
  | User ID                      | Input   | Displays the user ID.                                                                                                                                                                                                                                                                                                                                                | `input.userId`                                                    |
  | VG: Caller Country           | Input   | Displays the caller's country.                                                                                                                                                                                                                                                                                                                                       | `input.data.numberMetaData.country`                               |
  | VG: Caller Country Code      | Input   | Displays the caller's country code.                                                                                                                                                                                                                                                                                                                                  | `input.data.numberMetaData.countryCallingCode`                    |
  | VG: Caller Number            | Input   | Displays the caller's number.                                                                                                                                                                                                                                                                                                                                        | `input.data.numberMetaData.number`                                |
  | VG: Caller Number Type       | Input   | Displays the caller's number type.                                                                                                                                                                                                                                                                                                                                   | `input.data.numberMetaData.type`                                  |
  | VG: Caller URL               | Input   | Displays the caller's URL.                                                                                                                                                                                                                                                                                                                                           | `input.data.numberMetaData.url`                                   |
  | xApp PIN Page URL            | Input   | Displays the PIN Page URL.                                                                                                                                                                                                                                                                                                                                           | `input.apps.baseUrl`                                              |
  | xApp Session PIN             | Input   | Displays the Session PIN.                                                                                                                                                                                                                                                                                                                                            | `input.apps.session.pin`                                          |
  | xApp Session URL             | Input   | Displays the Session URL.                                                                                                                                                                                                                                                                                                                                            | `input.apps.url[]`                                                |
  | Short-Term Memory            | Context | Displays the short-term memory of the conversation, as defined in the [AI Agent Node settings](/ai/agents/develop/node-reference/ai/ai-agent).                                                                                                                                                                                                                       | `context.shortTermMemory`                                         |
  | Accepted GDPR                | Profile | Indicates `true`/`false`; means accepted / not accepted.                                                                                                                                                                                                                                                                                                             | `profile.accepted_gdpr`                                           |
  | Age                          | Profile | Displays the customer's age.                                                                                                                                                                                                                                                                                                                                         | `profile.age`                                                     |
  | Birthday                     | Profile | Displays the customer's birthday.                                                                                                                                                                                                                                                                                                                                    | `profile.birthday`                                                |
  | Customer Number              | Profile | Displays the customer number.                                                                                                                                                                                                                                                                                                                                        | `profile.customerNumber`                                          |
  | Email                        | Profile | Displays the customer's email address.                                                                                                                                                                                                                                                                                                                               | `profile.email`                                                   |
  | First Name                   | Profile | Displays the customer's first name.                                                                                                                                                                                                                                                                                                                                  | `profile.firstname`                                               |
  | Gender                       | Profile | Displays the customer's gender.                                                                                                                                                                                                                                                                                                                                      | `profile.gender`                                                  |
  | Goals                        | Profile | Displays the defined tasks.                                                                                                                                                                                                                                                                                                                                          | `profile.tasks`                                                   |
  | Last Name                    | Profile | Displays the customer's last name.                                                                                                                                                                                                                                                                                                                                   | `profile.lastname`                                                |
  | Location                     | Profile | Displays the customer's location.                                                                                                                                                                                                                                                                                                                                    | `profile.location`                                                |
  | Prevent Data Collection      | Profile | Indicates, if available, `true`/`false`; shows whether data collection is active or not.                                                                                                                                                                                                                                                                             | `profile.prevent_data_collection`                                 |
  | Profile Picture              | Profile | Contains the URL to a profile picture.                                                                                                                                                                                                                                                                                                                               | `profile.profilepic`                                              |
</Accordion>

### Create and Use Tokens in the Node Editor

By clicking <img alt="open token list" /> in a text field, you open the Token list. You can search for a Token in the search field at the top or scroll through the Token types. To create a Token, click <img alt="create token" />.

### Create Tokens from JSON Editors

To create a Token from a JSON editor, right-click a CognigyScript expression and select **Create Token from JSON Path**. The selected CognigyScript expression is added to the **Script** field. If you highlight any code snippet and create a Token from the JSON path, the highlighted code is added to the **Name** field.

## Example

The default Token named **Text** represents `{{input.text}}`. The following image shows the **Text** Token in a Node and the corresponding CognigyScript expression after it:

<Frame>
  <img alt="Text Token in a Node alongside its CognigyScript expression" />
</Frame>


# Automatic Translation
Source: https://docs.cognigy.com/ai/platform-features/translation-and-localization/auto-translation

Localize Cognigy Flows and Intents for multiple languages with automatic or supervised machine translation, enabling multilingual voice and text AI Agents.

Cognigy has integrated *Automatic Translation* capabilities to streamline the process of customizing Flows and Intents for various Locales. These capabilities are helpful for users who need multilingual support for voice and text AI Agents without the need for manual translations.

## Key Features

* **Freedom of Language Choice**. You have the flexibility to choose your preferred translation language.
* **Supervised or Automatic Translation**. You can trigger the translation manually or automatically, depending on your needs.
* **Voice-Ready Translation**. Cognigy's translation capabilities are equipped to handle all modes, making them voice-ready and adaptable to various communication methods.

## Prerequisites

* Add a [machine translation provider](#configure-machine-translation-providers).
* Add a [Locale](#how-to-use-automatic-translation).

## Restrictions

* For [Intents](#machine-translation-for-flow-localization), machine translation doesn't match the [Slots](/ai/platform-features/nlu/intents/overview) you set for your primary Locale to the Keyphrases in the localized Example Sentences. After performing an Intent or Multi-Intent localization, check and fix the [Slot annotation](/ai/platform-features/nlu/intents/annotations) in the localized Example Sentences.

## Compatibility with Machine Translation Providers

Cognigy supports the following machine translation providers:

* Microsoft Translator
  * Microsoft Azure Cognitive Services Translator<sup><a href="#footnote1">1</a></sup>
* Google Cloud Translation
* DeepL Translation Pro

## Configure Machine Translation Providers

You can configure a machine translation provider in **Manage > Settings > Translation Settings**. To do so, you need the following information:

<Tabs>
  <Tab title="Microsoft Translator">
    | Information                  | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
    | ---------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Translation Provider API Key | The API key provided by Microsoft Translator.                                                                                                                                                                                                                                                                                                                                                                                                                     |
    | Custom API Base URL          | This setting is optional. If you have specific requirements or need to connect to a custom instance of the Microsoft Translator API different from the default one `https://api.cognitive.microsofttranslator.com/`, you can use this field to specify a custom base URL. This setting allows you to define the URL, including the protocol scheme, for example, `https://api-eur.cognitive.microsofttranslator.com`, to tailor the API connection to your needs. |
    | Custom Subscription Region   | This setting is optional. It represents the region of your Azure MS Translator resource. You may need to use this setting when calling this API, especially if you have set a specific region in your Azure account.                                                                                                                                                                                                                                              |
  </Tab>

  <Tab title="Google Cloud Translation">
    | Information                  | Description                                       |
    | ---------------------------- | ------------------------------------------------- |
    | Translation Provider API Key | The API key provided by Google Cloud Translation. |
  </Tab>

  <Tab title="DeepL Translate Pro">
    | Information                  | Description                                                                                                                                                                                                                                                                                                                                                |
    | ---------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Translation Provider API Key | The API key provided by DeepL Translate Pro.                                                                                                                                                                                                                                                                                                               |
    | Custom API Base URL          | This setting is optional. If you have specific requirements or need to connect to a custom instance of the DeepL Translate Pro API, you can use this field to specify a custom base URL. This setting allows you to define the URL, including the protocol scheme, for example, `https://api-free.deepl.com/`, to tailor the API connection to your needs. |
  </Tab>
</Tabs>

## How to Use Automatic Translation

<Tabs>
  <Tab title="GUI">
    You can add, edit, and delete Locales in **Manage > Localization**. Also, you can copy their Reference ID.
  </Tab>

  <Tab title="API">
    You can view, create, edit, and delete Locales using the [Cognigy API](https://api-trial.cognigy.ai/openapi#tag--Locales-v2.0).
  </Tab>
</Tabs>

### Machine Translation for Flow Localization

You can automatically translate [Nodes](/ai/agents/develop/nodes/overview) and [Intents](/ai/platform-features/nlu/intents/overview) in the [Flow editor](/ai/agents/develop/projects-and-flows/editor). To do so, select a Locale in the top bar, go to the Node editor or the **NLU** tab, and click **Add Localization**.

<Accordion title="Switching a Locale during a Conversation">
  You can change a Flow's Locale during a conversation using the [Switch Locale](/ai/agents/develop/node-reference/logic/switch-locale) Node.
</Accordion>

## More Information

* [Localization](/ai/platform-features/translation-and-localization/localization)
* [Switch Locale Node](/ai/agents/develop/node-reference/logic/switch-locale)

***

<span>
  <sup>1</sup> To use **Azure Cognitive Services Translator**, select Microsoft Translator in **Manage > Settings > Translation Settings**.
</span>

<br />


# Localization
Source: https://docs.cognigy.com/ai/platform-features/translation-and-localization/localization



Cognigy offers a robust localization feature for easy customization and content reuse. You can configure multiple Locales and add localized content to your Flow. Alternatively, you can use [machine translation to automatically translate content](/ai/platform-features/translation-and-localization/auto-translation). If one Locale doesn't have configured content, Cognigy automatically falls back to another Locale.

## Restrictions

* You can't change the primary Locale after you've selected it.

## Limitations

* The maximum number of Locales per Project is 10.

## How to Use

<Tabs>
  <Tab title="GUI">
    You can add, edit, and delete Locales in **Manage > Localization**. Also, you can copy their Reference ID.
  </Tab>

  <Tab title="API">
    You can view, create, edit, and delete Locales using the [Cognigy API](https://api-trial.cognigy.ai/openapi#tag--Locales-v2.0).
  </Tab>
</Tabs>

### Localize Flows

You can localize [Nodes](/ai/agents/develop/nodes/overview) and [Intents](/ai/platform-features/nlu/intents/overview) in the [Flow editor](/ai/agents/develop/projects-and-flows/editor) after selecting a Locale in the top bar.

<Accordion title="Switching a Locale during a Conversation">
  You can change a Flow's Locale during a conversation using the [Switch Locale](/ai/agents/develop/node-reference/logic/switch-locale) Node.
</Accordion>

## More Information

* [Automatic Translation](/ai/platform-features/translation-and-localization/auto-translation)
* [Switch Locale Node](/ai/agents/develop/node-reference/logic/switch-locale)


# Real-Time Translation
Source: https://docs.cognigy.com/ai/platform-features/translation-and-localization/real-time-translation



The *Real-Time Translation* feature translates conversations in real time, eliminating the need for pre-translation, as in typical [localization](/ai/platform-features/translation-and-localization/localization) scenarios.

With real-time translation, you don't need human agents for every language nor to localize your Flow beforehand. You can focus on a few key languages while still serving customers in many others.

## Key Features

* [Multilingual AI](#users-and-ai-agents). Your voice and chat [AI Agents](/ai/agents/quickstart) can switch languages in real time, even if you only have configured your Flow in English. With this approach, you can reach a greater audience in their native language with low effort.
* [Real-time translation between users and human agents](#users-and-human-agents). Users and human agents can write in their languages after a [handover](/ai/escalate/handovers), but they receive messages translated into their own language.

## Prerequisites

* You have configured a [machine translation provider](#configure-machine-translation-providers).

## Compatibility with Machine Translation Providers

Cognigy supports the following machine translation providers:

* Microsoft Translator
  * Microsoft Azure Cognitive Services Translator<sup><a href="#footnote1">1</a></sup>
* Google Cloud Translation
* DeepL Translation Pro

## Configure Machine Translation Providers

You can configure a machine translation provider in **Manage > Settings > Translation Settings**. To do so, you need the following information:

<Tabs>
  <Tab title="Microsoft Translator">
    | Information                  | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                       |
    | ---------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Translation Provider API Key | The API key provided by Microsoft Translator.                                                                                                                                                                                                                                                                                                                                                                                                                     |
    | Custom API Base URL          | This setting is optional. If you have specific requirements or need to connect to a custom instance of the Microsoft Translator API different from the default one `https://api.cognitive.microsofttranslator.com/`, you can use this field to specify a custom base URL. This setting allows you to define the URL, including the protocol scheme, for example, `https://api-eur.cognitive.microsofttranslator.com`, to tailor the API connection to your needs. |
    | Custom Subscription Region   | This setting is optional. It represents the region of your Azure MS Translator resource. You may need to use this setting when calling this API, especially if you have set a specific region in your Azure account.                                                                                                                                                                                                                                              |
  </Tab>

  <Tab title="Google Cloud Translation">
    | Information                  | Description                                       |
    | ---------------------------- | ------------------------------------------------- |
    | Translation Provider API Key | The API key provided by Google Cloud Translation. |
  </Tab>

  <Tab title="DeepL Translate Pro">
    | Information                  | Description                                                                                                                                                                                                                                                                                                                                                |
    | ---------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Translation Provider API Key | The API key provided by DeepL Translate Pro.                                                                                                                                                                                                                                                                                                               |
    | Custom API Base URL          | This setting is optional. If you have specific requirements or need to connect to a custom instance of the DeepL Translate Pro API, you can use this field to specify a custom base URL. This setting allows you to define the URL, including the protocol scheme, for example, `https://api-free.deepl.com/`, to tailor the API connection to your needs. |
  </Tab>
</Tabs>

## How to Use Real-Time Translation

You can configure the real-time translation feature in the [Real Time Translation](/ai/agents/deploy/endpoints/real-time-translation-settings) settings in the Endpoint editor. All Endpoints, including voice Endpoints, such as Voice Gateway, support real-time translation.

### Users and AI Agents

Real-time translation between a user and an AI Agent works for text-based and voice conversations.

### Users and Human Agents

Real-time translation between a user and a human agent works for text-based conversations after a user triggers a [Handover to Human Agent Node](/ai/agents/develop/node-reference/service/handover-to-agent) in the [Flow](/ai/agents/develop/projects-and-flows/overview). Conversations with real-time translation between the user and the human agent are recorded for compliance purposes.

After the human agent takes over the conversation, the user can keep writing in their original language while the human agent receives real-time translation of the user input. The human agent can respond in their language, and the user receives real-time translation of the human agent's response. The following image shows real-time translation example, where the user is writing in German and the human agent is responding in English.

<Frame>
  <img alt="Real-time translation between a German-speaking user and English-speaking agent" />
</Frame>

## Supported Languages by Machine Translation Provider

The following tables show which languages are supported by the supported machine translation providers and the respective language code to use when configuring Nodes:

<AccordionGroup>
  <Accordion title="Microsoft">
    | Language                | Language Code |
    | ----------------------- | ------------- |
    | Afrikaans               | `af`          |
    | Albanian                | `sq`          |
    | Amharic                 | `am`          |
    | Arabic                  | `ar`          |
    | Armenian                | `hy`          |
    | Assamese                | `as`          |
    | Azerbaijani             | `az`          |
    | Bangla                  | `bn`          |
    | Bosnian (Latin)         | `bs`          |
    | Bulgarian               | `bg`          |
    | Cantonese (Traditional) | `yue`         |
    | Catalan                 | `ca`          |
    | Chinese Simplified      | `zh-Hans`     |
    | Chinese Traditional     | `zh-Hant`     |
    | Croatian                | `hr`          |
    | Czech                   | `cs`          |
    | Danish                  | `da`          |
    | Dari                    | `prs`         |
    | Dutch                   | `nl`          |
    | English                 | `en`          |
    | Estonian                | `et`          |
    | Fijian                  | `fj`          |
    | Filipino                | `fil`         |
    | Finnish                 | `fi`          |
    | French                  | `fr`          |
    | French (Canada)         | `fr-ca`       |
    | German                  | `de`          |
    | Greek                   | `el`          |
    | Gujarati                | `gu`          |
    | Haitian Creole          | `ht`          |
    | Hebrew                  | `he`          |
    | Hindi                   | `hi`          |
    | Hmong Daw               | `mww`         |
    | Hungarian               | `hu`          |
    | Icelandic               | `is`          |
    | Indonesian              | `id`          |
    | Inuktitut               | `iu`          |
    | Irish                   | `ga`          |
    | Italian                 | `it`          |
    | Japanese                | `ja`          |
    | Kannada                 | `kn`          |
    | Kazakh                  | `kk`          |
    | Khmer                   | `km`          |
    | Klingon                 | `tlh-Latn`    |
    | Klingon (plqaD)         | `tlh-Piqd`    |
    | Korean                  | `ko`          |
    | Kurdish (Central)       | `ku`          |
    | Kurdish (Northern)      | `kmr`         |
    | Lao                     | `lo`          |
    | Latvian                 | `lv`          |
    | Lithuanian              | `lt`          |
    | Macedonian              | `mk`          |
    | Malagasy                | `mg`          |
    | Malay                   | `ms`          |
    | Malayalam               | `ml`          |
    | Maltese                 | `mt`          |
    | Maori                   | `mi`          |
    | Marathi                 | `mr`          |
    | Myanmar                 | `my`          |
    | Nepali                  | `ne`          |
    | Norwegian               | `nb`          |
    | Odia                    | `or`          |
    | Pashto                  | `ps`          |
    | Persian                 | `fa`          |
    | Polish                  | `pl`          |
    | Portuguese (Brazil)     | `pt`          |
    | Portuguese (Portugal)   | `pt-pt`       |
    | Punjabi                 | `pa`          |
    | Queretaro Otomi         | `otq`         |
    | Romanian                | `ro`          |
    | Russian                 | `ru`          |
    | Samoan                  | `sm`          |
    | Serbian (Cyrillic)      | `sr-Cyrl`     |
    | Serbian (Latin)         | `sr-Latn`     |
    | Slovak                  | `sk`          |
    | Slovenian               | `sl`          |
    | Spanish                 | `es`          |
    | Swahili                 | `sw`          |
    | Swedish                 | `sv`          |
    | Tahitian                | `ty`          |
    | Tamil                   | `ta`          |
    | Telugu                  | `te`          |
    | Thai                    | `th`          |
    | Tigrinya                | `ti`          |
    | Tongan                  | `to`          |
    | Turkish                 | `tr`          |
    | Ukrainian               | `uk`          |
    | Urdu                    | `ur`          |
    | Vietnamese              | `vi`          |
    | Welsh                   | `cy`          |
    | Yucatec Maya            | `yua`         |
  </Accordion>

  <Accordion title="Google">
    | Language                      | Language Code |
    | ----------------------------- | ------------- |
    | Afrikaans                     | `af`          |
    | Albanian                      | `sq`          |
    | Amharic                       | `am`          |
    | Arabic                        | `ar`          |
    | Armenian                      | `hy`          |
    | Azerbaijani                   | `az`          |
    | Basque                        | `eu`          |
    | Belarusian                    | `be`          |
    | Bengali                       | `bn`          |
    | Bosnian                       | `bs`          |
    | Bulgarian                     | `bg`          |
    | Catalan                       | `ca`          |
    | Chinese (Simplified)          | `zh-CN`       |
    | Chinese (Traditional)         | `zh`          |
    | Chinese (Traditional)         | `zh-TW`       |
    | Corsican                      | `co`          |
    | Croatian                      | `hr`          |
    | Czech                         | `cs`          |
    | Danish                        | `da`          |
    | Dutch                         | `nl`          |
    | English                       | `en`          |
    | Esperanto                     | `eo`          |
    | Estonian                      | `et`          |
    | Finnish                       | `fi`          |
    | French                        | `fr`          |
    | Frisian                       | `fy`          |
    | Galician                      | `gl`          |
    | Georgian                      | `ka`          |
    | German                        | `de`          |
    | Greek                         | `el`          |
    | Gujarati                      | `gu`          |
    | Haitian Creole                | `ht`          |
    | Hausa                         | `ha`          |
    | Hawaiian                      | `haw`         |
    | Hebrew                        | `he`          |
    | Hebrew                        | `iw`          |
    | Hindi                         | `hi`          |
    | Hmong                         | `hmn`         |
    | Hungarian                     | `hu`          |
    | Icelandic                     | `is`          |
    | Igbo                          | `ig`          |
    | Indonesian                    | `id`          |
    | Irish                         | `ga`          |
    | Italian                       | `it`          |
    | Japanese                      | `ja`          |
    | Javanese                      | `jv`          |
    | Kannada                       | `kn`          |
    | Kazakh                        | `kk`          |
    | Khmer                         | `km`          |
    | Kinyarwanda                   | `rw`          |
    | Korean                        | `ko`          |
    | Kurdish                       | `ku`          |
    | Kyrgyz                        | `ky`          |
    | Lao                           | `lo`          |
    | Latin                         | `la`          |
    | Latvian                       | `lv`          |
    | Lithuanian                    | `lt`          |
    | Luxembourgish                 | `lb`          |
    | Macedonian                    | `mk`          |
    | Malagasy                      | `mg`          |
    | Malay                         | `ms`          |
    | Malayalam                     | `ml`          |
    | Maltese                       | `mt`          |
    | Maori                         | `mi`          |
    | Marathi                       | `mr`          |
    | Mongolian                     | `mn`          |
    | Myanmar (Burmese)             | `my`          |
    | Nepali                        | `ne`          |
    | Norwegian                     | `no`          |
    | Nyanja (Chichewa)             | `ny`          |
    | Odia (Oriya)                  | `or`          |
    | Pashto                        | `ps`          |
    | Persian                       | `fa`          |
    | Polish                        | `pl`          |
    | Portuguese (Portugal, Brazil) | `pt`          |
    | Punjabi                       | `pa`          |
    | Romanian                      | `ro`          |
    | Russian                       | `ru`          |
    | Samoan                        | `sm`          |
    | Scots Gaelic                  | `gd`          |
    | Serbian                       | `sr`          |
    | Sesotho                       | `st`          |
    | Shona                         | `sn`          |
    | Sindhi                        | `sd`          |
    | Sinhala (Sinhalese)           | `si`          |
    | Slovak                        | `sk`          |
    | Slovenian                     | `sl`          |
    | Somali                        | `so`          |
    | Spanish                       | `es`          |
    | Sundanese                     | `su`          |
    | Swahili                       | `sw`          |
    | Swedish                       | `sv`          |
    | Tagalog (Filipino)            | `tl`          |
    | Tajik                         | `tg`          |
    | Tamil                         | `ta`          |
    | Tatar                         | `tt`          |
    | Telugu                        | `te`          |
    | Thai                          | `th`          |
    | Turkish                       | `tr`          |
    | Turkmen                       | `tk`          |
    | Ukrainian                     | `uk`          |
    | Urdu                          | `ur`          |
    | Uyghur                        | `ug`          |
    | Uzbek                         | `uz`          |
    | Vietnamese                    | `vi`          |
    | Welsh                         | `cy`          |
    | Xhosa                         | `xh`          |
    | Yiddish                       | `yi`          |
    | Yoruba                        | `yo`          |
    | Zulu                          | `zu`          |
  </Accordion>

  <Accordion title="DeepL Translate Pro">
    | Language               | Language Code |
    | ---------------------- | ------------- |
    | Arabic                 | `AR`          |
    | Bulgarian              | `BG`          |
    | Czech                  | `CS`          |
    | Danish                 | `DA`          |
    | German                 | `DE`          |
    | Greek                  | `EL`          |
    | English (British)      | `EN-GB`       |
    | English (American)     | `EN-US`       |
    | Spanish                | `ES`          |
    | Estonian               | `ET`          |
    | Finnish                | `FI`          |
    | French                 | `FR`          |
    | Hungarian              | `HU`          |
    | Indonesian             | `ID`          |
    | Italian                | `IT`          |
    | Japanese               | `JA`          |
    | Korean                 | `KO`          |
    | Lithuanian             | `LT`          |
    | Latvian                | `LV`          |
    | Norwegian              | `NB`          |
    | Dutch                  | `NL`          |
    | Polish                 | `PL`          |
    | Portuguese (Brazilian) | `PT-BR`       |
    | Portuguese (European)  | `PT-PT`       |
    | Romanian               | `RO`          |
    | Russian                | `RU`          |
    | Slovak                 | `SK`          |
    | Slovenian              | `SL`          |
    | Swedish                | `SV`          |
    | Turkish                | `TR`          |
    | Ukrainian              | `UK`          |
    | Chinese                | `ZH`          |
  </Accordion>
</AccordionGroup>

## More Information

* [Real-Time Translation Settings](/ai/agents/deploy/endpoints/real-time-translation-settings)
* [Automatic Translation](/ai/platform-features/translation-and-localization/auto-translation)
* [Localization](/ai/platform-features/translation-and-localization/localization)

***

<span>
  <sup>1</sup> To use **Azure Cognitive Services Translator**, select Microsoft Translator in **Manage > Settings > Translation Settings**.
</span>

<br />


# Download a packaged Logs file
Source: https://docs.cognigy.com/api-reference-logs/logs-v21/download-a-packaged-logs-file

https://api-trial.cognigy.ai/logs/openapi/openapi-viewer.json get /v2.1/projects/{projectId}/logs/package/{logsFileId}/download
Downloads the logs file previously created via `/v2.1/projects/{projectId}/logs/package`. The file contains one log entry per line (JSON Lines). While the file is still being packaged, a 409 error is returned — retry later. Packaged files expire after a configured retention time; after expiry a 404 is returned and the logs have to be packaged again.



# Get a Log Entry
Source: https://docs.cognigy.com/api-reference-logs/logs-v21/get-a-log-entry

https://api-trial.cognigy.ai/logs/openapi/openapi-viewer.json get /v2.1/projects/{projectId}/logs/{logEntryId}
Gets a [log entry](https://docs.cognigy.com/ai/agents/test/logs) in a Project. When the request header `x-cognigy-log-model-version` is set to `v2`, the response uses the V2 shape with a nested `metadata` object containing `organizationId`, `projectId`, and `type` instead of a top-level `type` field.



# Get Log Entries
Source: https://docs.cognigy.com/api-reference-logs/logs-v21/get-log-entries

https://api-trial.cognigy.ai/logs/openapi/openapi-viewer.json get /v2.1/projects/{projectId}/logs
Gets [log entries](https://docs.cognigy.com/ai/agents/test/logs) in a Project. By default, 25 log entries are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results. If you provide an invalid `cursor` value, a 400 error will be returned. When `startDate` or `endDate` are not provided, a default 1-hour time window is applied. When the request header `x-cognigy-log-model-version` is set to `v2`, `total` is only an accurate count on the first page (no `next`/`previous` cursor); subsequent cursor-paginated pages return `-1` for `total`, so callers must not rely on it there.



# Get the Latest Log Entries
Source: https://docs.cognigy.com/api-reference-logs/logs-v21/get-the-latest-log-entries

https://api-trial.cognigy.ai/logs/openapi/openapi-viewer.json get /v2.1/projects/{projectId}/logs/tail
Gets the latest 100 [log entries](https://docs.cognigy.com/ai/agents/test/logs) in a Project. Results are sorted descending by timestamp and `_id`. Supports cursor-based pagination via opaque base64url tokens returned in `nextCursor` and `previousCursor`. The `total` field is always `-1` because counting rows in this collection is intentionally skipped for performance reasons. If you provide an invalid `cursor` value, a 400 error will be returned. When the request header `x-cognigy-log-model-version` is set to `v2`, the response items use the V2 shape with a nested `metadata` object instead of a top-level `type` field.



# Package Log Entries into a downloadable file
Source: https://docs.cognigy.com/api-reference-logs/logs-v21/package-log-entries-into-a-downloadable-file

https://api-trial.cognigy.ai/logs/openapi/openapi-viewer.json post /v2.1/projects/{projectId}/logs/package
Starts packaging all [log entries](https://docs.cognigy.com/ai/agents/test/logs) matching the given filters into a file on disk and returns immediately with the `logsFileId` of that file. Download the file via `/v2.1/projects/{projectId}/logs/package/{logsFileId}/download` once packaging has finished. Time ranges larger than one hour are queried in one-hour windows to keep the database load low. If an identical packaging request (same filters) is already in progress, the `logsFileId` of that packaging is returned instead of starting a new one. Packaged files expire after a configured retention time.



# Get a Log Entry
Source: https://docs.cognigy.com/api-reference-logs/logs/get-a-log-entry

https://api-trial.cognigy.ai/logs/openapi/openapi-viewer.json get /v2.0/projects/{projectId}/logs/{logEntryId}
Gets a [log entry](https://docs.cognigy.com/ai/agents/test/logs) in a Project. This endpoint is deprecated, use `/v2.1/projects/{projectId}/logs/{logEntryId}` instead.



# Get Log Entries
Source: https://docs.cognigy.com/api-reference-logs/logs/get-log-entries

https://api-trial.cognigy.ai/logs/openapi/openapi-viewer.json get /v2.0/projects/{projectId}/logs
Gets [log entries](https://docs.cognigy.com/ai/agents/test/logs) in a Project. By default, 25 log entries are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results. If you provide an invalid `cursor` value, a 400 error will be returned. This endpoint is deprecated, use `/v2.1/projects/{projectId}/logs` instead.



# Get the Latest Log Entries
Source: https://docs.cognigy.com/api-reference-logs/logs/get-the-latest-log-entries

https://api-trial.cognigy.ai/logs/openapi/openapi-viewer.json get /v2.0/projects/{projectId}/logs/tail
Gets the latest 100 [log entries](https://docs.cognigy.com/ai/agents/test/logs) in a Project. Uses aggregation to first get the latest rows by `_id`, then sorts by timestamp. Supports `cursor`-based pagination using the `_id` field. If you provide an invalid `cursor` value, a 400 error will be returned. This endpoint is deprecated, use `/v2.1/projects/{projectId}/logs/tail` instead.



# Get dashboard overview metrics
Source: https://docs.cognigy.com/api-reference-simulator/dashboard/get-dashboard-overview-metrics

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json get /dashboard/overview
Retrieve key metrics for the dashboard overview cards including total scenarios, simulations, average success rate, and scheduled runs count.



# Get success rate trend data
Source: https://docs.cognigy.com/api-reference-simulator/dashboard/get-success-rate-trend-data

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json get /dashboard/success-rate-trend
Retrieve 7-day rolling average trend of simulation outcomes for visualization. Returns daily aggregated data showing successful vs failed runs.



# Get upcoming scheduled runs
Source: https://docs.cognigy.com/api-reference-simulator/dashboard/get-upcoming-scheduled-runs

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json get /dashboard/upcoming-scheduled-runs
Retrieve list of upcoming scheduled simulation runs with scenario name, next run time, and frequency information.



# Bulk generate separated personas and missions
Source: https://docs.cognigy.com/api-reference-simulator/personas/bulk-generate-separated-personas-and-missions

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json post /personas/generate-bulk
Generate separated persona and mission information for all provided persona types and mission types.
Returns personas and missions in separated structures for independent management.

**Key Feature**: This endpoint generates ALL combinations of provided persona types and mission types, 
returning them in a structured format where:
- Personas are keyed by persona type name
- Missions are keyed by mission type name




# Create scenario from transcript
Source: https://docs.cognigy.com/api-reference-simulator/personas/create-scenario-from-transcript

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json post /personas/from-transcript
Create a scenario draft by analyzing a conversation transcript. The system uses LLM to extract user intent, 
conversation goals, and user characteristics from the transcript, then generates an appropriate persona package.

The endpoint supports:
- Fetching transcript via sessionId from analytics service (RPC call)
- Direct transcript content via transcriptContent
- Optional flow context for enhanced persona generation
- Content safety validation (warnings only, non-blocking)

Priority: If both sessionId and transcriptContent are provided, sessionId takes precedence.

Returns a complete persona package and metadata about the analysis.




# Generate persona packages
Source: https://docs.cognigy.com/api-reference-simulator/personas/generate-persona-packages

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json post /personas/generate
Generate persona packages based on selected mission type and persona type with auto-discovery of AI agents and tools from the flow.
This endpoint creates complete persona packages including scenario names, persona descriptions, missions, and success criteria.




# Get persona options (mission types and persona types)
Source: https://docs.cognigy.com/api-reference-simulator/personas/get-persona-options-mission-types-and-persona-types

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json post /personas/options
Smart endpoint that handles multiple scenarios for getting persona options:
- **Legacy**: Provide only `flowId` to analyze the entire flow
- **AI Agent Specific**: Provide `flowId` + `aiagentReferenceId` to analyze a specific AI agent
- **Job Node Specific**: Provide `flowId` + `aiagentReferenceId` + `jobNodeId` to analyze a specific job node




# Regenerate a specific field of a persona package
Source: https://docs.cognigy.com/api-reference-simulator/personas/regenerate-a-specific-field-of-a-persona-package

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json post /personas/regenerate-field
Regenerate a specific field of a persona package using LLM. Requires the full persona package context to regenerate any individual field.
Valid fields to regenerate: simulationName, personaName, personaDescription, mission, successCriteria




# Get simulation run by ID
Source: https://docs.cognigy.com/api-reference-simulator/simulation-runs/get-simulation-run-by-id

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json get /simulations/{simulationReference}/batches/{simulationRunBatchReference}/runs/{simulationRunReference}
Retrieve a specific simulation run by its reference ID.



# Get simulation runs by simulation
Source: https://docs.cognigy.com/api-reference-simulator/simulation-runs/get-simulation-runs-by-simulation

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json get /simulations/{simulationReference}/batches/{simulationRunBatchReference}/runs
Retrieve all simulation runs for a specific simulation.



# Clone scenario
Source: https://docs.cognigy.com/api-reference-simulator/simulations/clone-scenario

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json post /simulations/{simulationReference}/clone
Clone a scenario and optionally provide a new name for the cloned scenario.



# Create a scheduler for a simulation
Source: https://docs.cognigy.com/api-reference-simulator/simulations/create-a-scheduler-for-a-simulation

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json post /simulations/{simulationReference}/schedules
Create a scheduler associated with a scenario to run simulations automatically.



# Create scenario
Source: https://docs.cognigy.com/api-reference-simulator/simulations/create-scenario

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json post /simulations
Create a scenario with the provided configuration.



# Delete scenario
Source: https://docs.cognigy.com/api-reference-simulator/simulations/delete-scenario

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json delete /simulations/{simulationReference}
Delete a scenario and all its associated data.



# Get all scenarios
Source: https://docs.cognigy.com/api-reference-simulator/simulations/get-all-scenarios

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json get /simulations
Retrieve a paginated list of scenarios for the authenticated user's organization and project.



# Get all simulations
Source: https://docs.cognigy.com/api-reference-simulator/simulations/get-all-simulations

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json get /simulations/batches
Retrieve a paginated list of all simulations.



# Get scenario by ID
Source: https://docs.cognigy.com/api-reference-simulator/simulations/get-scenario-by-id

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json get /simulations/{simulationReference}
Retrieve a specific scenario by its reference ID.



# Get scheduler configuration for a simulation
Source: https://docs.cognigy.com/api-reference-simulator/simulations/get-scheduler-configuration-for-a-simulation

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json get /simulations/{simulationReference}/schedules
Retrieve the scheduler configuration associated with a specific simulation.



# Get simulation by ID
Source: https://docs.cognigy.com/api-reference-simulator/simulations/get-simulation-by-id

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json get /simulations/{simulationReference}/batches/{simulationRunBatchReference}
Retrieve a specific simulation by its reference ID.



# Get simulations by scenarios
Source: https://docs.cognigy.com/api-reference-simulator/simulations/get-simulations-by-scenarios

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json get /simulations/{simulationReference}/batches
Retrieve all simulations for a specific scenario.



# Run scenario
Source: https://docs.cognigy.com/api-reference-simulator/simulations/run-scenario

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json post /simulations/{simulationReference}/schedule
Run multiple executions of a scenario.



# Stop simulation
Source: https://docs.cognigy.com/api-reference-simulator/simulations/stop-simulation

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json post /simulations/{simulationReference}/batches/{simulationRunBatchReference}
Stop a running simulation and all its active runs.



# Update a scheduler by reference
Source: https://docs.cognigy.com/api-reference-simulator/simulations/update-a-scheduler-by-reference

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json patch /simulations/{simulationReference}/schedules/{schedulerReference}
Update fields of an existing scheduler configuration by its reference.



# Update scenario
Source: https://docs.cognigy.com/api-reference-simulator/simulations/update-scenario

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json patch /simulations/{simulationReference}
Update an existing scenario with new configuration.



# Create transcript upload request
Source: https://docs.cognigy.com/api-reference-simulator/transcript-scenario-generation/create-transcript-upload-request

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json post /v1/transcript-request
Create a new transcript upload session. This is the first step in the three-step workflow
for generating scenarios from transcript files.

This endpoint:
- Creates a new request with a unique ID
- Checks rate limits per project (max active requests)
- Returns configuration details (max files, max file size, supported formats)

The returned requestId must be passed in subsequent upload and generation calls.

**projectId is required.** Provide it via the `project_id` header, or in the request body/query.




# Generate scenarios from uploaded transcripts
Source: https://docs.cognigy.com/api-reference-simulator/transcript-scenario-generation/generate-scenarios-from-uploaded-transcripts

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json post /v1/generate-from-transcript
Generate 3 distinct simulation scenarios from previously uploaded and parsed transcript files.
This is the third and final step in the three-step workflow.

The endpoint:
1. Retrieves all uploaded files from Redis for the given requestId
2. Combines all transcript content
3. Validates minimum word count
4. Calls the LLM via RPC to generate 3 scenarios
5. Cleans up Redis data after successful generation

Each generated scenario includes a simulationName, personaName, personaDescription,
mission, and success criteria - ready to be used for creating simulations.




# Upload and parse a transcript file
Source: https://docs.cognigy.com/api-reference-simulator/transcript-scenario-generation/upload-and-parse-a-transcript-file

https://api-trial.cognigy.ai/testing/openapi/openapi-viewer.json post /v1/transcript-upload
Upload a single transcript file for parsing and content extraction. This is the second step
in the three-step workflow.

The file is streamed directly to Redis via the streamUpload middleware, then:
1. Read back from Redis
2. Parsed based on file type (PDF, CSV, DOCX)
3. Analyzed for speaker labels (Agent/Customer)
4. Stored as structured data in Redis (keyed by requestId)

Multiple files can be uploaded by calling this endpoint multiple times with the same requestId.
The generate endpoint retrieves all files for the request by requestId.




# About Cognigy API
Source: https://docs.cognigy.com/api-reference/about

Introduction to the Cognigy REST API, covering authentication, CORS support, and RESTful endpoints for managing AI Agents and platform resources.

The Cognigy API follows a [RESTful](https://en.wikipedia.org/wiki/REST) design and uses standard HTTP methods, status codes, and authentication mechanisms.
All endpoints support JSON-encoded responses and predictable resource-oriented URLs.

### Cross-Origin Resource Sharing (CORS)

The API fully supports CORS in accordance with the [W3C specification](https://fetch.spec.whatwg.org/). All responses include a wildcard `Access-Control-Allow-Origin` header, enabling cross-domain browser requests.

### Authentication

Cognigy supports the following authentication methods:

<AccordionGroup>
  <Accordion title="API Key">
    For Cognigy installations, use an [API key](/ai/administer/user-menu/my-profile#api-keys) in the request path or HTTP headers. Keep it secure and revoke it if exposed.
  </Accordion>

  <Accordion title="BasicAuth">
    Use a username and password for API requests related to the [Management UI](/api-reference/management-ui/create-an-api-key).
  </Accordion>
</AccordionGroup>

### Error Handling

The API uses standard HTTP status codes (400 and higher) to indicate errors. Error responses follow [RFC 7807](https://datatracker.ietf.org/doc/html/rfc7807) and include a traceId that you can provide to [Cognigy technical support](/help/get-help) for troubleshooting.

### API Server

Cognigy API:

* **Core Resources:** `https://api-trial.cognigy.ai/new/`
* **Simulator:** `https://api-trial.cognigy.ai/testing/`
* **Logs:** `https://api-trial.cognigy.ai/logs/`

For constructing the API base URL for your environment, refer to [API Base URL](/ai/for-developers/developers/api-and-cli#api-base-url).

### API Scope

The Cognigy API is split into independently versioned services:

| API            | Base Path   | Purpose                                                                                                                   |
| -------------- | ----------- | ------------------------------------------------------------------------------------------------------------------------- |
| Core Resources | `/new/`     | Manages the lifecycle of AI Agents, including Flows, Endpoints, Snapshots, Knowledge Stores, and platform administration. |
| Simulator      | `/testing/` | Manages automated testing resources such as scenarios, simulation runs, schedules, and evaluation results.                |
| Logs           | `/logs/`    | Retrieves Flow execution logs and debug entries.                                                                          |

All services share the same authentication methods and error-handling conventions described above.


# Get Knowledge Chunk Count
Source: https://docs.cognigy.com/api-reference/administration/get-knowledge-chunk-count

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/organisations/chunkcount
Gets the total number of Knowledge Chunks created across Projects in the organization.



# Get Organization Policies
Source: https://docs.cognigy.com/api-reference/administration/get-organization-policies

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/organisations/policy
Gets all policies for the organization as shown in the Management UI, including quotas, password and session rules, conversation and Contact Profile settings, data privacy options, billing time zone, and AI Ops Center status.



# Get System Maintenance Messages
Source: https://docs.cognigy.com/api-reference/administration/get-system-maintenance-messages

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/system/message
Gets system-wide maintenance messages that are preconfigured in the Management UI, including their display timing and visibility settings.



# Get the License State for the Organization
Source: https://docs.cognigy.com/api-reference/administration/get-the-license-state-for-the-organization

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/system/license
Gets the current license state of the organization, including its system capabilities and usage quotas.



# Initiate Organization Deletion
Source: https://docs.cognigy.com/api-reference/administration/initiate-organization-deletion

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/organisations/requestdeletion
Initiates the process to delete an organisation. A confirmation token will be sent via email to the system administrator to proceed with the deletion.



# Set the License
Source: https://docs.cognigy.com/api-reference/administration/set-the-license

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/system/license
Sets the system license for the organization, updating license details such as access and quotas.



# Create an Agent Copilot Configuration
Source: https://docs.cognigy.com/api-reference/agent-copilot-configuration/create-an-agent-copilot-configuration

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/aicopilotconfigs
Creates an Agent Copilot configuration in a Project. The Agent Copilot configuration defines the grid layout, which serves as the structure for your AI Copilot Workspace. After creation, select the configuration in the Cognigy.AI Endpoint under the Copilot section.



# Delete an Agent Copilot Configuration
Source: https://docs.cognigy.com/api-reference/agent-copilot-configuration/delete-an-agent-copilot-configuration

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/aicopilotconfigs/{configId}
Deletes an Agent Copilot configuration from a Project. If the configuration is still in use in the Copilot section of any Endpoint in this Project, replace the configuration with another one before using this method.



# Get Agent Copilot Configurations
Source: https://docs.cognigy.com/api-reference/agent-copilot-configurations/get-agent-copilot-configurations

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/aicopilotconfigs
Gets Agent Copilot configurations for a Project. By default, 25 audit events are returned. Use the `limit` query parameter to change this value.



# Get an Agent Copilot Configuration
Source: https://docs.cognigy.com/api-reference/agent-copilot-configurations/get-an-agent-copilot-configuration

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/aicopilotconfigs/{configId}
Gets an Agent Copilot Configuration from a Project.



# Update an Agent Copilot Configuration
Source: https://docs.cognigy.com/api-reference/agent-copilot-configurations/update-an-agent-copilot-configuration

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/aicopilotconfigs/{configId}
Updates an existing Agent Copilot configuration in a Project. Allows modifying the grid layout, tiles, or metadata of a configuration. You must provide the relevant project ID and the configuration ID to apply changes.



# Check AI Agent Name Availability
Source: https://docs.cognigy.com/api-reference/ai-agents/check-ai-agent-name-availability

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/aiagents/validatename
Checks whether an AI Agent name is already used in a Project.



# Create an AI Agent
Source: https://docs.cognigy.com/api-reference/ai-agents/create-an-ai-agent

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/aiagents
Creates an AI Agent in a Project. The request defines the AI Agent persona (name, description, avatar), behavioral instructions, speaking style, and safety settings. You can optionally link the AI Agent to a Knowledge Store, configure voice and text-to-speech settings, and control whether voice configurations are activated. The method returns the newly created AI Agent, including its unique ID and metadata such as creation timestamps and author information.



# Delete an AI Agent
Source: https://docs.cognigy.com/api-reference/ai-agents/delete-an-ai-agent

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/aiagents/{aiAgentId}
Deletes an AI Agent from a Project.



# Get AI Agent Jobs and their Tools
Source: https://docs.cognigy.com/api-reference/ai-agents/get-ai-agent-jobs-and-their-tools

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/aiagents/{aiAgentId}/jobs
Gets all jobs and associated tools for an AI Agent by its ID.



# Get AI Agents
Source: https://docs.cognigy.com/api-reference/ai-agents/get-ai-agents

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/aiagents
Gets AI Agents from a Project. By default, returns 25 AI Agents from all Projects the user has access to. Use the `limit` query parameter to change the default value. Specify the Project ID to get AI Agents from a specific Project.



# Get AI Agents from the Job Market
Source: https://docs.cognigy.com/api-reference/ai-agents/get-ai-agents-from-the-job-market

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/aiagents/hire/templates
Gets a list of AI Agents from the job market that are ready for hire.



# Get an AI Agent
Source: https://docs.cognigy.com/api-reference/ai-agents/get-an-ai-agent

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/aiagents/{aiAgentId}
Gets an AI Agent from a Project. Returns the AI Agent configuration, including identity, description, avatar, instructions, speaking style, voice, safety settings, and metadata such as creation and modification timestamps.



# Hire an AI Agent from the Job Market
Source: https://docs.cognigy.com/api-reference/ai-agents/hire-an-ai-agent-from-the-job-market

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/aiagents/hire
Hires an AI Agent in a Project using a ready-made template (`templateId`) with predefined skills for a specific job. You can optionally override an existing AI Agent using `overrideAiAgentReferenceId` to adapt it to your needs. The method returns a task tracking the status of the hiring process.



# Update an AI Agent
Source: https://docs.cognigy.com/api-reference/ai-agents/update-an-ai-agent

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/aiagents/{aiAgentId}
Updates an AI Agent in a Project. Allows modifying AI Agent persona, instructions, speaking style, voice, and safety settings. Returns no content on success



# Create an Amazon User
Source: https://docs.cognigy.com/api-reference/alexa/create-an-amazon-user

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/external/alexa/users
Creates an authenticated Amazon user by submitting their access token and refresh token. This method enables secure integration with Amazon Alexa services for the current Cognigy.AI user.



# Delete an Amazon User
Source: https://docs.cognigy.com/api-reference/alexa/delete-an-amazon-user

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/external/alexa/users/me
Deletes an already registered Amazon user from the Cognigy.AI platform.



# Deploy an Alexa Endpoint to an Alexa Skill
Source: https://docs.cognigy.com/api-reference/alexa/deploy-an-alexa-endpoint-to-an-alexa-skill

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /external/alexa/skills/{skillId}/deploy
Deploys a Cognigy.AI Alexa Endpoint to an Alexa skill.



# Get Alexa Skills
Source: https://docs.cognigy.com/api-reference/alexa/get-alexa-skills

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/external/alexa/skills
Gets a list of all Alexa skills associated with the current user, including details like skill ID, name per locale, last update, and lifecycle stage.



# Get Amazon User Details
Source: https://docs.cognigy.com/api-reference/alexa/get-amazon-user-details

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/external/alexa/users/me
Gets details about an already registered Amazon user.



# Get Call Counter Metrics for a Project
Source: https://docs.cognigy.com/api-reference/analytics-30/get-call-counter-metrics-for-a-project

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v3.0/projects/{projectId}/callcounter
Gets call counter metrics for a Project by its ID. Data is available up to two days ago, and only for 2025 onwards.



# Get Call Counter Metrics for the Organization
Source: https://docs.cognigy.com/api-reference/analytics-30/get-call-counter-metrics-for-the-organization

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v3.0/callcounter
Gets call counter metrics for the organization. Data is available up to two days ago, and only for 2025 onwards.



# Get Conversation Counter Metrics for a Project
Source: https://docs.cognigy.com/api-reference/analytics-30/get-conversation-counter-metrics-for-a-project

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v3.0/projects/{projectId}/conversationcounter
Gets conversation counter metrics for a Project by its ID. Data is available up to one day ago, and only for 2025 onwards.



# Get Conversation Counter Metrics for the Organization
Source: https://docs.cognigy.com/api-reference/analytics-30/get-conversation-counter-metrics-for-the-organization

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v3.0/conversationcounter
Gets conversation counter metrics for the organization. Data is available up to one day ago, and only for 2025 onwards.



# Delete Analytics Records
Source: https://docs.cognigy.com/api-reference/analytics/delete-analytics-records

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/analytics/{projectId}
Deletes analytics records for a Project.



# Get Call Counter Metrics for a Project
Source: https://docs.cognigy.com/api-reference/analytics/get-call-counter-metrics-for-a-project

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/projects/{projectId}/callcounter
Gets call counter metrics for a Project by its ID.



# Get Call Counter Metrics for the Organization
Source: https://docs.cognigy.com/api-reference/analytics/get-call-counter-metrics-for-the-organization

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/callcounter
Gets call counter metrics for the organization.



# Get Conversation Counter Metrics for a Project
Source: https://docs.cognigy.com/api-reference/analytics/get-conversation-counter-metrics-for-a-project

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/projects/{projectId}/conversationcounter
Gets conversation counter metrics for a Project by its ID.



# Get Conversation Counter Metrics for the Organization
Source: https://docs.cognigy.com/api-reference/analytics/get-conversation-counter-metrics-for-the-organization

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/conversationcounter
Gets conversation counter metrics for the organization.



# Get Knowledge Query Counter Metrics for a Project.
Source: https://docs.cognigy.com/api-reference/analytics/get-knowledge-query-counter-metrics-for-a-project

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/projects/{projectId}/knowledgequerycounter
Gets knowledge query counter metrics for a Project by its ID.



# Get Knowledge Query Counter Metrics for an Organization.
Source: https://docs.cognigy.com/api-reference/analytics/get-knowledge-query-counter-metrics-for-an-organization

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/knowledgequerycounter
Gets knowledge query counter metrics for the organization.



# Update an Analytics Record
Source: https://docs.cognigy.com/api-reference/analytics/update-an-analytics-record

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/analytics
Updates an analytics record for a specific user input. Allows modifying conversation properties, Flow and Intent details, ratings, handover information, Slots, tracked goals, and custom fields.



# Get an Audit Event
Source: https://docs.cognigy.com/api-reference/audit-events/get-an-audit-event

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/auditevents/{auditEventId}
Gets an audit event by its ID.



# Get Audit Events
Source: https://docs.cognigy.com/api-reference/audit-events/get-audit-events

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/auditevents
Gets audit events for the organization. By default, 25 audit events are returned. Use the `limit` query parameter to change this value.



# Create a Connection
Source: https://docs.cognigy.com/api-reference/connections/create-a-connection

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/connections
Creates a Connection in a Project.



# Create a Connection Field
Source: https://docs.cognigy.com/api-reference/connections/create-a-connection-field

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/connections/{connectionId}/fields
Creates a Connection field, allowing you to store additional key-value data in a Connection. This method is useful for adding custom settings required by external services or integrations.



# Delete a Connection
Source: https://docs.cognigy.com/api-reference/connections/delete-a-connection

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/connections/{connectionId}
Deletes a Connection from a Project. Deleting a Connection stops any dependent features from working. Flows that use the deleted Connection aren't updated automatically and may hang at the affected Node, causing errors.



# Delete a Connection Field
Source: https://docs.cognigy.com/api-reference/connections/delete-a-connection-field

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/connections/{connectionId}/fields/{fieldName}
Deletes a Connection field, removing the associated key-value data from a Connection. This method is useful for removing unused or sensitive data from a Connection.



# Get a Connection
Source: https://docs.cognigy.com/api-reference/connections/get-a-connection

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/connections/{connectionId}
Gets a Connection from a Project.



# Get Connection Schemas
Source: https://docs.cognigy.com/api-reference/connections/get-connection-schemas

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/connections/schemas
Gets Connection schemas from a Project. This method helps understand the structure and fields for creating or validating Connections. By default, 25 schemas are returned. Use the `limit` query parameter to change this value.



# Get Connections
Source: https://docs.cognigy.com/api-reference/connections/get-connections

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/connections
Gets Connections from a Project. By default, 25 Connections are returned. To change this value, use the `limit` query parameter.



# Perform Batch Operations on Connections
Source: https://docs.cognigy.com/api-reference/connections/perform-batch-operations-on-connections

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/connections
Performs batch operations on Connections. Executes batch create, update, and delete operations, returning the affected Connection IDs.



# Update a Connection
Source: https://docs.cognigy.com/api-reference/connections/update-a-connection

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/connections/{connectionId}
Updates a Connection in a Project, including type, credential and custom fields.



# Create a Contact Profile
Source: https://docs.cognigy.com/api-reference/contact-profiles/create-a-contact-profile

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/profiles
Creates a [Contact Profile](https://docs.cognigy.com/ai/agents/analyze/contact-profiles).



# Delete a Contact Profile
Source: https://docs.cognigy.com/api-reference/contact-profiles/delete-a-contact-profile

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/profiles/{profileId}
Deletes a [Contact Profile](https://docs.cognigy.com/ai/agents/analyze/contact-profiles).



# Delete Contact Profile Data
Source: https://docs.cognigy.com/api-reference/contact-profiles/delete-contact-profile-data

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/profiles/{profileId}/data/remove
Deletes data associated with a Contact Profile.



# Export Contact Profile
Source: https://docs.cognigy.com/api-reference/contact-profiles/export-contact-profile

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/profiles/{profileId}/data/export
Exports Contact Profile data in JSON format.



# Get a Contact Profile
Source: https://docs.cognigy.com/api-reference/contact-profiles/get-a-contact-profile

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/profiles/{profileId}
Gets the data of a [Contact Profile](https://docs.cognigy.com/ai/agents/analyze/contact-profiles).



# Get Contact Profiles
Source: https://docs.cognigy.com/api-reference/contact-profiles/get-contact-profiles

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/profiles
Gets [Contact Profiles](https://docs.cognigy.com/ai/agents/analyze/contact-profiles) in a Project. By default, 25 Contact Profiles are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results. This endpoint is deprecated, use `/v2.1/profiles` instead.



# Get Contact Profiles (performance)
Source: https://docs.cognigy.com/api-reference/contact-profiles/get-contact-profiles-performance

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.1/profiles
Faster variant of Get Contact Profiles. Numeric filter values are matched exactly against contact IDs; other values match first name, last name, or email. Alternatively, use the `exactMatch.*` parameters for an exact (indexed) match on a single field — mutually exclusive with `filter`. By default the total result count is not computed; pass `withTotal=true` to compute it.



# Get Schema
Source: https://docs.cognigy.com/api-reference/contact-profiles/get-schema

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/profiles/schema/{projectId}
Gets the [Contact Profile schema](https://docs.cognigy.com/ai/agents/analyze/contact-profiles#profile-schema).



# Merge Contact Profiles
Source: https://docs.cognigy.com/api-reference/contact-profiles/merge-contact-profiles

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/profiles/{profileId}/merge
Merges a [Contact Profile](https://docs.cognigy.com/ai/agents/analyze/contact-profiles) with another one.



# Remove the Contact ID
Source: https://docs.cognigy.com/api-reference/contact-profiles/remove-the-contact-id

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/profiles/{profileId}/contactids/remove
Removes the contact ID from a Contact Profile.



# Set Schema
Source: https://docs.cognigy.com/api-reference/contact-profiles/set-schema

https://api-trial.cognigy.ai/openapi/openapi-viewer.json put /v2.0/profiles/schema/{projectId}
Sets the [Contact Profile schema](https://docs.cognigy.com/ai/agents/analyze/contact-profiles#profile-schema).



# Unmerge Contact Profiles
Source: https://docs.cognigy.com/api-reference/contact-profiles/unmerge-contact-profiles

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/profiles/{profileId}/unmerge
Splits a merged [Contact Profile](https://docs.cognigy.com/ai/agents/analyze/contact-profiles) into two separate ones.



# Update a Contact Profile
Source: https://docs.cognigy.com/api-reference/contact-profiles/update-a-contact-profile

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/profiles/{profileId}
Updates Contact Profile data.



# Get a Conversation
Source: https://docs.cognigy.com/api-reference/conversations-v20/get-a-conversation

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/conversations/{sessionId}
Gets a conversation by its `sessionId`.



# Deletes conversations records for a sessionId.
Source: https://docs.cognigy.com/api-reference/conversations-v21/deletes-conversations-records-for-a-sessionid

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.1/conversations/{sessionId}
Deletes conversations records for a sessionId.



# Deletes a Conversation
Source: https://docs.cognigy.com/api-reference/conversations/deletes-a-conversation

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/conversations/{sessionId}
Deletes a conversation by its `sessionId`.
**Deprecated.** Use the Insights API [DELETE /v2.1/conversations/{sessionId}](#delete-/v2.1/conversations/-sessionId-) instead.




# Get Conversations
Source: https://docs.cognigy.com/api-reference/conversations/get-conversations

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/conversations
Gets conversations from a Project. You can also get conversations for a specific user by specifying their contact ID.



# Create an Endpoint
Source: https://docs.cognigy.com/api-reference/endpoints/create-an-endpoint

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/endpoints
Creates an Endpoint in a Project.



# Create an Endpoint API Key
Source: https://docs.cognigy.com/api-reference/endpoints/create-an-endpoint-api-key

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/endpoints/{endpointId}/apikeys
Generates a new API key for the given Webhook/Generic Endpoint. Returns the plaintext key in the response body exactly once — the server stores only the SHA-256 hash and a masked preview, and the plaintext is never retrievable again.




# Delete an Endpoint
Source: https://docs.cognigy.com/api-reference/endpoints/delete-an-endpoint

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/endpoints/{endpointId}
Deletes an Endpoint from a Project.



# Get an Endpoint
Source: https://docs.cognigy.com/api-reference/endpoints/get-an-endpoint

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/endpoints/{endpointId}
Gets an Endpoint from a Project.



# Get Endpoints
Source: https://docs.cognigy.com/api-reference/endpoints/get-endpoints

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/endpoints
Gets Endpoints from a Project. By default, 25 Endpoints are returned. To change this value, use the `limit` query parameter.



# List Endpoint API Keys
Source: https://docs.cognigy.com/api-reference/endpoints/list-endpoint-api-keys

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/endpoints/{endpointId}/apikeys
Lists all API keys configured on the given Webhook/Generic Endpoint. The response never contains the SHA-256 hash or the plaintext key — only the masked preview, name, id, and creation timestamp.




# Perform Batch Operations on Endpoints
Source: https://docs.cognigy.com/api-reference/endpoints/perform-batch-operations-on-endpoints

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/endpoints
Performs batch operations on Endpoints in a Project. Use this method to create, update, or delete multiple Endpoints in a request by specifying the operation type (`create`, `update`, or `delete`) for each item in the operations array.



# Revoke an Endpoint API Key
Source: https://docs.cognigy.com/api-reference/endpoints/revoke-an-endpoint-api-key

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/endpoints/{endpointId}/apikeys/{keyId}
Permanently revokes the given API key by hard-deleting it. Any client still using the revoked key will start failing immediately on the next inbound webhook request (subject to cache propagation).




# Send a Message as a User Input
Source: https://docs.cognigy.com/api-reference/endpoints/send-a-message-as-a-user-input

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/endpoint/inject
Sends a message as a user input in an active conversation, so an AI Agent can process the input and continue the conversation. For example, to follow up after user inactivity or trigger a reply after an external API event.



# Send a Message as an AI Agent Output
Source: https://docs.cognigy.com/api-reference/endpoints/send-a-message-as-an-ai-agent-output

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/endpoint/notify
Sends a message as an AI Agent Output. It bypasses the Flow's input processing and delivers a message to the user. For example, sending `Your order 12345 has shipped` as a status update.



# Update an Endpoint
Source: https://docs.cognigy.com/api-reference/endpoints/update-an-endpoint

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/endpoints/{endpointId}
Updates Endpoint settings in a Project. Allows modifying the Endpoint's activation status, Transformers, handover, analytics, and privacy options.



# Create an Example Sentence
Source: https://docs.cognigy.com/api-reference/example-sentences/create-an-example-sentence

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/flows/{flowId}/intents/{intentId}/sentences
Creates an [Example Sentence](https://docs.cognigy.com/ai/platform-features/nlu/intents/example-sentences) for an Intent.



# Delete an Example Sentence
Source: https://docs.cognigy.com/api-reference/example-sentences/delete-an-example-sentence

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/flows/{flowId}/intents/{intentId}/sentences/{sentenceId}
Deletes an [Example Sentence](https://docs.cognigy.com/ai/platform-features/nlu/intents/example-sentences) from an Intent.



# Generate Example Sentences
Source: https://docs.cognigy.com/api-reference/example-sentences/generate-example-sentences

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/flows/{flowId}/intents/{intentId}/sentences/generate
Generates [Example Sentences](https://docs.cognigy.com/ai/platform-features/nlu/intents/example-sentences) for an Intent using generative AI.



# Get an Example Sentence
Source: https://docs.cognigy.com/api-reference/example-sentences/get-an-example-sentence

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/flows/{flowId}/intents/{intentId}/sentences/{sentenceId}
Gets an [Example Sentence](https://docs.cognigy.com/ai/platform-features/nlu/intents/example-sentences) from an Intent.



# Get Example Sentences
Source: https://docs.cognigy.com/api-reference/example-sentences/get-example-sentences

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/flows/{flowId}/intents/{intentId}/sentences
Gets [Example Sentences](https://docs.cognigy.com/ai/platform-features/nlu/intents/example-sentences) from an Intent. By default, 25 Example Sentences are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Perform Batch Operations
Source: https://docs.cognigy.com/api-reference/example-sentences/perform-batch-operations

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/flows/{flowId}/intents/{intentId}/sentences
Performs batch operations on [Sentences](https://docs.cognigy.com/ai/platform-features/nlu/intents/example-sentences) in an Intent.



# Update an Example Sentence
Source: https://docs.cognigy.com/api-reference/example-sentences/update-an-example-sentence

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/flows/{flowId}/intents/{intentId}/sentences/{sentenceId}
Updates an [Example Sentence](https://docs.cognigy.com/ai/platform-features/nlu/intents/example-sentences) for an Intent.



# Delete an Extension
Source: https://docs.cognigy.com/api-reference/extensions/delete-an-extension

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/extensions/{extensionId}
Deletes an Extension from a Project.



# Get an Extension
Source: https://docs.cognigy.com/api-reference/extensions/get-an-extension

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/extensions/{extensionId}
Gets an Extension from a Project.



# Get Extensions
Source: https://docs.cognigy.com/api-reference/extensions/get-extensions

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/extensions
Gets Extensions from a Project. By default, 25 Extensions are returned. To change this value, use the `limit` query parameter.



# Update an Extension
Source: https://docs.cognigy.com/api-reference/extensions/update-an-extension

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/extensions/update
Updates an existing uploaded Extension in a Project by replacing it with a new version. You can provide the Extension either as a file upload or via a URL. The method creates a Task to process the update, and the response returns the task status and progress so you can track when the replacement is completed.



# Update Extension Settings
Source: https://docs.cognigy.com/api-reference/extensions/update-extension-settings

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/extensions/{extensionId}
Updates Extension settings, such as enabling or disabling trusted code behavior.



# Upload an Extension
Source: https://docs.cognigy.com/api-reference/extensions/upload-an-extension

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/extensions/upload
Upload an Extension to a Project. The Extension can be provided as a file or via a URL. The method creates a Task to process the creation, and the response includes the task status and progress so you can track completion. This upload always creates a new Extension and doesn't replace any existing ones.



# Get Binary Data of a File
Source: https://docs.cognigy.com/api-reference/files/get-binary-data-of-a-file

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/files/{fileToken}
Gets the binary data of a file object using its file token.



# Get File Metadata
Source: https://docs.cognigy.com/api-reference/files/get-file-metadata

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/files/{fileId}
Gets file metadata without the binary content.



# Add a Locale to a Flow
Source: https://docs.cognigy.com/api-reference/flows/add-a-locale-to-a-flow

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/flows/{flowId}/addlocalization
Adds a Locale to a Flow.



# Clone a Flow
Source: https://docs.cognigy.com/api-reference/flows/clone-a-flow

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/flows/{flowId}/clone
Creates a copy of an existing Flow in a Project.



# Create a Flow
Source: https://docs.cognigy.com/api-reference/flows/create-a-flow

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/flows
Creates a Flow in a Project.



# Create a Flow from a Child Node
Source: https://docs.cognigy.com/api-reference/flows/create-a-flow-from-a-child-node

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/flows/{flowId}/chart/nodes/childflow
Creates a Flow from a child Node. Moves the Nodes following the child Node to the newly created Flow. The original Flow receives an Execute Flow Node pointing to the new Flow. This method only works if the child Node is followed by other Nodes.



# Delete a Flow
Source: https://docs.cognigy.com/api-reference/flows/delete-a-flow

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/flows/{flowId}
Deletes a Flow from a Project.



# Get a Flow
Source: https://docs.cognigy.com/api-reference/flows/get-a-flow

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/flows/{flowId}
Gets a Flow from a Project.



# Get Flow Settings
Source: https://docs.cognigy.com/api-reference/flows/get-flow-settings

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/flows/{flowId}/settings
Gets Flow settings, including execution behavior, Slot parsing, Intent mapping, default replies, locale settings, thresholds, Yes/No logic, and Slot usage. All metadata, such as creation and modification timestamps and user information, is also included.



# Get Flows
Source: https://docs.cognigy.com/api-reference/flows/get-flows

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/flows
Gets Flows from a Project. By default, 25 Flows are returned. To change this value, use the `limit` query parameter. Additionally, use `withAiAgents=true` to include AI Agent information for each Flow.



# Get the Chart of specified Flow
Source: https://docs.cognigy.com/api-reference/flows/get-the-chart-of-specified-flow

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/flows/{flowId}/chart
Gets the Chart of the specified Flow.



# Perform Batch Operations on Flows
Source: https://docs.cognigy.com/api-reference/flows/perform-batch-operations-on-flows

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/flows
Performs batch operations on Flows.



# Remove a Locale from a Flow
Source: https://docs.cognigy.com/api-reference/flows/remove-a-locale-from-a-flow

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/flows/{flowId}/removelocalization
Removes a Locale from a Flow in a Project.



# Update a Flow
Source: https://docs.cognigy.com/api-reference/flows/update-a-flow

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/flows/{flowId}
Updates a Flow in a Project. Allows modifying of a Flow's name, description, and locale.



# Update Flow Settings
Source: https://docs.cognigy.com/api-reference/flows/update-flow-settings

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/flows/{flowId}/settings
Updates Flow settings, including execution behavior, Slot parsing, Intent mapping, default replies, Locale settings, thresholds, Yes/No logic, and Slot usage.



# Create a Function
Source: https://docs.cognigy.com/api-reference/functions/create-a-function

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/functions
Creates a Function in a Project.



# Create a Function Instance
Source: https://docs.cognigy.com/api-reference/functions/create-a-function-instance

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/functions/{functionId}/trigger
Creates a Function instance. This method triggers the function scheduler to execute the Function immediately.



# Delete a Function
Source: https://docs.cognigy.com/api-reference/functions/delete-a-function

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/functions/{functionId}
Deletes a Function from a Project.



# Get a Function
Source: https://docs.cognigy.com/api-reference/functions/get-a-function

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/functions/{functionId}
Gets a Function from a Project.



# Get a Function Instance
Source: https://docs.cognigy.com/api-reference/functions/get-a-function-instance

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/functions/{functionId}/instances/{functionInstanceId}
Gets a Function instance from a Function.



# Get Function Instances
Source: https://docs.cognigy.com/api-reference/functions/get-function-instances

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/functions/{functionId}/instances
Gets Function instances from a Function. By default, 25 Function instances are returned. To change this value, use the `limit` query parameter.



# Get Functions
Source: https://docs.cognigy.com/api-reference/functions/get-functions

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/functions
Gets Functions from a Project. By default, 25 Functions are returned. To change this value, use the `limit` query parameter.



# Stop a Function Instance
Source: https://docs.cognigy.com/api-reference/functions/stop-a-function-instance

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/functions/{functionId}/instances/{functionInstanceId}/stop
Stops a Function instance. This method immediately halts the running function.



# Update a Function
Source: https://docs.cognigy.com/api-reference/functions/update-a-function

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/functions/{functionId}
Updates a Function in a Project. Allows modifying of a Function’s name, source code, or enabled state. You can also disable the Function without deleting it.



# Create a Handover Provider
Source: https://docs.cognigy.com/api-reference/handover-providers/create-a-handover-provider

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/handover/providers
Creates a handover provider in a Project.



# Delete a Handover Provider
Source: https://docs.cognigy.com/api-reference/handover-providers/delete-a-handover-provider

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/handover/providers/{providerId}
Deletes a handover provider from a Project.



# Get a Handover Provider
Source: https://docs.cognigy.com/api-reference/handover-providers/get-a-handover-provider

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/handover/providers/{providerId}
Gets a handover provider from a Project.



# Get Handover Providers
Source: https://docs.cognigy.com/api-reference/handover-providers/get-handover-providers

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/handover/providers
Gets handover providers for a Project.



# Update a Handover Provider
Source: https://docs.cognigy.com/api-reference/handover-providers/update-a-handover-provider

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/handover/providers/{providerId}
Updates a handover provider in a Project.



# Get a Handover Service
Source: https://docs.cognigy.com/api-reference/handover-services/get-a-handover-service

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/handover/services/{serviceId}
Gets a handover service from a Project.



# Get Handover Services
Source: https://docs.cognigy.com/api-reference/handover-services/get-handover-services

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/handover/services
Gets handover services.



# Reset an Identity Provider Setup
Source: https://docs.cognigy.com/api-reference/identity-providers/reset-an-identity-provider-setup

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/identityprovider/reset
Resets an identity provider setup for the organization by removing the configured identity provider and all related authentication settings. Use this method to clear the existing SSO configuration.



# Set Up an Identity Provider
Source: https://docs.cognigy.com/api-reference/identity-providers/set-up-an-identity-provider

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/identityprovider/configure
Sets up an identity provider (SAML or OIDC) for your organization, including login/logout URLs, certificates, client credentials, and security settings, enabling Single Sign-On (SSO) authentication.



# Cognigy API
Source: https://docs.cognigy.com/api-reference/index



The Cognigy API allows developers to manage the lifecycle of AI Agents, including deployment, configuration, and real-time interactions. It also provides administrative tools for monitoring resource usage and managing users.

<CardGroup>
  <Card title="Resource Monitoring" icon="eye" href="/api-reference/administration/get-knowledge-chunk-count">
    Track key administrative metrics, such as Knowledge Base resource usage and counts.
  </Card>

  <Card title="Deployment & Snapshots" icon="cloud-download" href="/api-reference/snapshots/get-snapshots">
    Manage the configuration lifecycle: deploy AI Agents, create and upload Snapshots.
  </Card>

  <Card title="Flow Interaction" icon="message-square" href="/api-reference/endpoints/send-a-message-as-a-user-input">
    Inject messages, reset session context, and control Flows.
  </Card>

  <Card title="Administration & Setup" icon="settings" href="/api-reference/management-ui/create-an-api-key">
    Manage organization-level security and access, including creating users and API Keys.
  </Card>

  <Card title="Flow Configuration" icon="layout-list" href="/api-reference/flows/get-flow-settings">
    Access and modify the core settings and structure of individual Flows.
  </Card>

  <Card title="Simulator" icon="flask-conical" href="/api-reference-simulator">
    Manage scenarios, run simulations, schedule automated tests, and retrieve evaluation results.
  </Card>

  <Card title="Logs" icon="bug" href="/api-reference-logs">
    Retrieve the latest execution logs and debug entries to troubleshoot Flow behavior.
  </Card>
</CardGroup>


# Create a Download Link for Intent Trainer Records
Source: https://docs.cognigy.com/api-reference/intent-trainer/create-a-download-link-for-intent-trainer-records

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/trainer/downloadlink
Creates a link to download packaged [Intent Trainer records](https://docs.cognigy.com/ai/platform-features/nlu/intents/intent-trainer).



# Get an Intent Trainer Record
Source: https://docs.cognigy.com/api-reference/intent-trainer/get-an-intent-trainer-record

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/trainer/{recordId}
Gets an [Intent Trainer record](https://docs.cognigy.com/ai/platform-features/nlu/intents/intent-trainer) from a Project.



# Get Trainer Records
Source: https://docs.cognigy.com/api-reference/intent-trainer/get-trainer-records

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/trainer
Gets [Intent Trainer records](https://docs.cognigy.com/ai/platform-features/nlu/intents/intent-trainer) in a Project. By default, 25 Trainer Records are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Package Intent Trainer Records
Source: https://docs.cognigy.com/api-reference/intent-trainer/package-intent-trainer-records

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/trainer/package
Creates a task that packages [Intent Trainer records](https://docs.cognigy.com/ai/platform-features/nlu/intents/intent-trainer).



# Perform Batch Operations on Intent Trainer Records
Source: https://docs.cognigy.com/api-reference/intent-trainer/perform-batch-operations-on-intent-trainer-records

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/trainer/batch
Performs batch operations on [Intent Trainer records](https://docs.cognigy.com/ai/platform-features/nlu/intents/intent-trainer).



# Upload an Intent Trainer Records Package
Source: https://docs.cognigy.com/api-reference/intent-trainer/upload-an-intent-trainer-records-package

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/trainer/upload
Uploads a package with [Intent Trainer records](https://docs.cognigy.com/ai/platform-features/nlu/intents/intent-trainer).



# Create an Intent
Source: https://docs.cognigy.com/api-reference/intents/create-an-intent

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/flows/{flowId}/intents
Creates an Intent in a Flow.



# Delete an Intent
Source: https://docs.cognigy.com/api-reference/intents/delete-an-intent

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/flows/{flowId}/intents/{intentId}
Deletes an Intent from a Flow.



# Delete Intent Localization
Source: https://docs.cognigy.com/api-reference/intents/delete-intent-localization

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/flows/{flowId}/intents/{intentId}/removelocalization
Deletes a localization from an Intent. Use this method to delete translations or Locale-specific data associated with the Intent.



# Export Intents
Source: https://docs.cognigy.com/api-reference/intents/export-intents

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/flows/{flowId}/intents/export
Exports the Intents of a Flow. You can specify a Locale and select the export format, CSV or JSON file.



# Get an Intent
Source: https://docs.cognigy.com/api-reference/intents/get-an-intent

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/flows/{flowId}/intents/{intentId}
Gets an Intent from a Flow.



# Get Intents
Source: https://docs.cognigy.com/api-reference/intents/get-intents

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/flows/{flowId}/intents
Gets Intents from a Flow. By default, 25 Intents are returned. To change this value, use the `limit` query parameter.



# Localize an Intent
Source: https://docs.cognigy.com/api-reference/intents/localize-an-intent

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/flows/{flowId}/intents/{intentId}/addlocalization
Adds a localized version of an Intent in a Flow. You can specify the target Locale, optionally inherit content from another Locale, and provide translation settings to adapt the Intent for different languages or regions.



# Perform Batch Operations on Intents
Source: https://docs.cognigy.com/api-reference/intents/perform-batch-operations-on-intents

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/flows/{flowId}/intents
Performs batch operations on Intents in a Flow. Use this method to create, update, or delete multiple Intents in a request by specifying the operation type (`create`, `update`, or `delete`) for each item in the operations array.



# Train Intents
Source: https://docs.cognigy.com/api-reference/intents/train-intents

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/flows/{flowId}/intents/train
Starts a Task to train the Intents of a Flow.



# Update an Intent
Source: https://docs.cognigy.com/api-reference/intents/update-an-intent

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/flows/{flowId}/intents/{intentId}
Update an Intent in a Flow. Allows modifying of an Intent's name, description, rules, tags, and status. You can also enable or disable the Intent, mark it as a reject Intent, adjust child/parent intent relationships, and override default reply settings.



# Upload Intents
Source: https://docs.cognigy.com/api-reference/intents/upload-intents

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/flows/{flowId}/intents/import
Uploads Intents and creates a Task to import the Intents. The supported file format is CSV.



# Create a Knowledge Chunk
Source: https://docs.cognigy.com/api-reference/knowledge-chunks/create-a-knowledge-chunk

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/knowledgestores/{knowledgeStoreId}/sources/{sourceId}/chunks
Creates a Knowledge Chunk in a Knowledge Resource.



# Delete a Knowledge Chunk
Source: https://docs.cognigy.com/api-reference/knowledge-chunks/delete-a-knowledge-chunk

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/knowledgestores/{knowledgeStoreId}/sources/{sourceId}/chunks/{chunkId}
Deletes a Knowledge Chunk from a Knowledge Source.



# Get a Knowledge Chunk
Source: https://docs.cognigy.com/api-reference/knowledge-chunks/get-a-knowledge-chunk

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/knowledgestores/{knowledgeStoreId}/sources/{sourceId}/chunks/{chunkId}
Gets a Knowledge Chunk from a Knowledge Source.



# Get Knowledge Chunks
Source: https://docs.cognigy.com/api-reference/knowledge-chunks/get-knowledge-chunks

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/knowledgestores/{knowledgeStoreId}/sources/{sourceId}/chunks
Gets Knowledge Chunks from a Knowledge Source. By default, 25 Knowledge Chunks are returned. To change this value, use the `limit` query parameter.



# Update a Knowledge Chunk
Source: https://docs.cognigy.com/api-reference/knowledge-chunks/update-a-knowledge-chunk

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/knowledgestores/{knowledgeStoreId}/sources/{sourceId}/chunks/{chunkId}
Updates a Knowledge Chunk in a Knowledge Source. Allows modifying the Chunk's text, data, order, and status.



# Create a Knowledge Connector
Source: https://docs.cognigy.com/api-reference/knowledge-connectors/create-a-knowledge-connector

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/knowledgestores/{knowledgeStoreId}/connectors
Creates a Knowledge Connector.



# Delete a Knowledge Connector
Source: https://docs.cognigy.com/api-reference/knowledge-connectors/delete-a-knowledge-connector

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/knowledgestores/{knowledgeStoreId}/connectors/{connectorId}
Deletes a Knowledge Connector from a Knowledge Store.



# Get a Knowledge Connector
Source: https://docs.cognigy.com/api-reference/knowledge-connectors/get-a-knowledge-connector

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/knowledgestores/{knowledgeStoreId}/connectors/{connectorId}
Gets a Knowledge Connector from a Project.



# Get Knowledge Connectors
Source: https://docs.cognigy.com/api-reference/knowledge-connectors/get-knowledge-connectors

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/knowledgestores/{knowledgeStoreId}/connectors
Gets a list of Knowledge Connectors.



# Get Knowledge Connectors
Source: https://docs.cognigy.com/api-reference/knowledge-connectors/get-knowledge-connectors-1

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/knowledgestores/{knowledgeStoreId}/descriptors
Gets information about the Knowledge Connectors from a Project.



# Run a Knowledge Connector
Source: https://docs.cognigy.com/api-reference/knowledge-connectors/run-a-knowledge-connector

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/knowledgestores/{knowledgeStoreId}/connectors/{connectorId}/run
Runs a Knowledge Connector to extract and import knowledge from external sources.



# Update a Knowledge Connector
Source: https://docs.cognigy.com/api-reference/knowledge-connectors/update-a-knowledge-connector

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/knowledgestores/{knowledgeStoreId}/connectors/{connectorId}
Updates a Knowledge Connector resource in a Project. Allows modifying a Knowledge Connector’s configuration and name.



# Create a Knowledge Source
Source: https://docs.cognigy.com/api-reference/knowledge-sources/create-a-knowledge-source

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/knowledgestores/{knowledgeStoreId}/sources
Creates a Knowledge Source in a Knowledge Store. You define the source metadata and type—no file is uploaded. Use this method for manual sources, URL sources (for web scraping), or when creating the source first and ingesting content later (pdf, txt, ctxt).



# Create a Knowledge Source from a File
Source: https://docs.cognigy.com/api-reference/knowledge-sources/create-a-knowledge-source-from-a-file

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/knowledgestores/{knowledgeStoreId}/sources/upload
Creates a Knowledge Source in the Knowledge Store by uploading a file and asynchronously ingesting its content. Supported file formats pdf, txt, ctxt.



# Delete a Knowledge Source
Source: https://docs.cognigy.com/api-reference/knowledge-sources/delete-a-knowledge-source

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/knowledgestores/{knowledgeStoreId}/sources/{sourceId}
Deletes a Knowledge Source from a Knowledge Store.



# Get a Knowledge Source
Source: https://docs.cognigy.com/api-reference/knowledge-sources/get-a-knowledge-source

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/knowledgestores/{knowledgeStoreId}/sources/{sourceId}
Gets a Knowledge Source from a Knowledge Store.



# Get Knowledge Sources
Source: https://docs.cognigy.com/api-reference/knowledge-sources/get-knowledge-sources

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/knowledgestores/{knowledgeStoreId}/sources
Gets Knowledge Sources from a Knowledge Store. By default, 25 Knowledge Sources are returned. You can change this number using the `limit` parameter.



# Run a Knowledge Extension
Source: https://docs.cognigy.com/api-reference/knowledge-sources/run-a-knowledge-extension

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/knowledgestores/{knowledgeStoreId}/extensions/run
Runs a Knowledge Extension to extract and import knowledge from external sources.



# Updates a Knowledge Source
Source: https://docs.cognigy.com/api-reference/knowledge-sources/updates-a-knowledge-source

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/knowledgestores/{knowledgeStoreId}/sources/{sourceId}
Updates a Knowledge Source in a Knowledge Store. Allows modifying a Source's name, description, status, and metadata.



# Create a Knowledge Store
Source: https://docs.cognigy.com/api-reference/knowledge-stores/create-a-knowledge-store

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/knowledgestores
Creates a Knowledge Store in a Project.



# Delete a Knowledge Store
Source: https://docs.cognigy.com/api-reference/knowledge-stores/delete-a-knowledge-store

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/knowledgestores/{knowledgeStoreId}
Deletes a Knowledge Store from a Project.



# Get a Knowledge Store
Source: https://docs.cognigy.com/api-reference/knowledge-stores/get-a-knowledge-store

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/knowledgestores/{knowledgeStoreId}
Gets a Knowledge Store from a Project.



# Get Knowledge Stores
Source: https://docs.cognigy.com/api-reference/knowledge-stores/get-knowledge-stores

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/knowledgestores
Gets Knowledge Stores from a Project. By default, 25 Knowledge Stores are returned. To change this value, use the `limit` query parameter.



# Update a Knowledge Store
Source: https://docs.cognigy.com/api-reference/knowledge-stores/update-a-knowledge-store

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/knowledgestores/{knowledgeStoreId}
Updates a Knowledge Store in a Project. Allows modifying a Store's name and description.



# Clone a large language model
Source: https://docs.cognigy.com/api-reference/large-language-models/clone-a-large-language-model

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/largelanguagemodels/{largeLanguageModelId}/clone
Clones a [large language model](https://docs.cognigy.com/ai/agents/develop/gen-ai-and-llms/llms) in a Project.



# Create a large language model
Source: https://docs.cognigy.com/api-reference/large-language-models/create-a-large-language-model

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/largelanguagemodels
Creates a large language model. The large language model can have the following scopes:
- Project (`resourceLevel: 'project'`)
- Global (`resourceLevel: 'organisation'`)

Depending on the scope, different fields are required or must be excluded:
- **Project-level resources**: `projectId` is required.
- **Global resources**: `isDefault` and `fallback` must be excluded.




# Delete a large language model
Source: https://docs.cognigy.com/api-reference/large-language-models/delete-a-large-language-model

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/largelanguagemodels/{largeLanguageModelId}
Deletes a [large language model](https://docs.cognigy.com/ai/agents/develop/gen-ai-and-llms/llms) from a Project.



# Get a large language model
Source: https://docs.cognigy.com/api-reference/large-language-models/get-a-large-language-model

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/largelanguagemodels/{largeLanguageModelId}
Gets data about a [large language model](https://docs.cognigy.com/ai/agents/develop/gen-ai-and-llms/llms).



# Get large language models
Source: https://docs.cognigy.com/api-reference/large-language-models/get-large-language-models

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/largelanguagemodels
Gets [large language models](https://docs.cognigy.com/ai/agents/develop/gen-ai-and-llms/llms) in a Project. By default, 25 large language models are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Get models for a large language model provider
Source: https://docs.cognigy.com/api-reference/large-language-models/get-models-for-a-large-language-model-provider

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/largelanguagemodels/{largeLanguageModelId}/availablemodels
Gets the available models for a large language model provider. Returns an empty array if the provider doesn't support model listing.



# Test large language model connection
Source: https://docs.cognigy.com/api-reference/large-language-models/test-large-language-model-connection

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/largelanguagemodels/{largeLanguageModelId}/test
Tests the connection to the large language model provider by validating the credentials.



# Update a large language model
Source: https://docs.cognigy.com/api-reference/large-language-models/update-a-large-language-model

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/largelanguagemodels/{largeLanguageModelId}
Updates a [large language model](https://docs.cognigy.com/ai/agents/develop/gen-ai-and-llms/llms#large-language-models-llms).



# Add a Keyphrase to a Lexicon
Source: https://docs.cognigy.com/api-reference/lexicons/add-a-keyphrase-to-a-lexicon

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/lexicons/{lexiconId}/entries/{entryId}/keyphrases
Adds a [Keyphrase](https://docs.cognigy.com/ai/platform-features/nlu/slots/overview#keyphrases) to a [Lexicon](https://docs.cognigy.com/ai/platform-features/nlu/slots/user-defined/lexicon).



# Add a Slot to a Lexicon
Source: https://docs.cognigy.com/api-reference/lexicons/add-a-slot-to-a-lexicon

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/lexicons/{lexiconId}/entries/{entryId}/slots
Adds a [Slot](https://docs.cognigy.com/ai/platform-features/nlu/slots/user-defined/lexicon) to a Lexicon.



# Create a Download Link for a Lexicon
Source: https://docs.cognigy.com/api-reference/lexicons/create-a-download-link-for-a-lexicon

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/lexicons/{lexiconId}/downloadlink
Creates a link to download a [Lexicon](https://docs.cognigy.com/ai/platform-features/nlu/slots/user-defined/lexicon) as a CSV file.



# Create a Lexicon Entry
Source: https://docs.cognigy.com/api-reference/lexicons/create-a-lexicon-entry

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/lexicons/{lexiconId}/entries
Creates a Lexicon entry with a Keyphrase and optional Lexicon Slots.



# Create a Lexicon Slot
Source: https://docs.cognigy.com/api-reference/lexicons/create-a-lexicon-slot

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/lexicons/{lexiconId}/slots
Creates a [Lexicon Slot](https://docs.cognigy.com/ai/platform-features/nlu/slots/user-defined/lexicon) in a Lexicon. You can assign Slots to Keyphrases.



# Create a New Lexicon
Source: https://docs.cognigy.com/api-reference/lexicons/create-a-new-lexicon

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/lexicons
Creates a [Lexicon](https://docs.cognigy.com/ai/platform-features/nlu/slots/user-defined/lexicon) in a Project.



# Delete a Keyphrase
Source: https://docs.cognigy.com/api-reference/lexicons/delete-a-keyphrase

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/lexicons/{lexiconId}/entries/{entryId}/keyphrases/{keyphraseId}
Deletes a [Keyphrase](https://docs.cognigy.com/ai/platform-features/nlu/slots/user-defined/lexicon) from a Lexicon entry.



# Delete a Lexicon
Source: https://docs.cognigy.com/api-reference/lexicons/delete-a-lexicon

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/lexicons/{lexiconId}
Deletes a [Lexicon](https://docs.cognigy.com/ai/platform-features/nlu/slots/user-defined/lexicon) from a Project.



# Delete a Lexicon Entry
Source: https://docs.cognigy.com/api-reference/lexicons/delete-a-lexicon-entry

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/lexicons/{lexiconId}/entries/{entryId}
Deletes a [Lexicon entry](https://docs.cognigy.com/ai/platform-features/nlu/slots/user-defined/lexicon) from a Lexicon.



# Delete a Lexicon Slot
Source: https://docs.cognigy.com/api-reference/lexicons/delete-a-lexicon-slot

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/lexicons/{lexiconId}/entries/{entryId}/slots/{slotId}
Deletes a [Lexicon Slot](https://docs.cognigy.com/ai/platform-features/nlu/slots/user-defined/lexicon) from a Lexicon entry.



# Delete a Lexicon Slot
Source: https://docs.cognigy.com/api-reference/lexicons/delete-a-lexicon-slot-1

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/lexicons/{lexiconId}/slots/{slotId}
Deletes a [Lexicon Slot](https://docs.cognigy.com/ai/platform-features/nlu/slots/user-defined/lexicon) from a Lexicon.



# Export a Lexicon
Source: https://docs.cognigy.com/api-reference/lexicons/export-a-lexicon

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/lexicons/{lexiconId}/export
Creates a task to export a [Lexicon](https://docs.cognigy.com/ai/platform-features/nlu/slots/user-defined/lexicon) as a CSV file.



# Get a Lexicon
Source: https://docs.cognigy.com/api-reference/lexicons/get-a-lexicon

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/lexicons/{lexiconId}
Gets a [Lexicon](https://docs.cognigy.com/ai/platform-features/nlu/slots/user-defined/lexicon) from a Project.



# Get a Lexicon Entry
Source: https://docs.cognigy.com/api-reference/lexicons/get-a-lexicon-entry

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/lexicons/{lexiconId}/entries/{entryId}
Gets a Lexicon entry from a [Lexicon](https://docs.cognigy.com/ai/platform-features/nlu/slots/user-defined/lexicon).



# Get Keyphrases from a Lexicon Entry
Source: https://docs.cognigy.com/api-reference/lexicons/get-keyphrases-from-a-lexicon-entry

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/lexicons/{lexiconId}/entries/{entryId}/keyphrases
Gets [Keyphrases](https://docs.cognigy.com/ai/platform-features/nlu/slots/user-defined/lexicon) from a Lexicon entry. By default, 25 Keyphrases are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Get Keyphrases in a Lexicon
Source: https://docs.cognigy.com/api-reference/lexicons/get-keyphrases-in-a-lexicon

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/lexicons/{lexiconId}/keyphrases
Gets [Keyphrases](https://docs.cognigy.com/ai/platform-features/nlu/slots/user-defined/lexicon) from a Lexicon. By default, 25 Keyphrases are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Get Lexicon Entries
Source: https://docs.cognigy.com/api-reference/lexicons/get-lexicon-entries

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/lexicons/{lexiconId}/entries
Gets entries from a [Lexicon](https://docs.cognigy.com/ai/platform-features/nlu/slots/user-defined/lexicon). By default, 25 entries are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Get Lexicon Slots
Source: https://docs.cognigy.com/api-reference/lexicons/get-lexicon-slots

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/lexicons/{lexiconId}/slots
Gets [Lexicon Slots](https://docs.cognigy.com/ai/platform-features/nlu/slots/user-defined/lexicon) from a Lexicon. By default, 25 Lexicon Slots are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Get Lexicons
Source: https://docs.cognigy.com/api-reference/lexicons/get-lexicons

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/lexicons
Gets [Lexicons](https://docs.cognigy.com/ai/resources/build/lexicons/) in a Project. By default, 25 Lexicons are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Import a Lexicon
Source: https://docs.cognigy.com/api-reference/lexicons/import-a-lexicon

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/lexicons/{lexiconId}/import
Create a task to import a [Lexicon](https://docs.cognigy.com/ai/platform-features/nlu/slots/user-defined/lexicon) from a CSV file.



# Perform Batch Operations in a Lexicon
Source: https://docs.cognigy.com/api-reference/lexicons/perform-batch-operations-in-a-lexicon

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/lexicons/{lexiconId}/entries
Performs batch create, update, or delete operations on Keyphrases and Lexicon Slots in a single request.



# Perform Batch Operations in a Lexicon
Source: https://docs.cognigy.com/api-reference/lexicons/perform-batch-operations-in-a-lexicon-1

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/lexicons/{lexiconId}/slots
Performs batch create, update, or delete operations on Lexicon Slots in a single request.



# Perform Batch Operations on Lexicons
Source: https://docs.cognigy.com/api-reference/lexicons/perform-batch-operations-on-lexicons

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/lexicons
Performs batch create, update, or delete operations on Lexicons in a single request.



# Update a Keyphrase
Source: https://docs.cognigy.com/api-reference/lexicons/update-a-keyphrase

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/lexicons/{lexiconId}/keyphrases/{keyphraseId}
Updates a [Keyphrase](https://docs.cognigy.com/ai/platform-features/nlu/slots/user-defined/lexicon) in a Lexicon.



# Update a Lexicon
Source: https://docs.cognigy.com/api-reference/lexicons/update-a-lexicon

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/lexicons/{lexiconId}
Updates a [Lexicon](https://docs.cognigy.com/ai/platform-features/nlu/slots/user-defined/lexicon) in a Project.



# Update a Lexicon Entry
Source: https://docs.cognigy.com/api-reference/lexicons/update-a-lexicon-entry

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/lexicons/{lexiconId}/entries/{entryId}
Updates a [Lexicon entry](https://docs.cognigy.com/ai/platform-features/nlu/slots/user-defined/lexicon).



# Update a Lexicon Slot
Source: https://docs.cognigy.com/api-reference/lexicons/update-a-lexicon-slot

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/lexicons/{lexiconId}/slots/{slotId}
Updates a [Lexicon Slot](https://docs.cognigy.com/ai/platform-features/nlu/slots/user-defined/lexicon) in a Lexicon.



# Get the Default Live Agent Inbox
Source: https://docs.cognigy.com/api-reference/live-agent/get-the-default-live-agent-inbox

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/projects/{projectId}/inbox
Gets the default Live Agent Inbox from a Project.



# Get the Live Agent Account
Source: https://docs.cognigy.com/api-reference/live-agent/get-the-live-agent-account

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/organisations/liveagentaccount
Gets the Live Agent Account for the organization.



# Set Up a Live Agent Inbox
Source: https://docs.cognigy.com/api-reference/live-agent/set-up-a-live-agent-inbox

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/projects/{projectId}/setupinbox
Sets up a default Live Agent Inbox for a Project.



# Set Up Live Agent
Source: https://docs.cognigy.com/api-reference/live-agent/set-up-live-agent

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/organisations/setupliveagent
Sets up Cognigy Live Agent for the organization.



# Update a Live Agent Inbox
Source: https://docs.cognigy.com/api-reference/live-agent/update-a-live-agent-inbox

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/projects/{projectId}/setupinbox
Updates the default Live Agent Inbox assigned to a Project.



# Create a Locale
Source: https://docs.cognigy.com/api-reference/locales/create-a-locale

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/locales
Creates a [Locale](https://docs.cognigy.com/ai/platform-features/translation-and-localization/localization) in a Project.



# Delete a Locale
Source: https://docs.cognigy.com/api-reference/locales/delete-a-locale

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/locales/{localeId}
Deletes a non-primary [Locale](https://docs.cognigy.com/ai/platform-features/translation-and-localization/localization) in a Project.



# Get a Locale
Source: https://docs.cognigy.com/api-reference/locales/get-a-locale

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/locales/{localeId}
Gets a [Locale](https://docs.cognigy.com/ai/platform-features/translation-and-localization/localization).



# Get Locales
Source: https://docs.cognigy.com/api-reference/locales/get-locales

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/locales
Gets [Locales](https://docs.cognigy.com/ai/resources/manage/localization/) in a Project. By default, 25 Locales are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Update a Locale
Source: https://docs.cognigy.com/api-reference/locales/update-a-locale

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/locales/{localeId}
Updates a [Locale](https://docs.cognigy.com/ai/platform-features/translation-and-localization/localization).



# Create a System Message
Source: https://docs.cognigy.com/api-reference/management-ui/create-a-system-message

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /management/v2.0/system/createsystemmessage
Creates a system message or overwrites the existing one.



# Create a User
Source: https://docs.cognigy.com/api-reference/management-ui/create-a-user

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /management/v2.0/users
Creates a user in Management UI.



# Create an API Key
Source: https://docs.cognigy.com/api-reference/management-ui/create-an-api-key

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /management/v2.0/organisations/{organisationId}/apikeys
Creates an API Key with admin permissions. The API Key is valid for 15 minutes. To activate this feature, set the FEATURE_USE_SUPERAPIKEY_API to "true".




# Create an Organization
Source: https://docs.cognigy.com/api-reference/management-ui/create-an-organization

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /management/v2.0/organisations
Creates an organization in Management UI.



# Delete a User
Source: https://docs.cognigy.com/api-reference/management-ui/delete-a-user

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /management/v2.0/users/{userId}
Deletes a user from Management UI.



# Delete an Organization
Source: https://docs.cognigy.com/api-reference/management-ui/delete-an-organization

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /management/v2.0/organisations/{organisationId}
Deletes an organization and the associated data.



# Enforce Password Policy
Source: https://docs.cognigy.com/api-reference/management-ui/enforce-password-policy

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /management/v2.0/organisations/{organisationId}/enforcepasswordpolicy
Enforces a password policy and requests all users to reset their passwords.



# Get a User
Source: https://docs.cognigy.com/api-reference/management-ui/get-a-user

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /management/v2.0/users/{userId}
Gets information about a user.



# Get an Organization
Source: https://docs.cognigy.com/api-reference/management-ui/get-an-organization

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /management/v2.0/organisations/{organisationId}
Gets information about an organization.



# Get Billing Information
Source: https://docs.cognigy.com/api-reference/management-ui/get-billing-information

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /management/v2.0/billing
Gets the billing information of organizations from Management UI. By default, 25 organizations are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Get Billing Information
Source: https://docs.cognigy.com/api-reference/management-ui/get-billing-information-1

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /management/v3.0/billing
Gets the billing information of organizations from Management UI. By default, 25 organizations are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Get Organizations
Source: https://docs.cognigy.com/api-reference/management-ui/get-organizations

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /management/v2.0/organisations
Gets organizations from Management UI. By default, 25 organizations are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Get the License via Management UI
Source: https://docs.cognigy.com/api-reference/management-ui/get-the-license-via-management-ui

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /management/v2.0/system/license
Returns the current license state along with company, expiration, signed date and features. Authenticated via Management UI credentials.



# Get Users
Source: https://docs.cognigy.com/api-reference/management-ui/get-users

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /management/v2.0/users
Gets users of the organization. By default, 25 users are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Get Vendor Information
Source: https://docs.cognigy.com/api-reference/management-ui/get-vendor-information

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /management/v2.0/vendor
Gets information about a vendor from Management UI.



# Impersonate a User
Source: https://docs.cognigy.com/api-reference/management-ui/impersonate-a-user

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /management/v2.0/users/{userId}/impersonate
Creates an access and refresh token pair to log in as a user.



# Set the License via Management UI
Source: https://docs.cognigy.com/api-reference/management-ui/set-the-license-via-management-ui

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /management/v2.0/system/license
Sets the system license, updating license details such as access and quotas. Authenticated via Management UI credentials.



# Update a User
Source: https://docs.cognigy.com/api-reference/management-ui/update-a-user

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /management/v2.0/users/{userId}
Updates information of a user.



# Update an Organization
Source: https://docs.cognigy.com/api-reference/management-ui/update-an-organization

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /management/v2.0/organisations/{organisationId}
Updates information of an organization.



# Create an NLU Connector
Source: https://docs.cognigy.com/api-reference/nlu-connectors/create-an-nlu-connector

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/nluconnectors
Creates an [NLU Connector](https://docs.cognigy.com/ai/platform-features/nlu/external/nlu-connectors/overview#nlu-connectors) in a Project.



# Delete an NLU Connector
Source: https://docs.cognigy.com/api-reference/nlu-connectors/delete-an-nlu-connector

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/nluconnectors/{nluConnectorId}
Deletes an NLU Connector.



# Get an NLU Connector
Source: https://docs.cognigy.com/api-reference/nlu-connectors/get-an-nlu-connector

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/nluconnectors/{nluConnectorId}
Gets information about an NLU Connector in a Project.



# Get NLU Connectors
Source: https://docs.cognigy.com/api-reference/nlu-connectors/get-nlu-connectors

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/nluconnectors
Gets NLU Connectors in a Project. By default, 25 NLU Connectors are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Perform Batch Operations
Source: https://docs.cognigy.com/api-reference/nlu-connectors/perform-batch-operations

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/nluconnectors
Performs batch operations on [NLU Connectors](https://docs.cognigy.com/ai/platform-features/nlu/external/nlu-connectors/overview#nlu-connectors).



# Update an NLU Connector
Source: https://docs.cognigy.com/api-reference/nlu-connectors/update-an-nlu-connector

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/nluconnectors/{nluConnectorId}
Updates an NLU Connector in a Project.



# Add a New Locale to a Node
Source: https://docs.cognigy.com/api-reference/nodes/add-a-new-locale-to-a-node

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/flows/{flowId}/chart/nodes/{nodeId}/addlocalization
Adds a new Locale to a Node in a Flow. Use `localeId` to specify the target Locale and `inheritFromLocaleId` to inherit content from an existing Locale.



# Create a Node
Source: https://docs.cognigy.com/api-reference/nodes/create-a-node

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/flows/{flowId}/chart/nodes
Creates a new Node in a Flow.



# Delete a Node
Source: https://docs.cognigy.com/api-reference/nodes/delete-a-node

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/flows/{flowId}/chart/nodes/{nodeId}
Delete a Node from a Flow.



# Generate Output for a Node
Source: https://docs.cognigy.com/api-reference/nodes/generate-output-for-a-node

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/flows/{flowId}/chart/nodes/output/generate
Generates content for a Node using generative AI. The generated output, either text or an adaptive card, is returned in the response and can be copied into the Node for testing or development. No node ID is required.



# Get a Node
Source: https://docs.cognigy.com/api-reference/nodes/get-a-node

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/flows/{flowId}/chart/nodes/{nodeId}
Gets a Node from a Flow.



# Get AI Agents Nodes
Source: https://docs.cognigy.com/api-reference/nodes/get-ai-agents-nodes

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/flows/{flowId}/chart/nodes/aiagents
Gets all AI Agent Nodes from the Flow with full AI Agent details.



# Get all Node Descriptors
Source: https://docs.cognigy.com/api-reference/nodes/get-all-node-descriptors

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/flows/{flowId}/chart/descriptors
Gets all Node descriptors of a Flow. A descriptor defines a node type, its capabilities, appearance, and behavior in the Flow Chart. The descriptor acts as a blueprint for creating Nodes.



# Get Nodes
Source: https://docs.cognigy.com/api-reference/nodes/get-nodes

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/flows/{flowId}/chart/nodes
Gets Nodes from a Flow. By default, 25 Nodes are returned. Use the `limit` query parameter to change this value.



# Move a Node
Source: https://docs.cognigy.com/api-reference/nodes/move-a-node

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/flows/{flowId}/chart/nodes/{nodeId}/move
Moves a Node in a Flow Chart. Allows modifying a Node's position relative to other Nodes using modes such as append, prepend, or insert at a specific position.



# Remove a Locale from a Node
Source: https://docs.cognigy.com/api-reference/nodes/remove-a-locale-from-a-node

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/flows/{flowId}/chart/nodes/{nodeId}/removelocalization
Removes a Locale from a Node in a Flow.



# Search for Nodes
Source: https://docs.cognigy.com/api-reference/nodes/search-for-nodes

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/flows/{flowId}/chart/nodes/search
Searches for Nodes in a flow using a given filter word. Fill in the `filter` parameter to search for Nodes.



# Update a Node
Source: https://docs.cognigy.com/api-reference/nodes/update-a-node

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/flows/{flowId}/chart/nodes/{nodeId}
Updates a Node in a Flow. Allows modifying the Node's type, label, comment, Entrypoint status, and configuration properties.



# Resolve Options for a Field
Source: https://docs.cognigy.com/api-reference/options-resolver/resolve-options-for-a-field

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/optionsresolver
Resolves the selected option of a field in an Extension Node editor. This endpoint is accessible only in dedicated SaaS and on-premises installations.



# Create a Download Link
Source: https://docs.cognigy.com/api-reference/packages/create-a-download-link

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/packages/{packageId}/downloadlink
Creates a link to download a [Package](https://docs.cognigy.com/ai/platform-features/packages).



# Create a Package
Source: https://docs.cognigy.com/api-reference/packages/create-a-package

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/packages
Creates a task to create a [Package](https://docs.cognigy.com/ai/resources/build/packages).



# Delete a Package
Source: https://docs.cognigy.com/api-reference/packages/delete-a-package

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/packages/{packageId}
Creates a task to delete a [Package](https://docs.cognigy.com/ai/resources/build/packages).



# Get a Package
Source: https://docs.cognigy.com/api-reference/packages/get-a-package

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/packages/{packageId}
Gets information about a [Package](https://docs.cognigy.com/ai/resources/build/packages).



# Get Packages
Source: https://docs.cognigy.com/api-reference/packages/get-packages

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/packages
Gets Packages in a Project. By default, 25 Packages are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Merge a Package
Source: https://docs.cognigy.com/api-reference/packages/merge-a-package

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/packages/{packageId}/merge
Creates a task to merge a [Package](https://docs.cognigy.com/ai/resources/build/packages).



# Upload a Package
Source: https://docs.cognigy.com/api-reference/packages/upload-a-package

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/packages/upload
Uploads a [Package](https://docs.cognigy.com/ai/resources/build/packages) and creates a task to unpack it in the Project.



# Change Step Order
Source: https://docs.cognigy.com/api-reference/playbooks/change-step-order

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/playbooks/{playbookId}/steps/order
Changes the order of [Playbook steps](https://docs.cognigy.com/ai/agents/test/playbooks).



# Create a Playbook
Source: https://docs.cognigy.com/api-reference/playbooks/create-a-playbook

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/playbooks
Creates a [Playbook](https://docs.cognigy.com/ai/agents/build/playbooks) in a Project.



# Create a Playbook Step
Source: https://docs.cognigy.com/api-reference/playbooks/create-a-playbook-step

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/playbooks/{playbookId}/steps
Creates a step in a [Playbook](https://docs.cognigy.com/ai/agents/test/playbooks).



# Create an Assertion
Source: https://docs.cognigy.com/api-reference/playbooks/create-an-assertion

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/playbooks/{playbookId}/steps/{stepId}/asserts
Creates an [assertion](https://docs.cognigy.com/ai/agents/test/assertions) for a Playbook step.



# Delete a Playbook
Source: https://docs.cognigy.com/api-reference/playbooks/delete-a-playbook

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/playbooks/{playbookId}
Deletes a [Playbook](https://docs.cognigy.com/ai/agents/test/playbooks) from a Project.



# Delete a Playbook Run
Source: https://docs.cognigy.com/api-reference/playbooks/delete-a-playbook-run

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/playbooks/{playbookId}/runs/{playbookRunId}
Deletes a [Playbook run](https://docs.cognigy.com/ai/agents/test/playbooks) from a Project.



# Delete a Playbook Step
Source: https://docs.cognigy.com/api-reference/playbooks/delete-a-playbook-step

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/playbooks/{playbookId}/steps/{stepId}
Deletes a step from a [Playbook](https://docs.cognigy.com/ai/agents/test/playbooks).



# Delete a Playbook Step Assertion
Source: https://docs.cognigy.com/api-reference/playbooks/delete-a-playbook-step-assertion

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/playbooks/{playbookId}/steps/{stepId}/asserts/{assertId}
Deletes an assertion from a [Playbook step](https://docs.cognigy.com/ai/agents/test/playbooks).



# Get a Playbook
Source: https://docs.cognigy.com/api-reference/playbooks/get-a-playbook

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/playbooks/{playbookId}
Gets data about a [Playbook](https://docs.cognigy.com/ai/agents/test/playbooks).



# Get Playbooks
Source: https://docs.cognigy.com/api-reference/playbooks/get-playbooks

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/playbooks
Gets [Playbooks](https://docs.cognigy.com/ai/agents/test/playbooks) in a Project. By default, 25 Playbooks are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# List Playbook Runs
Source: https://docs.cognigy.com/api-reference/playbooks/list-playbook-runs

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/playbooks/{playbookId}/runs
Lists runs from a Playbook.



# Perform Batch Operations
Source: https://docs.cognigy.com/api-reference/playbooks/perform-batch-operations

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/playbooks
Performs batch operations on Playbooks.



# Retrieve a Playbook Run
Source: https://docs.cognigy.com/api-reference/playbooks/retrieve-a-playbook-run

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/playbooks/{playbookId}/runs/{playbookRunId}
Retrieves data about a [Playbook run](https://docs.cognigy.com/ai/agents/test/playbooks).



# Schedule a Playbook Run
Source: https://docs.cognigy.com/api-reference/playbooks/schedule-a-playbook-run

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/playbooks/{playbookId}/schedule
Schedules a Playbook run.



# Update a Playbook
Source: https://docs.cognigy.com/api-reference/playbooks/update-a-playbook

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/playbooks/{playbookId}
Updates Playbook data in a Project.



# Update a Playbook Step
Source: https://docs.cognigy.com/api-reference/playbooks/update-a-playbook-step

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/playbooks/{playbookId}/steps/{stepId}
Updates a step in a [Playbook](https://docs.cognigy.com/ai/agents/test/playbooks).



# Update a Playbook Step Assertion
Source: https://docs.cognigy.com/api-reference/playbooks/update-a-playbook-step-assertion

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/playbooks/{playbookId}/steps/{stepId}/asserts/{assertId}
Updates an [assertion](https://docs.cognigy.com/ai/agents/test/assertions) from a [Playbook step](https://docs.cognigy.com/ai/agents/test/playbooks).



# Create a Project
Source: https://docs.cognigy.com/api-reference/projects/create-a-project

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/projects
Creates a [Project](https://docs.cognigy.com/ai/agents/develop/projects-and-flows/projects).



# Create a Project from a Template
Source: https://docs.cognigy.com/api-reference/projects/create-a-project-from-a-template

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/projects/createbytemplate
Creates a [Project](https://docs.cognigy.com/ai/agents/develop/projects-and-flows/projects#projects) from a template.



# Delete a Project
Source: https://docs.cognigy.com/api-reference/projects/delete-a-project

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/projects/{projectId}
Deletes a [Project](https://docs.cognigy.com/ai/agents/develop/projects-and-flows/projects#projects).



# Generate NLU Scores
Source: https://docs.cognigy.com/api-reference/projects/generate-nlu-scores

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/projects/{projectId}/nlu/scores
Generate a NLU score for an sentence by analyzing Intents in a Flow and Locale.



# Get a Graph
Source: https://docs.cognigy.com/api-reference/projects/get-a-graph

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/projects/{projectId}/graph
Gets a graph showing all resources of a Project, Snapshot or Package.



# Get a Project
Source: https://docs.cognigy.com/api-reference/projects/get-a-project

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/projects/{projectId}
Gets data from a [Project](https://docs.cognigy.com/ai/agents/develop/projects-and-flows/projects#projects).



# Get Project Settings
Source: https://docs.cognigy.com/api-reference/projects/get-project-settings

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/projects/{projectId}/settings
Gets the settings of a [Project](https://docs.cognigy.com/ai/agents/develop/projects-and-flows/projects).



# Get Projects
Source: https://docs.cognigy.com/api-reference/projects/get-projects

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/projects
Gets [Projects](https://docs.cognigy.com/ai/resources/agents/agents/) in an organization. By default, 25 Projects are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Set Up Cognigy Generative AI
Source: https://docs.cognigy.com/api-reference/projects/set-up-cognigy-generative-ai

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/projects/{projectId}/settings/setupgenerativeai
Sets up Cognigy Generative AI for a Project.



# Set Up Proactive Follow-Up
Source: https://docs.cognigy.com/api-reference/projects/set-up-proactive-follow-up

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/projects/{projectId}/settings/proactive/setup
Provisions a Proactive tenant, creates a journey, and generates a follow-up flow.



# Test a Translation Provider Connection
Source: https://docs.cognigy.com/api-reference/projects/test-a-translation-provider-connection

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/projects/{projectId}/settings/translation/test
Validates the connection with a translation provider.



# Train Flow Models
Source: https://docs.cognigy.com/api-reference/projects/train-flow-models

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/projects/{projectId}/trainallflows
Trains the NLU model of all Flows in a [Project](https://docs.cognigy.com/ai/agents/develop/projects-and-flows/projects#projects).



# Update a Project
Source: https://docs.cognigy.com/api-reference/projects/update-a-project

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/projects/{projectId}
Updates the data of a [Project](https://docs.cognigy.com/ai/agents/develop/projects-and-flows/projects#projects).



# Update Project Settings
Source: https://docs.cognigy.com/api-reference/projects/update-project-settings

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/projects/{projectId}/settings
Updates the settings of a [Project](https://docs.cognigy.com/ai/agents/develop/projects-and-flows/projects).



# Validate a Project Name
Source: https://docs.cognigy.com/api-reference/projects/validate-a-project-name

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/projects/validateprojectname
Checks whether Project name is already used in the organization.



# Search Resources
Source: https://docs.cognigy.com/api-reference/search/search-resources

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/search
Searches resources across the organization and list the first 25 results by default. This endpoint searches only for Extensions, Projects, Lexicons, Endpoints, Flows, Functions, Playbooks, NLU Connectors, Snapshots, and Simulations. To change the default number of listed results, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Inject a Context
Source: https://docs.cognigy.com/api-reference/sessions/inject-a-context

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/sessions/{sessionId}/context/inject
Injects a context into a session.



# Reset a Context
Source: https://docs.cognigy.com/api-reference/sessions/reset-a-context

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/sessions/{sessionId}/context/reset
Resets a context in a session.



# Create a Slot Filler
Source: https://docs.cognigy.com/api-reference/slot-fillers/create-a-slot-filler

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/flows/{flowId}/slotfillers
Creates a Slot Filler in the Project.



# Delete a Slot Filler
Source: https://docs.cognigy.com/api-reference/slot-fillers/delete-a-slot-filler

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/flows/{flowId}/slotfillers/{slotFillerId}
Deletes a Slot Filler from the Flow.



# Get a Slot Filler
Source: https://docs.cognigy.com/api-reference/slot-fillers/get-a-slot-filler

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/flows/{flowId}/slotfillers/{slotFillerId}
Gets a Slot Filler from a Flow.



# Get Slot Fillers
Source: https://docs.cognigy.com/api-reference/slot-fillers/get-slot-fillers

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/flows/{flowId}/slotfillers
Gets Slot Fillers from a Flow. By default, 25 Slot Fillers are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Perform Batch Operations
Source: https://docs.cognigy.com/api-reference/slot-fillers/perform-batch-operations

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/flows/{flowId}/slotfillers
Performs batch operations on Slot Fillers.



# Update a Slot Filler
Source: https://docs.cognigy.com/api-reference/slot-fillers/update-a-slot-filler

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/flows/{flowId}/slotfillers/{slotFillerId}
Updates a Slot Filler from a Flow.



# Create a Download Link
Source: https://docs.cognigy.com/api-reference/snapshots/create-a-download-link

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/snapshots/{snapshotId}/downloadlink
Creates a link to download a [Snapshot](https://docs.cognigy.com/ai/agents/deploy/snapshots).



# Create a Snapshot
Source: https://docs.cognigy.com/api-reference/snapshots/create-a-snapshot

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/snapshots
Creates a task to create a [Snapshot](https://docs.cognigy.com/ai/agents/deploy/snapshots) in a Project.



# Delete a Snapshot
Source: https://docs.cognigy.com/api-reference/snapshots/delete-a-snapshot

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/snapshots/{snapshotId}
Creates a task to delete a [Snapshot](https://docs.cognigy.com/ai/agents/deploy/snapshots).



# Get Resources in a Snapshot
Source: https://docs.cognigy.com/api-reference/snapshots/get-resources-in-a-snapshot

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/snapshots/{snapshotId}/resources
Gets resources of a specific type in a [Snapshot](https://docs.cognigy.com/ai/agents/deploy/snapshots). By default, 25 resources are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Get Snapshots
Source: https://docs.cognigy.com/api-reference/snapshots/get-snapshots

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/snapshots
Gets [Snapshots](https://docs.cognigy.com/ai/agents/deploy/snapshots) in a Project. By default, 25 Snapshots are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Gets a Snapshot
Source: https://docs.cognigy.com/api-reference/snapshots/gets-a-snapshot

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/snapshots/{snapshotId}
Gets information about a [Snapshot](https://docs.cognigy.com/ai/agents/deploy/snapshots) in a Project.



# Package a Snapshot
Source: https://docs.cognigy.com/api-reference/snapshots/package-a-snapshot

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/snapshots/{snapshotId}/package
Creates a task to package a [Snapshot](https://docs.cognigy.com/ai/agents/deploy/snapshots) in a Project.



# Restore a Snapshot
Source: https://docs.cognigy.com/api-reference/snapshots/restore-a-snapshot

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/snapshots/{snapshotId}/restore
Creates a task to restore a [Snapshot](https://docs.cognigy.com/ai/agents/deploy/snapshots).



# Upload a Snapshot Package
Source: https://docs.cognigy.com/api-reference/snapshots/upload-a-snapshot-package

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/snapshots/upload
Uploads a [Snapshot](https://docs.cognigy.com/ai/agents/deploy/snapshots) package and creates a task to unpack it in the Project.



# Cancel a Task
Source: https://docs.cognigy.com/api-reference/tasks/cancel-a-task

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/tasks/{taskId}/cancel
Cancels a [task](https://docs.cognigy.com/ai/agents/develop/projects-and-flows/projects#task-panel) in a Project.



# Get a Task
Source: https://docs.cognigy.com/api-reference/tasks/get-a-task

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/tasks/{taskId}
Gets a [task](https://docs.cognigy.com/ai/agents/develop/projects-and-flows/projects#task-panel) from a Project.



# Get Tasks
Source: https://docs.cognigy.com/api-reference/tasks/get-tasks

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/tasks
Gets [tasks](https://docs.cognigy.com/ai/agents/develop/projects-and-flows/projects#task-panel) from a Project. By default, 25 tasks are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Create a Token
Source: https://docs.cognigy.com/api-reference/tokens/create-a-token

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/projects/{projectId}/snippets
Creates a [Token](https://docs.cognigy.com/ai/platform-features/tokens) in a Project.



# Delete a Token
Source: https://docs.cognigy.com/api-reference/tokens/delete-a-token

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/projects/{projectId}/snippets/{snippetId}
Deletes a [Token](https://docs.cognigy.com/ai/platform-features/tokens) from a Project.



# Get Tokens
Source: https://docs.cognigy.com/api-reference/tokens/get-tokens

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/projects/{projectId}/snippets
Gets custom [Tokens](https://docs.cognigy.com/ai/platform-features/tokens) in a Project. By default, 25 Tokens are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results. Default Tokens aren't included in the results.



# Create an API Key
Source: https://docs.cognigy.com/api-reference/users-21/create-an-api-key

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.1/users/me/apikeys
Creates an [API key](https://docs.cognigy.com/ai/tools/user-menu/my-profile/#api-keys) for the current user.



# Get API Keys of the Current User
Source: https://docs.cognigy.com/api-reference/users-21/get-api-keys-of-the-current-user

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.1/users/me/apikeys
Gets API keys of the current user. By default, 25 API keys are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Accept the Terms of Service
Source: https://docs.cognigy.com/api-reference/users/accept-the-terms-of-service

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/users/me/accepttermsofservice
Accepts the terms of service.



# Add a Project Member
Source: https://docs.cognigy.com/api-reference/users/add-a-project-member

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/projects/{projectId}/addprojectmember
Adds a user to a Project as a member.



# Assign a Global Role
Source: https://docs.cognigy.com/api-reference/users/assign-a-global-role

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/users/{userId}/addroletouser
Assigns a global role to a user.



# Assign a Project Role
Source: https://docs.cognigy.com/api-reference/users/assign-a-project-role

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/projects/{projectId}/members/{userId}/addroletomember
Assigns a project role to a user.



# Assign a Project to a User
Source: https://docs.cognigy.com/api-reference/users/assign-a-project-to-a-user

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/users/{userId}/addproject
Gives a user access to a Project.



# Change a User Password
Source: https://docs.cognigy.com/api-reference/users/change-a-user-password

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/users/changepassword
Changes the password of a user.



# Create a User
Source: https://docs.cognigy.com/api-reference/users/create-a-user

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/users
Creates a [user](https://docs.cognigy.com/ai/administer/access/management-ui#create-a-user) in the installation.



# Create an API Key
Source: https://docs.cognigy.com/api-reference/users/create-an-api-key

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/users/me/apikeys
Creates an [API key](https://docs.cognigy.com/ai/administer/user-menu/my-profile#api-keys) for the current user.



# Delete a User
Source: https://docs.cognigy.com/api-reference/users/delete-a-user

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/users/{userId}
Deletes a user from the system completely.



# Delete an API Key
Source: https://docs.cognigy.com/api-reference/users/delete-an-api-key

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/users/me/apikeys/{apiKeyId}
Deletes an [API key](https://docs.cognigy.com/ai/administer/user-menu/my-profile#api-keys) for the current user.



# Delete Own User
Source: https://docs.cognigy.com/api-reference/users/delete-own-user

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/users/me
Deletes the current [user](https://docs.cognigy.com/ai/administer/access/management-ui#other-operations) from the installation.



# Deprecate User Passwords
Source: https://docs.cognigy.com/api-reference/users/deprecate-user-passwords

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/users/deprecatepassword
Deprecates the password of a single user or all users of the organization.



# Get a Project Member
Source: https://docs.cognigy.com/api-reference/users/get-a-project-member

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/projects/{projectId}/members/{userId}
Gets all information about a member of a Project.



# Get a User
Source: https://docs.cognigy.com/api-reference/users/get-a-user

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/users/{userId}
Gets all the information from a user.



# Get API Keys of the Current User
Source: https://docs.cognigy.com/api-reference/users/get-api-keys-of-the-current-user

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/users/me/apikeys
Gets API keys of the current user. By default, 25 API keys are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Get Current User's ACL
Source: https://docs.cognigy.com/api-reference/users/get-current-users-acl

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/users/me/acl
Gets the access control list (ACL) information for the current user's profile.



# Get Login Attempts of the Current User
Source: https://docs.cognigy.com/api-reference/users/get-login-attempts-of-the-current-user

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/users/me/loginattempts
Gets login attempts for the current user. By default, 25 login attempts are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Get Pinned Resources
Source: https://docs.cognigy.com/api-reference/users/get-pinned-resources

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/users/me/pinned
Gets pinned Cognigy.AI resources by resource type.



# Get Project Members
Source: https://docs.cognigy.com/api-reference/users/get-project-members

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/projects/{projectId}/members
Gets [members](https://docs.cognigy.com/ai/resources/manage/members/) of a Project. By default, 25 Project members are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Get the Last Login Attempt
Source: https://docs.cognigy.com/api-reference/users/get-the-last-login-attempt

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/users/me/lastloginattempt
Gets the current user's last login attempt.



# Get Users
Source: https://docs.cognigy.com/api-reference/users/get-users

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/users
Gets [users](https://docs.cognigy.com/voicegateway/webapp/users) in the organization. By default, 25 users are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Gets the Current User's Profile
Source: https://docs.cognigy.com/api-reference/users/gets-the-current-users-profile

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/users/me
Gets the current user's profile information.



# Log Out a User
Source: https://docs.cognigy.com/api-reference/users/log-out-a-user

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/users/{userId}/logout
Logs out a user and removes the access/refresh token pair.



# Pin or Unpin Resources
Source: https://docs.cognigy.com/api-reference/users/pin-or-unpin-resources

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/users/me/pin
Pins or unpins Cognigy.AI resources, such as Projects, to organize them based on your preferences.



# Remove a Project Member
Source: https://docs.cognigy.com/api-reference/users/remove-a-project-member

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0//projects/{projectId}/removeprojectmember
Removes a user from a Project.



# Request a Password Reset
Source: https://docs.cognigy.com/api-reference/users/request-a-password-reset

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/users/requestpasswordreset
Requests a password reset email.



# Reset a User Password
Source: https://docs.cognigy.com/api-reference/users/reset-a-user-password

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/users/resetpassword
Resets a user's password using a provided token.



# Reset Failed Login Attempts
Source: https://docs.cognigy.com/api-reference/users/reset-failed-login-attempts

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/users/{userId}/resetfailedloginattempts
Resets the failed login attempts for a user. Requires the 'admin' role.



# Unassign a Global Role
Source: https://docs.cognigy.com/api-reference/users/unassign-a-global-role

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/users/{userId}/removerolefromuser
Unassigns a global role from a user.



# Unassign a Project from a User
Source: https://docs.cognigy.com/api-reference/users/unassign-a-project-from-a-user

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/users/{userId}/removeproject
Removes the access to a Project from a user.



# Unassign a Project Role
Source: https://docs.cognigy.com/api-reference/users/unassign-a-project-role

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/projects/{projectId}/members/{userId}/removerolefrommember
Unassigns a Project role from a Project member.



# Update a Project Member
Source: https://docs.cognigy.com/api-reference/users/update-a-project-member

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/projects/{projectId}/members/{userId}
Updates information of a Project member.



# Update a User Profile
Source: https://docs.cognigy.com/api-reference/users/update-a-user-profile

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/users/{userId}
Updates information of a user profile.



# Update Own User
Source: https://docs.cognigy.com/api-reference/users/update-own-user

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/users/me
Updates the profile of the current [user](doc:users).



# Validate a Password Reset Token
Source: https://docs.cognigy.com/api-reference/users/validate-a-password-reset-token

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/users/validatepasswordresettoken
Checks if the user password reset token is valid.



# Get Voice Gateway Account
Source: https://docs.cognigy.com/api-reference/voice-gateway/get-voice-gateway-account

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/organisations/voicegateway/account
Gets information from the Voice Gateway account.



# Get Voice Gateway Speech Credentials
Source: https://docs.cognigy.com/api-reference/voice-gateway/get-voice-gateway-speech-credentials

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/voicegateway/speechcredentials
Gets speech credentials associated with the Voice Gateway account for the organisation.



# Get Voice Gateway Supported Voices
Source: https://docs.cognigy.com/api-reference/voice-gateway/get-voice-gateway-supported-voices

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/voicegateway/speechcredentials/voices
Gets supported languages and voices for a given speech credential.



# Set Up a Voice Gateway Account
Source: https://docs.cognigy.com/api-reference/voice-gateway/set-up-a-voice-gateway-account

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/organisations/voicegateway/account
Sets up a Voice Gateway account for the organization.



# Get Voices
Source: https://docs.cognigy.com/api-reference/voice-preview/get-voices

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/audio/preview/languages
Gets available preview voices for the speech provider selected in the Project settings.



# Validate Speech Provider
Source: https://docs.cognigy.com/api-reference/voice-preview/validate-speech-provider

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/audio/voiceprovider/test
Validates the credentials of a speech provider connection.



# Get User or Organization Info via Webfinger for SSO
Source: https://docs.cognigy.com/api-reference/webfinger/get-user-or-organization-info-via-webfinger-for-sso

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /.well-known/webfinger
Gets data from an Identity Provider (IdP) configuration for SSO authentication based on a user email or organization ID via the [WebFinger protocol](https://en.wikipedia.org/wiki/WebFinger).



# Create a Sentence
Source: https://docs.cognigy.com/api-reference/yesno-intents/create-a-sentence

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/locales/{localeId}/yesnointents/{intentId}/sentences
Creates a Sentence for a [Yes/No Intent](docs.cognigy.com/ai/platform-features/nlu/intents/yes-no-intents).



# Delete a Sentence
Source: https://docs.cognigy.com/api-reference/yesno-intents/delete-a-sentence

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/locales/{localeId}/yesnointents/{intentId}/sentences/{sentenceId}
Deletes a Sentence from a [Yes/No Intent](docs.cognigy.com/ai/platform-features/nlu/intents/yes-no-intents).



# Delete a Yes/No Intent
Source: https://docs.cognigy.com/api-reference/yesno-intents/delete-a-yesno-intent

https://api-trial.cognigy.ai/openapi/openapi-viewer.json delete /v2.0/locales/{localeId}/yesnointents/{intentId}
Deletes a Yes/No Intent. Once deleted, you can't recreate the Yes/No Intent. You need a new Locale to have Yes/No Intents.



# Get a Sentence
Source: https://docs.cognigy.com/api-reference/yesno-intents/get-a-sentence

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/locales/{localeId}/yesnointents/{intentId}/sentences/{sentenceId}
Gets a Sentence from a Yes/No Intent.



# Get Sentences
Source: https://docs.cognigy.com/api-reference/yesno-intents/get-sentences

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/locales/{localeId}/yesnointents/{intentId}/sentences
Gets sentences from a [Yes/No Intent](docs.cognigy.com/ai/platform-features/nlu/intents/yes-no-intents). By default, 25 sentences are returned. To change this value, use the `limit` query parameter. Additionally, you can filter, sort, and paginate the results.



# Get Yes/No Intents
Source: https://docs.cognigy.com/api-reference/yesno-intents/get-yesno-intents

https://api-trial.cognigy.ai/openapi/openapi-viewer.json get /v2.0/locales/{localeId}/yesnointents
Gets the [Yes/No Intents](docs.cognigy.com/ai/platform-features/nlu/intents/yes-no-intents) of a Locale.



# Train Yes/No Intents
Source: https://docs.cognigy.com/api-reference/yesno-intents/train-yesno-intents

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/locales/{localeId}/yesnointents/train
Creates a task to train the [Yes/No Intents](https://docs.cognigy.com/ai/nlu/nlu-overview/yes-no-intents/) of a Locale.



# Train Yes/No Intents for All Locales
Source: https://docs.cognigy.com/api-reference/yesno-intents/train-yesno-intents-for-all-locales

https://api-trial.cognigy.ai/openapi/openapi-viewer.json post /v2.0/projects/{projectId}/yesnointents/train
Creates tasks to train the [Yes/No Intents](https://docs.cognigy.com/ai/platform-features/nlu/intents/yes-no-intents#yes-no-intents) for all Locales in a Project.



# Update a Sentence
Source: https://docs.cognigy.com/api-reference/yesno-intents/update-a-sentence

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/locales/{localeId}/yesnointents/{intentId}/sentences/{sentenceId}
Updates a Sentence in a [Yes/No Intent](docs.cognigy.com/ai/platform-features/nlu/intents/yes-no-intents).



# Update a Yes/No Intent
Source: https://docs.cognigy.com/api-reference/yesno-intents/update-a-yesno-intent

https://api-trial.cognigy.ai/openapi/openapi-viewer.json patch /v2.0/locales/{localeId}/yesnointents/{intentId}
Updates a Yes/No Intent of a Locale.



# Configure the Click To Call Widget
Source: https://docs.cognigy.com/click-to-call/configure



To deploy a Click To Call widget, you need to create a [Voice Gateway Endpoint](/ai/agents/deploy/endpoint-reference/voice-gateway) and set up the Click To Call widget in the Endpoint settings. The default setup includes only voice communication and live transcription. For a more advanced setup, read the [Tutorials](/click-to-call/tutorials/overview) section.

## Prerequisites

* Configure a [speech provider](/ai/agents/test/voice-preview#set-up-a-speech-provider) in the [Voice Preview Settings](/ai/administer/access/project-settings#voice-preview-settings) section of the [Project settings](/ai/administer/access/project-settings).
* Access to [Voice Gateway](/voice-gateway/getting-started#set-up-voice-gateway-for-your-organization).
* Configure a [Flow](/ai/agents/develop/projects-and-flows/overview).
* For testing purposes, allow microphone access in your browser and, if needed, in your device settings.

## Restrictions

* Click To Call currently doesn't support DTMF input by default.
* Click To Call supports only the Dial transfer type for [call transfers](#transfer-calls).

## Configure a Click To Call Widget

<Note>
  When you generate the embedding code for the Click To Call widget in the Voice Gateway Endpoint settings, Cognigy automatically creates an [Application](/voice-gateway/webapp/applications) in the Voice Gateway Self-Service Portal. You can change the voice settings in [the Application or Flow editor](#voice-configuration).
</Note>

<AccordionGroup>
  <Accordion title="Configure the Widget">
    To embed a Click To Call widget into your website, follow these steps:

    1. In **Deploy > Endpoints**, click **+ New Endpoint** and select **Voice Gateway**.

    2. In the **New Endpoint** section, configure the following:
       * **Name** — enter a unique name.
       * **Flow** — select a Flow you previously created.

    3. Click the **Set Up Click To Call Integration** button inside the **Click To Call Embedding HTML** field. The embedding code for the Click To Call widget is generated in a code editor. This step generates an Application in the Voice Gateway Self-Service Portal. The Application receives a name with the following pattern `Widget - <Cognigy Project Name> - <Voice Gateway Endpoint Name>`.

    4. In the **Click To Call Widget Settings** section, configure the following:

    * **Enable Click To Call Widget** — activate the toggle.
    * **AI Agent Name** — enter the name displayed on the Click To Call widget, for example, `Package Support`.
    * **Tagline** — enter the text displayed under the AI Agent name on the widget, for example, `Get help with your package`.
    * **AI Agent Avatar Logo URL** — URL of the image to be displayed on the widget, for example, `https://www.<domain>.com/logo.png`. If empty, the Cognigy logo is used.
    * (Optional) **Theme** — select the color theme of the widget, for example, **Clean white** (default), **Dark mode**, or **AI Purple**.
    * **Display Live Transcription** — toggle to show conversation transcript in the widget.
  </Accordion>

  <Accordion title="(Optional) Configure a Privacy Notice">
    In **Click To Call Widget Config** in the Voice Gateway Endpoint settings, you can add a privacy notice with the following parameters:

    | Parameter                 | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
    | ------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | Enable Privacy Notice     | Displays a privacy notice at the beginning of each chat session to inform users about data handling practices and prompts the users to accept or decline the privacy policy. Accepting the privacy policy is required to start the conversation, ensuring compliance and user trust. If the user declines the privacy policy, the conversation doesn't start.                                                                                                                   |
    | Privacy Notice Title      | This parameter is displayed only if you toggle on **Enable Privacy Notice**. Sets the heading that appears at the top of the privacy notice screen. By default, it is set to `Privacy Notice`.                                                                                                                                                                                                                                                                                  |
    | Privacy Notice Text       | This parameter is displayed only if you toggle on **Enable Privacy Notice**. Sets the message that users see in the privacy notice prompt. This parameter supports Markdown syntax. The default text is `Please accept our privacy policy to start your chat`. You can customize this message to provide more detailed information about your data collection, usage practices, and users' rights, ensuring clarity and building trust from the outset of the chat interaction. |
    | Submit Button             | This parameter is displayed only if you toggle on **Enable Privacy Notice**. Sets the text displayed on the button users click to accept the Privacy Notice. By default, the button is labeled `Submit`.                                                                                                                                                                                                                                                                        |
    | Cancel Button Text        | This parameter is displayed only if you toggle on **Enable Privacy Notice**. Sets the text displayed on the button users click to decline the Privacy Notice. By default, the button is labeled `Cancel`.                                                                                                                                                                                                                                                                       |
    | Privacy Policy Link Title | This parameter is displayed only if you toggle on **Enable Privacy Notice**. Sets the text that will be displayed as the title for the Privacy Policy link. The link title allows you to tailor the label of the link to better match the tone and context of your brand.                                                                                                                                                                                                       |
    | Privacy Policy Link URL   | This parameter is displayed only if you toggle on **Enable Privacy Notice**. Sets the URL that directs users to your privacy policy.                                                                                                                                                                                                                                                                                                                                            |
  </Accordion>
</AccordionGroup>

After configuring the Click To Call widget in the Endpoint settings, you can test it by clicking the **Open Demo Widget** button in the upper-right corner of the Endpoint settings.

### Voice Configuration

When you [generate the Click To Call embedding code](#configure-the-widget), an application is created in the Voice Gateway Self-Service Portal. The speech provider label is automatically generated from a lowercase version of the name of the [speech provider connection you configured](#prerequisites), with spaces replaced by underscores. For example, if you named the connection `Click To Call EU`, the label is `click_to_call_eu`.

The generated Application inherits all speech providers configured in the **Voice Preview Settings** section in the [Project settings](/ai/administer/access/project-settings) in Cognigy. By default, the Application uses the speech provider configured in the Project settings for STT and TTS. For TTS, the Application uses the speech provider's default voice. You can change the voice and speech provider of your AI Agent, for example, in the AI Agent Node or [Set Session Config Node](/ai/agents/develop/node-reference/voice/voice-gateway/set-session-config), without switching the default speech provider. If you set up a new speech provider in Cognigy after the Click To Call application is created, make sure the new provider is configured in the widget Application in the Voice Gateway Self-Service Portal.

### Transfer Calls

You can transfer the call with Click To Call by adding a [Transfer Node](/ai/agents/develop/node-reference/voice/voice-gateway/transfer) to your Flow. Click To Call supports only Dial transfer calls to SIP targets.

For call transfers, Click To Call uses any carrier configured in your Voice Gateway account. If you have only one carrier, Click To Call uses the configured carrier as the default outbound carrier. If you have multiple carriers, make sure you configure the carrier for outbound calls from the **Default carrier for Click To Call calls** list on your [Account page](/voice-gateway/webapp/accounts) in the Voice Gateway Self-Service Portal.

## Next Steps

* [Embed the Click To Call Widget in your website](/click-to-call/embed)

## More Information

* [Tutorials](/click-to-call/tutorials/overview)


# Embed the Click To Call Widget in Your Website
Source: https://docs.cognigy.com/click-to-call/embed



Embedding a Click To Call widget lets you integrate the widget into your website. By connecting the widget to your Voice Gateway Endpoint and configuring the widget's appearance and behavior, you can offer users seamless interaction and support without a complex setup.

The script can be hosted remotely, for example, on GitHub, or on your own server.

## Prerequisites

* Configure a [Click To Call Widget](/click-to-call/configure).
* Edit access to the HTML content of your website.

## Embed the Click To Call Widget

<Steps>
  <Step title="Copy the code from the Endpoint settings">
    1. Go to **Deploy > Endpoints** and select the Voice Gateway Endpoint you previously configured the Click To Call widget for.
    2. In the **Click To Call Embedding HTML** field of the Endpoint settings, hover over the code editor and click the **Copy to clipboard** button to copy the HTML code. The HTML code includes the following entities:

       * The Click To Call JavaScript bundle — hosted externally on GitHub.
       * The `addEventListener()` function — waits for the page to fully load, then initializes the Click To Call widget with `initWebRTCWidget()`, if it exists. The `initWebRTCWidget()` function is called with a single argument, which is the configuration URL of the Voice Gateway Endpoint to connect to.
         This embedding code contains the necessary JavaScript code to initialize and display the Click To Call widget on your website.

       ```html theme={null}
       <script src="https://github.com/Cognigy/click-to-call-widget/releases/latest/download/webRTCWidget.js"></script>
       <script>
           addEventListener("load", (event) => {
               if (window.initWebRTCWidget) {
                   window.initWebRTCWidget(
                       "<ENDPOINT_URL>/<YOUR_ENDPOINT_CONFIG_TOKEN>"
                   )
               }
           });
       </script>
       ```

    Where `<ENDPOINT_URL>` and `<YOUR_ENDPOINT_CONFIG_TOKEN>` are the URL and the Endpoint configuration token, respectively. For example, `https://endpoint-dev.cognigy.ai/ab8b929039b427fed7ee84bb799acd7f254fc9254be27c87a78fc8a70fb048ec` for the development environment or `https://endpoint.cognigy.ai/ab8b929039b427fed7ee84bb799acd7f254fc9254be27c87a78fc8a70fb048ec` for the production environment.
  </Step>

  <Step title="Paste the code into your website">
    Paste the code you copied previously from the Endpoint settings into your page's `<body>` or before `</body>`. The code looks similar to the following:

    ```html theme={null}
    <html lang="en">
    <body>
        <script src="https://github.com/Cognigy/click-to-call-widget/releases/latest/download/webRTCWidget.js"></script>
        <script>
            addEventListener("load", (event) => {
                if (window.initWebRTCWidget) {
                    window.initWebRTCWidget('https://endpoint-dev.cognigy.ai/<URLTOKEN>')
                }
            })
        </script>
    </body>
    </html>
    ```
  </Step>
</Steps>

## Customize the Click To Call Widget

You can customize the Click To Call widget by extending the [widget script](https://github.com/Cognigy/click-to-call-widget?tab=readme-ov-file#customizing-widget-styles), according to your requirements:

* Add event handling and callbacks.
* Use API methods for sending and receiving data.
* Customize the CSS styles and branding of your Click To Call widget.

### Configure Multimodal Communication

To configure multimodal communication for your voice AI Agents, use the [Click To Call API](https://github.com/Cognigy/click-to-call-widget).

<Accordion title="Configure Multimodal Input Fields">
  To open and close input fields on your website, you need to use the following from the Click To Call widget API:

  * **`widget.on('newRTCSession', callback)`** — uses the session from the callback to attach the following events.
  * **`session.on('newInfo', callback)`** — reads the INFO body from the Flow. Use this method to open input fields. For example, add this method to listen for the `"openInputField": true` property and open the input field by sending this property in the **Data** field of a [Say](/ai/agents/develop/node-reference/basic/say), [Question](/ai/agents/develop/node-reference/basic/question), or [Optional Question](/ai/agents/develop/node-reference/basic/optional-question) Node.
  * **`session.on('ended', callback)`** — close the input field when the call ends.
  * **`session.on('terminated', callback)`** — close the input field when the session is terminated.
</Accordion>


# Click To Call
Source: https://docs.cognigy.com/click-to-call/index



<div>
  <div>
    *Cognigy Click To Call* is a multimodal voice application that allows for seamless communication between users and your voice AI Agent with a single click.

    <p>
      The main function is to create easy to use voice AI Agent widgets that you can embed in your website. Additionally, you can build a custom experience with the Click To Call SDK.
    </p>

    <p>
      Click To Call is built with the web real-time communication (WebRTC) protocol and Opus codec, and delivers low-latency, high-quality audio communication.
    </p>
  </div>

  <div>
    <video aria-label="Cognigy Click to Call hero animation showing a user speaking to a voice AI agent using the Click to Call widget.">
      <source type="video/mp4" />
    </video>

    <video aria-label="Cognigy Click to Call hero animation showing a user speaking to a voice AI agent using the Click to Call widget.">
      <source type="video/mp4" />
    </video>

    <button type="button" aria-label="Pause video">
      ⏸
    </button>
  </div>
</div>

<CardGroup>
  <Card title="About Click To Call" icon="info" href="/click-to-call/overview">
    Learn about Click To Call, deployment options, key benefits, prerequisites, and restrictions.
  </Card>

  <Card title="Quickstart" icon="rocket" href="/click-to-call/quickstart">
    Set up a basic Click To Call widget in a few steps.
  </Card>

  <Card title="Configure" icon="settings" href="/click-to-call/configure">
    Configure the Click To Call widget in the Voice Gateway Endpoint settings.
  </Card>

  <Card title="Embed" icon="code" href="/click-to-call/embed">
    Embed the Click To Call widget in your website.
  </Card>

  <Card title="Tutorials" icon="book-open" href="/click-to-call/tutorials/overview">
    These tutorials help you with advanced setups for the Click To Call widget.
  </Card>

  <Card title="SDK" icon="hammer" href="/click-to-call/sdk/overview">
    Build a custom voice interface from scratch with the Click To Call SDK.
  </Card>
</CardGroup>


# About Click To Call
Source: https://docs.cognigy.com/click-to-call/overview



<a href="/release-notes/2026.6"><Badge>Added in 2026.6</Badge></a>

*Click To Call* lets you create a voice AI Agent that end users can talk to with a single click.

You can deploy Click To Call through the following options:

* **Click To Call Widget** — you can easily embed a widget into your website. Additionally, you can use the Click To Call [Widget API and CSS styles](https://github.com/Cognigy/click-to-call-widget) to customize your AI Agent.
* **Click To Call SDK** — you can build a custom application using the Click To Call [SDK](https://github.com/Cognigy/click-to-call-sdk).

<div>
  <video title="Click To Call SDK showing call flow" aria-label="Click To Call demo showing a user clicking the Click To Call widget and having a conversation with the AI Agent">
    <source />
  </video>
</div>

## Key Benefits

* **Optimized latency and high definition audio quality**. The underlying web real-time communication (WebRTC) protocol and Opus codec allow for low-latency and higher quality audio communication.
* **Multimodal communication**. Users can communicate primarily through voice with the AI Agent, but you can extend the capabilities of your voice AI Agents with [event handling and API methods](https://github.com/Cognigy/click-to-call-widget).
* **Centralized entry point**. Users can access your voice AI Agent directly through the widget while browsing through your website without the need to call a phone number.
* **Swift voice communication**. Users can start a conversation with your voice AI Agent with just one click.
* **Transfer calls**. The Click To Call widget supports [Dial call transfers](/ai/agents/develop/node-reference/voice/voice-gateway/transfer). This way, you can transfer the call to a human agent or a contact center.
* **Fully customizable**. You can customize the Click To Call widget with the CSS styles in the [Click To Call widget GitHub repository](https://github.com/Cognigy/click-to-call-widget) to match your brand or build a custom application using the [Click To Call SDK](https://github.com/Cognigy/click-to-call-sdk).

## How to Start with Click To Call

Start by reading the following articles and seeing what Click To Call is capable of:

* [Quickstart: Set up Your First Click To Call Widget](/click-to-call/quickstart)
* [Configure the Click To Call Widget](/click-to-call/configure)
* [Embed the Click To Call Widget in Your Website](/click-to-call/embed)

## More Information

* [Voice Gateway Endpoint](/ai/agents/deploy/endpoint-reference/voice-gateway)
* [Voice Gateway Self-Service Portal](/voice-gateway/webapp/overview)
* [Tutorials](/click-to-call/tutorials/overview)


# Quickstart: Set up Your First Click To Call Widget
Source: https://docs.cognigy.com/click-to-call/quickstart



This quickstart guide walks you through setting up a [Click To Call widget](/click-to-call/overview) for a basic setup. The basic setup includes voice communication and live transcription.

## Prerequisites

* Create an [AI Agent](/ai/agents/develop/manage-ai-agents).
* Configure a [speech provider](/ai/agents/test/voice-preview#set-up-a-speech-provider) in the [Voice Preview Settings](/ai/administer/access/project-settings#voice-preview-settings) section of the [Project settings](/ai/administer/access/project-settings).
* Ensure [Voice Gateway](/voice-gateway/getting-started#set-up-voice-gateway-for-your-organization) is active for your organization.
* For testing purposes, allow microphone access in your browser and, if needed, in your device settings.

## Set Up a Click To Call Widget

<Steps>
  <Step title="Create a Flow">
    1. Go to **Build > Flows**, click **+ New Flow**
    2. On the **New Flow** panel, enter a unique name for your Flow and click **Create**. You are redirected to the Flow editor.
    3. In the Flow editor, add an [AI Agent Node](/ai/agents/develop/node-reference/ai/ai-agent) and configure the AI Agent you previously created.
  </Step>

  <Step title="Configure a Voice Gateway Endpoint">
    1. Go to **Deploy > Endpoints**, click **+ New Endpoint** and select **Voice Gateway**.

    2. In the **New Endpoint** panel, configure the following:
       * **Name** — enter a unique name, for example, `Click To Call Endpoint`.
       * **Flow** — select the Flow you previously created, for example, `Click To Call Assistant`. Click **Save**. You are then redirected to the Endpoint settings.

    3. In the Endpoint settings, click the **Set Up Click To Call Integration** button in the **Click To Call Embedding HTML** field. The embedding code is generated in the code editor, and an [Application](/voice-gateway/webapp/applications) is created in the Voice Gateway Self-Service Portal. You can use this code snippet to [embed the Click To Call widget in your website](/click-to-call/embed).

    4. In the **Click To Call Widget Settings**, configure the following:
       * **Enable Click To Call Widget** — activate the toggle.
       * **AI Agent Name** — enter the name displayed on the Click To Call widget, for example, `Voice Support`.
       * *(Optional)* **Tagline** — enter the text displayed under the AI Agent name on the widget, for example, `Just ask me anything!`.

    5. Click **Save**.
  </Step>

  <Step title="Test the Click To Call widget">
    1. In the upper-right corner of the Endpoint settings, click the **Open Demo Widget** button. The Demo Widget opens in a new tab.
    2. Click <img alt="start voice call" /> to start a call with your AI Agent.
  </Step>
</Steps>

## Next Steps

* [Embed the Click To Call Widget in Your Website](/click-to-call/embed)

## More Information

* [Configure the Click To Call Widget](/click-to-call/configure)
* [Click To Call GitHub Reference](https://github.com/Cognigy/click-to-call-widget)
* [Click To Call SDK GitHub Reference](https://github.com/Cognigy/click-to-call-sdk)
* [Voice Gateway Endpoint](/ai/agents/deploy/endpoint-reference/voice-gateway#click-to-call-widget-settings)


# checkWebRTCSupport
Source: https://docs.cognigy.com/click-to-call/sdk/api-reference/check-click-to-call-support



Checks whether the current browser supports Click To Call. Call this function before creating a client to verify that all required browser APIs are available.

## Signature

```typescript theme={null}
import { checkWebRTCSupport } from '@cognigy/click-to-call-sdk';

const support = checkWebRTCSupport();
```

## Parameters

None.

## Returns

`{ supported: boolean; missing: string[] }`

| Property    | Type       | Description                                              |
| ----------- | ---------- | -------------------------------------------------------- |
| `supported` | `boolean`  | `true` if the browser supports all required WebRTC APIs. |
| `missing`   | `string[]` | List of missing browser features, if any.                |

## Example

```typescript theme={null}
import { checkWebRTCSupport } from '@cognigy/click-to-call-sdk';

const support = checkWebRTCSupport();
if (!support.supported) {
  console.error('WebRTC not supported. Missing features:', support.missing);
} else {
  console.log('WebRTC is supported — ready to create a client');
}
```

## More Information

* [createWebRTCClient](/click-to-call/sdk/api-reference/create-click-to-call-client) — create a client after verifying browser support


# connect
Source: https://docs.cognigy.com/click-to-call/sdk/api-reference/connect



Connects to the SIP server and registers the client. This method fetches the SIP configuration from the endpoint URL, validates the widget is active, initializes the SIP user agent, and waits for both WebSocket connection and SIP registration to complete (with a 15-second timeout). This must be called before [`startCall()`](/click-to-call/sdk/api-reference/start-call).

## Signature

```typescript theme={null}
await client.connect();
```

## Parameters

None.

## Returns

`Promise<void>`

## Throws

An error if the client has been destroyed, the configuration fetch fails, the Click To Call widget is not active, or the SIP connection or registration fails (15-second timeout).

## Example

```typescript theme={null}
await client.connect();
console.log('Connected and registered');
```

### With Error Handling

```typescript theme={null}
try {
  await client.connect();
  console.log('Connected and registered');
} catch (error) {
  console.error('Failed to connect:', error.message);
}
```

## More Information

* [disconnect](/click-to-call/sdk/api-reference/disconnect)
* [connectAndCall](/click-to-call/sdk/api-reference/connect-and-call)
* [startCall](/click-to-call/sdk/api-reference/start-call)
* [registered](/click-to-call/sdk/event-reference/registered) event


# connectAndCall
Source: https://docs.cognigy.com/click-to-call/sdk/api-reference/connect-and-call



Connects to the SIP server and starts a call in one operation. This is a convenience method that combines [`connect()`](/click-to-call/sdk/api-reference/connect) and [`startCall()`](/click-to-call/sdk/api-reference/start-call).

## Signature

```typescript theme={null}
await client.connectAndCall();
```

## Parameters

None.

## Returns

`Promise<void>`

## Throws

An error if the connection or call initiation fails.

## Example

```typescript theme={null}
await client.connectAndCall();
console.log('Connected and call started');
```

### With Error Handling

```typescript theme={null}
try {
  await client.connectAndCall();
} catch (error) {
  console.error('Failed to connect and call:', error.message);
}
```

## More Information

* [connect](/click-to-call/sdk/api-reference/connect)
* [startCall](/click-to-call/sdk/api-reference/start-call)
* [endCall](/click-to-call/sdk/api-reference/end-call)


# createWebRTCClient
Source: https://docs.cognigy.com/click-to-call/sdk/api-reference/create-click-to-call-client



Creates and returns a new `WebRTCClient` instance. This is the main entry point of the SDK.

## Signature

```typescript theme={null}
import { createWebRTCClient } from '@cognigy/click-to-call-sdk';

const client = await createWebRTCClient(config: WebRTCClientConfig);
```

## Parameters

| Parameter | Type                                                                              | Description                          | Required |
| --------- | --------------------------------------------------------------------------------- | ------------------------------------ | -------- |
| `config`  | [`WebRTCClientConfig`](/click-to-call/sdk/api-reference/types#webrtcclientconfig) | Configuration object for the client. | Yes      |

## Returns

`Promise<WebRTCClient>` — a configured client instance ready to connect and make calls.

## Example

```typescript theme={null}
import { createWebRTCClient } from '@cognigy/click-to-call-sdk';

const client = await createWebRTCClient({
  endpointUrl: 'https://your-api.com/voice-connect-config',
  userId: 'user-123',
});
```

### With Custom ICE Servers

```typescript theme={null}
const client = await createWebRTCClient({
  endpointUrl: 'https://your-api.com/voice-connect-config',
  userId: 'user-123',
  pcConfig: {
    iceServers: [
      { urls: 'stun:stun.l.google.com:19302' },
      { urls: 'turn:your-turn-server.com:3478', username: 'user', credential: 'password' }
    ]
  }
});
```

## More Information

* [WebRTCClientConfig](/click-to-call/sdk/api-reference/types#webrtcclientconfig)
* [checkWebRTCSupport](/click-to-call/sdk/api-reference/check-click-to-call-support)


# destroy
Source: https://docs.cognigy.com/click-to-call/sdk/api-reference/destroy



Disconnects from the SIP server, ends any active call, and releases all resources. This method calls `disconnect()`, then destroys all internal managers (`AudioManager`, `SessionManager`, `SipManager`), clears the cached configuration, and removes all event listeners. Call this method when the client is no longer needed, for example, on page unload. Once destroyed, the client instance cannot be reused.

## Signature

```typescript theme={null}
await client.destroy();
```

## Parameters

None.

## Returns

`Promise<void>`

## Throws

An error if the cleanup process fails.

## Example

```typescript theme={null}
await client.destroy();
console.log('Client destroyed');
```

### Cleanup on Page Unload

```typescript theme={null}
window.addEventListener('beforeunload', () => {
  void client.destroy().catch(() => {
    // Ignore errors during page unload
  });
});
```

## More Information

* [disconnect](/click-to-call/sdk/api-reference/disconnect)
* [endCall](/click-to-call/sdk/api-reference/end-call)


# disconnect
Source: https://docs.cognigy.com/click-to-call/sdk/api-reference/disconnect



Disconnects from the SIP server. This method stops audio playback, terminates any active call sessions, and closes the SIP WebSocket connection.

## Signature

```typescript theme={null}
await client.disconnect();
```

## Parameters

None.

## Returns

`Promise<void>`

## Throws

An error if the disconnection process encounters an issue, such as a failure to stop the SIP connection.

## Example

```typescript theme={null}
await client.disconnect();
console.log('Disconnected from SIP server');
```

## More Information

* [connect](/click-to-call/sdk/api-reference/connect)
* [destroy](/click-to-call/sdk/api-reference/destroy)
* [disconnected](/click-to-call/sdk/event-reference/disconnected) event


# endCall
Source: https://docs.cognigy.com/click-to-call/sdk/api-reference/end-call



Ends the current active call.

## Signature

```typescript theme={null}
await client.endCall();
```

## Parameters

None.

## Returns

`Promise<void>`

## Throws

An error if there is no active session to terminate or if the termination fails.

## Example

```typescript theme={null}
await client.endCall();
console.log('Call ended');
```

### With Error Handling

```typescript theme={null}
try {
  await client.endCall();
} catch (error) {
  console.error('Failed to end call:', error.message);
}
```

## More Information

* [startCall](/click-to-call/sdk/api-reference/start-call)
* [disconnect](/click-to-call/sdk/api-reference/disconnect)
* [ended](/click-to-call/sdk/event-reference/ended) event
* [failed](/click-to-call/sdk/event-reference/failed) event


# getCurrentSession
Source: https://docs.cognigy.com/click-to-call/sdk/api-reference/get-current-session



Returns the current active call session, or `null` if no call is active.

## Signature

```typescript theme={null}
const session = client.getCurrentSession();
```

## Parameters

None.

## Returns

[`CallSession`](/click-to-call/sdk/api-reference/types#callsession) `| null` — the active call session, or `null` if no call is in progress.

## Example

```typescript theme={null}
const session = client.getCurrentSession();
if (session) {
  console.log('Active call:', session.id);
  console.log('Muted:', session.muted);
} else {
  console.log('No active call');
}
```

### Toggle Mute Based on Current State

```typescript theme={null}
const session = client.getCurrentSession();
if (session) {
  if (session.muted) {
    await client.unmute();
  } else {
    await client.mute();
  }
}
```

## More Information

* [CallSession](/click-to-call/sdk/api-reference/types#callsession)
* [mute](/click-to-call/sdk/api-reference/mute)
* [unmute](/click-to-call/sdk/api-reference/unmute)


# isConnected
Source: https://docs.cognigy.com/click-to-call/sdk/api-reference/is-connected



Checks whether the client is connected and registered with the SIP server.

## Signature

```typescript theme={null}
const connected = client.isConnected();
```

## Parameters

None.

## Returns

`boolean` — `true` if the client is connected and registered, `false` otherwise.

## Example

```typescript theme={null}
if (client.isConnected()) {
  console.log('Client is ready to make calls');
  await client.startCall();
} else {
  console.log('Client is not connected');
  await client.connect();
}
```

## More Information

* [connect](/click-to-call/sdk/api-reference/connect)
* [registered](/click-to-call/sdk/event-reference/registered) event


# mute
Source: https://docs.cognigy.com/click-to-call/sdk/api-reference/mute



Mutes the local microphone for the current call. The remote party won't hear audio from the local user.

## Signature

```typescript theme={null}
await client.mute();
```

## Parameters

None.

## Returns

`Promise<void>`

## Throws

An error if there is no active session to mute or if the session is not established.

## Example

```typescript theme={null}
await client.mute();
console.log('Microphone muted');
```

### Toggle Mute

```typescript theme={null}
const session = client.getCurrentSession();
if (session) {
  if (session.muted) {
    await client.unmute();
  } else {
    await client.mute();
  }
}
```

## More Information

* [unmute](/click-to-call/sdk/api-reference/unmute)
* [getCurrentSession](/click-to-call/sdk/api-reference/get-current-session)
* [muted](/click-to-call/sdk/event-reference/muted) event


# off
Source: https://docs.cognigy.com/click-to-call/sdk/api-reference/off



Removes a previously registered event listener.

## Signature

```typescript theme={null}
client.off<T extends EventName>(event: T, callback: EventCallback<T>): this;
```

## Parameters

| Parameter  | Type               | Description                                  | Required |
| ---------- | ------------------ | -------------------------------------------- | -------- |
| `event`    | `EventName`        | The name of the event to stop listening for. | Yes      |
| `callback` | `EventCallback<T>` | The callback function to remove.             | Yes      |

## Returns

`this` — the client instance, allowing method chaining.

<Warning>
  You must pass the same function reference that was used with [`on()`](/click-to-call/sdk/api-reference/on). Anonymous functions can't be removed.
</Warning>

## Example

```typescript theme={null}
const onAnswered = (session) => {
  console.log('Call answered:', session.id);
};

// Subscribe
client.on('answered', onAnswered);

// Unsubscribe
client.off('answered', onAnswered);
```

## More Information

* [on](/click-to-call/sdk/api-reference/on)
* [Event Reference](/click-to-call/sdk/event-reference/overview)


# on
Source: https://docs.cognigy.com/click-to-call/sdk/api-reference/on



Adds an event listener for a specific event. See the [Event Reference](/click-to-call/sdk/event-reference/overview) for all available events.

## Signature

```typescript theme={null}
client.on<T extends EventName>(event: T, callback: EventCallback<T>): this;
```

## Parameters

| Parameter  | Type               | Description                                                | Required |
| ---------- | ------------------ | ---------------------------------------------------------- | -------- |
| `event`    | `EventName`        | The name of the event to listen for.                       | Yes      |
| `callback` | `EventCallback<T>` | The callback function to invoke when the event is emitted. | Yes      |

## Returns

`this` — the client instance, allowing method chaining.

## Example

```typescript theme={null}
client.on('answered', (session) => {
  console.log('Call answered:', session.id);
});

client.on('error', (error) => {
  console.error('SDK error:', error.message);
});
```

### Type-Safe Event Handling

```typescript theme={null}
import type { WebRTCClientEvents } from '@cognigy/click-to-call-sdk';

const handleAnswered: WebRTCClientEvents['answered'] = (session) => {
  console.log('Call answered:', session.id);
};

client.on('answered', handleAnswered);
```

## More Information

* [off](/click-to-call/sdk/api-reference/off)
* [Event Reference](/click-to-call/sdk/event-reference/overview)
* [WebRTCClientEvents](/click-to-call/sdk/api-reference/types#webrtcclientevents)


# API Reference
Source: https://docs.cognigy.com/click-to-call/sdk/api-reference/overview



This section provides a detailed reference for all functions, methods, configuration options, and types available in the Cognigy Click To Call SDK.

## Factory Functions

| Function                                                                           | Description                                                |
| ---------------------------------------------------------------------------------- | ---------------------------------------------------------- |
| [createWebRTCClient](/click-to-call/sdk/api-reference/create-click-to-call-client) | Creates and returns a new `WebRTCClient` instance.         |
| [checkWebRTCSupport](/click-to-call/sdk/api-reference/check-click-to-call-support) | Checks whether the current browser supports Click To Call. |

## Connection Methods

| Method                                                              | Description                                          |
| ------------------------------------------------------------------- | ---------------------------------------------------- |
| [connect](/click-to-call/sdk/api-reference/connect)                 | Connects to the SIP server and registers the client. |
| [disconnect](/click-to-call/sdk/api-reference/disconnect)           | Disconnects from the SIP server.                     |
| [connectAndCall](/click-to-call/sdk/api-reference/connect-and-call) | Connects and starts a call in one operation.         |

## Call Control Methods

| Method                                                   | Description                          |
| -------------------------------------------------------- | ------------------------------------ |
| [startCall](/click-to-call/sdk/api-reference/start-call) | Starts a new voice call.             |
| [endCall](/click-to-call/sdk/api-reference/end-call)     | Ends the current active call.        |
| [mute](/click-to-call/sdk/api-reference/mute)            | Mutes the local microphone.          |
| [unmute](/click-to-call/sdk/api-reference/unmute)        | Unmutes the local microphone.        |
| [sendInfo](/click-to-call/sdk/api-reference/send-info)   | Sends an info message during a call. |

## State Methods

| Method                                                                    | Description                                       |
| ------------------------------------------------------------------------- | ------------------------------------------------- |
| [isConnected](/click-to-call/sdk/api-reference/is-connected)              | Checks if the client is connected and registered. |
| [getCurrentSession](/click-to-call/sdk/api-reference/get-current-session) | Returns the current active call session.          |

## Event Methods

| Method                                      | Description                |
| ------------------------------------------- | -------------------------- |
| [on](/click-to-call/sdk/api-reference/on)   | Adds an event listener.    |
| [off](/click-to-call/sdk/api-reference/off) | Removes an event listener. |

## Cleanup Methods

| Method                                              | Description                                          |
| --------------------------------------------------- | ---------------------------------------------------- |
| [destroy](/click-to-call/sdk/api-reference/destroy) | Disconnects, ends calls, and releases all resources. |

## Types

| Type                                            | Description                                                                                |
| ----------------------------------------------- | ------------------------------------------------------------------------------------------ |
| [Types](/click-to-call/sdk/api-reference/types) | `WebRTCClientConfig`, `CallSession`, `SessionStatus`, `CallEndInfo`, `WebRTCClientEvents`. |


# sendInfo
Source: https://docs.cognigy.com/click-to-call/sdk/api-reference/send-info



Sends an info message during an active call. Info messages can carry text and optional structured data.

## Signature

```typescript theme={null}
await client.sendInfo(text: string, data?: Record<string, any>);
```

## Parameters

| Parameter | Type                  | Description                                          | Required |
| --------- | --------------------- | ---------------------------------------------------- | -------- |
| `text`    | `string`              | The text content of the info message.                | Yes      |
| `data`    | `Record<string, any>` | Optional key-value data to include with the message. | No       |

## Returns

`Promise<void>`

## Throws

An error if there is no active session to send the info message on, or if the session is not established.

## Example

```typescript theme={null}
await client.sendInfo('Hello', { key: 'value' });
```

### Sending Structured Data

```typescript theme={null}
await client.sendInfo('custom-action', {
  action: 'update',
  payload: { key: 'score', value: 42 }
});
```

## More Information

* [infoSent](/click-to-call/sdk/event-reference/info-sent) event
* [infoReceived](/click-to-call/sdk/event-reference/info-received) event


# startCall
Source: https://docs.cognigy.com/click-to-call/sdk/api-reference/start-call



Starts a new voice call. The client must be connected and registered before calling this method. Use [`connect()`](/click-to-call/sdk/api-reference/connect) or [`connectAndCall()`](/click-to-call/sdk/api-reference/connect-and-call) first.

## Signature

```typescript theme={null}
await client.startCall();
```

## Parameters

None.

## Returns

`Promise<void>`

## Throws

An error if the client isn't connected (`'Client not connected. Call connect() first.'`), the SIP client isn't registered, or the call fails to start.

## Example

```typescript theme={null}
await client.connect();
await client.startCall();
console.log('Call started');
```

### With Error Handling

```typescript theme={null}
try {
  await client.connect();
  await client.startCall();
} catch (error) {
  console.error('Failed to start call:', error.message);
}
```

## More Information

* [connect](/click-to-call/sdk/api-reference/connect)
* [connectAndCall](/click-to-call/sdk/api-reference/connect-and-call)
* [endCall](/click-to-call/sdk/api-reference/end-call)
* [answered](/click-to-call/sdk/event-reference/answered) event


# Types
Source: https://docs.cognigy.com/click-to-call/sdk/api-reference/types

Reference for TypeScript types and interfaces in the Cognigy Click To Call SDK.

This page documents all TypeScript types and interfaces exported by the Cognigy Click To Call SDK.

## WebRTCClientConfig

The configuration object passed to [`createWebRTCClient()`](/click-to-call/sdk/api-reference/create-click-to-call-client).

```typescript theme={null}
interface WebRTCClientConfig {
  endpointUrl: string;
  userId?: string;
  captureAudio?: boolean;
  pcConfig?: RTCConfiguration;
}
```

| Property       | Type               | Description                                                                                                                                                                      | Required |
| -------------- | ------------------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------- |
| `endpointUrl`  | `string`           | URL to fetch SIP configuration from the Cognigy Voice Gateway Endpoint.                                                                                                          | Yes      |
| `userId`       | `string`           | Optional user identifier for the session.                                                                                                                                        | No       |
| `captureAudio` | `boolean`          | When set to `true`, the SDK emits the [`captureAudio`](/click-to-call/sdk/event-reference/capture-audio) event with the remote `MediaStream` during a call. Defaults to `false`. | No       |
| `pcConfig`     | `RTCConfiguration` | Custom WebRTC peer connection configuration. Use this to specify STUN/TURN servers or other `RTCPeerConnection` settings.                                                        | No       |

## WebRTCClient

The main client interface returned by [`createWebRTCClient()`](/click-to-call/sdk/api-reference/create-click-to-call-client). Provides all call control, state query, and event handling methods.

| Method                | Reference                                                                 |
| --------------------- | ------------------------------------------------------------------------- |
| `connect()`           | [connect](/click-to-call/sdk/api-reference/connect)                       |
| `disconnect()`        | [disconnect](/click-to-call/sdk/api-reference/disconnect)                 |
| `connectAndCall()`    | [connectAndCall](/click-to-call/sdk/api-reference/connect-and-call)       |
| `startCall()`         | [startCall](/click-to-call/sdk/api-reference/start-call)                  |
| `endCall()`           | [endCall](/click-to-call/sdk/api-reference/end-call)                      |
| `mute()`              | [mute](/click-to-call/sdk/api-reference/mute)                             |
| `unmute()`            | [unmute](/click-to-call/sdk/api-reference/unmute)                         |
| `sendInfo()`          | [sendInfo](/click-to-call/sdk/api-reference/send-info)                    |
| `isConnected()`       | [isConnected](/click-to-call/sdk/api-reference/is-connected)              |
| `getCurrentSession()` | [getCurrentSession](/click-to-call/sdk/api-reference/get-current-session) |
| `on()`                | [on](/click-to-call/sdk/api-reference/on)                                 |
| `off()`               | [off](/click-to-call/sdk/api-reference/off)                               |
| `destroy()`           | [destroy](/click-to-call/sdk/api-reference/destroy)                       |

## CallSession

Represents an active call session.

```typescript theme={null}
interface CallSession {
  id: string;
  status: SessionStatus;
  direction: 'incoming' | 'outgoing';
  startTime: Date;
  answerTime?: Date;
  duration: number;
  muted: boolean;
  localHold: boolean;
  remoteHold: boolean;
}
```

| Property     | Type                       | Description                                                                            |
| ------------ | -------------------------- | -------------------------------------------------------------------------------------- |
| `id`         | `string`                   | Unique identifier for the call session.                                                |
| `status`     | `SessionStatus`            | Current session status: `'init'`, `'ringing'`, `'answered'`, `'failed'`, or `'ended'`. |
| `direction`  | `'incoming' \| 'outgoing'` | Whether the call was initiated locally or received from a remote party.                |
| `startTime`  | `Date`                     | Timestamp when the session was created.                                                |
| `answerTime` | `Date \| undefined`        | Timestamp when the call was answered, or `undefined` if not yet answered.              |
| `duration`   | `number`                   | Call duration in seconds since the call was answered.                                  |
| `muted`      | `boolean`                  | Whether the local microphone is muted.                                                 |
| `localHold`  | `boolean`                  | Whether the call is on hold locally.                                                   |
| `remoteHold` | `boolean`                  | Whether the call is on hold by the remote party.                                       |

### SessionStatus

```typescript theme={null}
type SessionStatus = 'init' | 'ringing' | 'answered' | 'failed' | 'ended';
```

## CallEndInfo

Contains information about why a call ended.

```typescript theme={null}
interface CallEndInfo {
  originator: 'local' | 'remote' | null;
  cause: string | null;
  description?: string | null;
}
```

| Property      | Type                          | Description                                                                          |
| ------------- | ----------------------------- | ------------------------------------------------------------------------------------ |
| `originator`  | `'local' \| 'remote' \| null` | Who initiated the call end — the local user, the remote party, or `null` if unknown. |
| `cause`       | `string \| null`              | The reason the call ended, for example, `Terminated`, `Canceled`, `Rejected`.        |
| `description` | `string \| null \| undefined` | Additional description, such as a SIP status code or SIP Reason header value.        |

## Transcription Data

The [`transcription`](/click-to-call/sdk/event-reference/transcription) event emits the value of the `_transcription` property from incoming SIP INFO messages. The structure depends on your Cognigy Voice Gateway configuration, but typically has the following shape:

```typescript theme={null}
interface TranscriptionData {
  messages: { text: string }[];
  originator: 'user' | 'bot';
}
```

| Property     | Type                 | Description                                                                  |
| ------------ | -------------------- | ---------------------------------------------------------------------------- |
| `originator` | `string`             | Indicates the source of the transcription, for example, `'user'` or `'bot'`. |
| `messages`   | `{ text: string }[]` | Array of message objects, each containing a `text` field.                    |

<Note>
  This is not a formally defined SDK interface. The transcription data is passed through as-is from the SIP INFO message body. The actual structure depends on your Voice Gateway configuration.
</Note>

## WebRTCClientEvents

Type map of all event names to their callback signatures. Use this for type-safe event handling.

The following events are emitted by the client:

| Event           | Callback Signature                                                              |
| --------------- | ------------------------------------------------------------------------------- |
| `connecting`    | `() => void`                                                                    |
| `connected`     | `() => void`                                                                    |
| `disconnected`  | `() => void`                                                                    |
| `registered`    | `() => void`                                                                    |
| `unregistered`  | `() => void`                                                                    |
| `ringing`       | `(session: CallSession) => void`                                                |
| `answered`      | `(session: CallSession) => void`                                                |
| `ended`         | `(session: CallSession, endInfo: CallEndInfo) => void`                          |
| `failed`        | `(session: CallSession, endInfo: CallEndInfo) => void`                          |
| `muted`         | `(session: CallSession) => void`                                                |
| `unmuted`       | `(session: CallSession) => void`                                                |
| `audioEnded`    | `() => void`                                                                    |
| `infoSent`      | `(text: string, data: Record<string, any>) => void`                             |
| `infoReceived`  | `(data: { originator: string, info: { body: string } }) => void`                |
| `transcription` | `(transcription: { originator: string; messages: { text: string }[] }) => void` |
| `error`         | `(error: Error) => void`                                                        |
| `captureAudio`  | `(stream: MediaStream) => void`                                                 |

```typescript theme={null}
import type { WebRTCClientEvents } from '@cognigy/click-to-call-sdk';

const handleAnswered: WebRTCClientEvents['answered'] = (session) => {
  console.log('Call answered:', session.id);
};

client.on('answered', handleAnswered);
```

For all available events, see the [Event Reference](/click-to-call/sdk/event-reference/overview).

## TypeScript Usage

The SDK is written in TypeScript and ships comprehensive type definitions. All interfaces and types can be imported directly:

```typescript theme={null}
import type {
  WebRTCClient,
  CallSession,
  CallEndInfo,
  WebRTCClientEvents
} from '@cognigy/click-to-call-sdk';
```

### Type-Safe Event Handling

Use the `WebRTCClientEvents` type map for fully typed event callbacks:

```typescript theme={null}
const handleAnswered: WebRTCClientEvents['answered'] = (session) => {
  console.log('Call answered:', session.id);
};

client.on('answered', handleAnswered);
```

### Type-Safe Info Messages

When sending structured data via [`sendInfo()`](/click-to-call/sdk/api-reference/send-info), you can define your data shape:

```typescript theme={null}
interface MyInfoData {
  action: string;
  payload: { key: string; value: number };
}

const data: MyInfoData = {
  action: 'update',
  payload: { key: 'score', value: 42 }
};

await client.sendInfo('custom-action', data);
```


# unmute
Source: https://docs.cognigy.com/click-to-call/sdk/api-reference/unmute



Unmutes the local microphone for the current call.

## Signature

```typescript theme={null}
await client.unmute();
```

## Parameters

None.

## Returns

`Promise<void>`

## Example

```typescript theme={null}
await client.unmute();
console.log('Microphone unmuted');
```

## More Information

* [mute](/click-to-call/sdk/api-reference/mute)
* [getCurrentSession](/click-to-call/sdk/api-reference/get-current-session)
* [unmuted](/click-to-call/sdk/event-reference/unmuted) event


# Custom Audio Handling
Source: https://docs.cognigy.com/click-to-call/sdk/custom-audio

Configure custom audio handling in the Click To Call SDK. Learn how to capture raw `MediaStream` for advanced audio processing, visualization, or custom WebRTC routing.

The Cognigy Click To Call SDK handles remote audio playback automatically by creating and managing an internal `Audio` element. For advanced use cases, you can opt in to receive the raw `MediaStream` via the `captureAudio` event by setting `captureAudio: true` in the client configuration. The SDK also notifies when playback stops via the [`audioEnded`](/click-to-call/sdk/event-reference/audio-ended) event.

## Default Behavior

By default, the SDK:

1. Creates an internal `Audio` element with `autoplay` enabled and volume set to `1.0`.
2. When a remote audio stream arrives via the WebRTC peer connection, the SDK attaches it to the `Audio` element and starts playback automatically.
3. When the call ends, the SDK stops playback, releases the stream, and emits the `audioEnded` event.

No manual audio setup is required for basic call functionality.

<Note>
  Browser autoplay policies may block audio playback if the call is not initiated by a user gesture, for example, a button click. Always start calls in response to user interaction.
</Note>

## Accessing the Raw Audio Stream

To receive the raw `MediaStream` for advanced audio processing, enable the `captureAudio` option and subscribe to the event:

```typescript theme={null}
const client = await createWebRTCClient({
  endpointUrl: 'https://your-cognigy-environment.com/token',
  userId: 'user-123',
  captureAudio: true,
});

client.on('captureAudio', (stream) => {
  // The stream contains the remote audio track
  console.log('Audio stream received:', stream.getAudioTracks());
});
```

<Note>
  The `captureAudio` event emits in addition to the SDK's automatic audio playback. You don't need to handle playback yourself — the SDK still plays audio through its internal `Audio` element.
</Note>

## Volume Control

Use the Web Audio API to add programmatic volume control on top of the captured stream:

```typescript expandable theme={null}
let gainNode: GainNode;
let audioContext: AudioContext;

client.on('captureAudio', (stream) => {
  audioContext = new AudioContext();
  const source = audioContext.createMediaStreamSource(stream);

  gainNode = audioContext.createGain();
  gainNode.gain.value = 1.0; // Full volume

  source.connect(gainNode);
  // Don't connect to audioContext.destination — the SDK already handles playback.
  // Use gainNode only if you need to route audio to a custom destination.
});

// Adjust volume (0.0 = silent, 1.0 = full)
function setVolume(level: number) {
  if (gainNode) {
    gainNode.gain.value = Math.max(0, Math.min(1, level));
  }
}
```

<Note>
  The example above does not connect to `audioContext.destination` because the SDK already plays audio through its internal `Audio` element. If you connect to `audioContext.destination`, you will hear audio twice. If you want to fully replace the SDK's audio playback with your own Web Audio pipeline, mute the internal audio element first.
</Note>

## Audio Visualization

Create a real-time audio visualizer using the `AnalyserNode`:

```typescript expandable theme={null}
client.on('captureAudio', (stream) => {
  const audioContext = new AudioContext();
  const source = audioContext.createMediaStreamSource(stream);
  const analyser = audioContext.createAnalyser();
  analyser.fftSize = 256;

  source.connect(analyser);
  // Don't connect to audioContext.destination — the SDK already plays audio
  // Connecting again would cause double playback

  const dataArray = new Uint8Array(analyser.frequencyBinCount);

  function draw() {
    requestAnimationFrame(draw);
    analyser.getByteFrequencyData(dataArray);

    // Use dataArray values to render a visualization
    const average = dataArray.reduce((a, b) => a + b, 0) / dataArray.length;
    console.log('Audio level:', average);
  }

  draw();
});
```

<Note>
  When using the `captureAudio` stream for visualization, avoid connecting it to `audioContext.destination` since the SDK already handles playback. Doing so would cause the audio to play twice.
</Note>

## Audio Recording

Record the remote audio stream using the `MediaRecorder` API:

```typescript expandable theme={null}
let mediaRecorder: MediaRecorder;
const recordedChunks: Blob[] = [];

client.on('captureAudio', (stream) => {
  // Record audio — playback is handled automatically by the SDK
  mediaRecorder = new MediaRecorder(stream, { mimeType: 'audio/webm' });
  mediaRecorder.ondataavailable = (event) => {
    if (event.data.size > 0) {
      recordedChunks.push(event.data);
    }
  };
  mediaRecorder.start();
});

client.on('audioEnded', () => {
  if (mediaRecorder && mediaRecorder.state !== 'inactive') {
    mediaRecorder.stop();
  }
});

// Save recording
function saveRecording() {
  const blob = new Blob(recordedChunks, { type: 'audio/webm' });
  const url = URL.createObjectURL(blob);
  const a = document.createElement('a');
  a.href = url;
  a.download = 'call-recording.webm';
  a.click();
}
```

<Warning>
  Recording calls may be subject to legal requirements such as consent laws. Ensure your application complies with applicable regulations before recording audio.
</Warning>

## More Information

* [audioEnded](/click-to-call/sdk/event-reference/audio-ended)
* [Getting Started](/click-to-call/sdk/getting-started)
* [Troubleshooting — No Audio](/click-to-call/sdk/troubleshooting#no-audio)


# answered
Source: https://docs.cognigy.com/click-to-call/sdk/event-reference/answered

The answered event fires when a Click To Call call is accepted by the remote party.

Emitted when the call is answered by the remote party.

## Category

Call

## Callback Parameters

| Parameter | Type                                                                | Description              |
| --------- | ------------------------------------------------------------------- | ------------------------ |
| `session` | [`CallSession`](/click-to-call/sdk/api-reference/types#callsession) | The active call session. |

## Example

```typescript theme={null}
client.on('answered', (session) => {
  console.log('Call answered. Session ID:', session.id);
});
```

## More Information

* [ended](/click-to-call/sdk/event-reference/ended)
* [failed](/click-to-call/sdk/event-reference/failed)
* [startCall](/click-to-call/sdk/api-reference/start-call)
* [CallSession](/click-to-call/sdk/api-reference/types#callsession)


# audioEnded
Source: https://docs.cognigy.com/click-to-call/sdk/event-reference/audio-ended

The audioEnded event fires when the remote audio stream stops during a Click To Call session.

Emitted when the remote audio stream ends.

## Category

Audio

## Callback Parameters

None.

## Example

```typescript theme={null}
client.on('audioEnded', () => {
  console.log('Remote audio playback has stopped');
});
```

## More Information

* [captureAudio](/click-to-call/sdk/event-reference/capture-audio)
* [ended](/click-to-call/sdk/event-reference/ended)
* [Custom Audio Handling](/click-to-call/sdk/custom-audio)


# captureAudio
Source: https://docs.cognigy.com/click-to-call/sdk/event-reference/capture-audio

The captureAudio event exposes the raw MediaStream when remote audio playback begins.

Emitted when the SDK receives a remote audio stream and audio playback begins. This event exposes the raw `MediaStream` for advanced audio processing such as visualization, recording, or custom routing.

This event is opt-in. To enable it, set `captureAudio: true` in the client configuration.

<Note>
  The SDK plays remote audio automatically via an internal `Audio` element. You don't need to handle the `captureAudio` event for basic playback — it's only needed for advanced use cases.
</Note>

## Category

Audio

## Callback Parameters

| Parameter | Type          | Description              |
| --------- | ------------- | ------------------------ |
| `stream`  | `MediaStream` | The remote audio stream. |

## Example

```typescript theme={null}
const client = await createWebRTCClient({
  endpointUrl: 'https://your-cognigy-environment.com/token',
  userId: 'user-123',
  captureAudio: true,
});

client.on('captureAudio', (stream) => {
  console.log('Audio stream available:', stream.getAudioTracks());
});
```

### Advanced: Audio Processing with Web Audio API

```typescript theme={null}
client.on('captureAudio', (stream) => {
  const audioContext = new AudioContext();
  const source = audioContext.createMediaStreamSource(stream);

  const analyser = audioContext.createAnalyser();
  source.connect(analyser);
  // Don't connect to destination — the SDK already handles playback
});
```

## More Information

* [audioEnded](/click-to-call/sdk/event-reference/audio-ended)
* [Custom Audio Handling](/click-to-call/sdk/custom-audio)
* [Troubleshooting — No Audio](/click-to-call/sdk/troubleshooting#no-audio)


# connected
Source: https://docs.cognigy.com/click-to-call/sdk/event-reference/connected

The connected event fires when the WebSocket connection to the SIP server is established.

Emitted when the WebSocket connection to the SIP server is established. This event is emitted before SIP registration — wait for the [`registered`](/click-to-call/sdk/event-reference/registered) event before making calls.

## Category

Connection

## Callback Parameters

None.

## Example

```typescript theme={null}
client.on('connected', () => {
  console.log('WebSocket connection established');
});
```

## More Information

* [connecting](/click-to-call/sdk/event-reference/connecting)
* [registered](/click-to-call/sdk/event-reference/registered)
* [disconnected](/click-to-call/sdk/event-reference/disconnected)


# connecting
Source: https://docs.cognigy.com/click-to-call/sdk/event-reference/connecting

The connecting event fires when Click To Call starts establishing a WebSocket connection to the SIP server.

Emitted when the SDK begins establishing a WebSocket connection to the SIP server.

## Category

Connection

## Callback Parameters

None.

## Example

```typescript theme={null}
client.on('connecting', () => {
  console.log('Connecting to SIP server...');
});
```

## More Information

* [connected](/click-to-call/sdk/event-reference/connected)
* [registered](/click-to-call/sdk/event-reference/registered)
* [connect](/click-to-call/sdk/api-reference/connect)


# disconnected
Source: https://docs.cognigy.com/click-to-call/sdk/event-reference/disconnected

The disconnected event fires when the WebSocket connection to the SIP server is lost or closed.

Emitted when the WebSocket connection to the SIP server is lost or closed.

## Category

Connection

## Callback Parameters

None.

## Example

```typescript theme={null}
client.on('disconnected', () => {
  console.log('Disconnected from SIP server');
});
```

## More Information

* [connected](/click-to-call/sdk/event-reference/connected)
* [disconnect](/click-to-call/sdk/api-reference/disconnect)
* [destroy](/click-to-call/sdk/api-reference/destroy)


# ended
Source: https://docs.cognigy.com/click-to-call/sdk/event-reference/ended

The ended event fires when a Click To Call call terminates normally.

Emitted when the call ends normally, for example, when either party hangs up.

## Category

Call

## Callback Parameters

| Parameter | Type                                                                | Description                           |
| --------- | ------------------------------------------------------------------- | ------------------------------------- |
| `session` | [`CallSession`](/click-to-call/sdk/api-reference/types#callsession) | The ended call session.               |
| `endInfo` | [`CallEndInfo`](/click-to-call/sdk/api-reference/types#callendinfo) | Information about why the call ended. |

## Example

```typescript theme={null}
client.on('ended', (session, endInfo) => {
  console.log('Call ended. Cause:', endInfo.cause);
  console.log('Originator:', endInfo.originator);
  console.log('Description:', endInfo.description);
});
```

## More Information

* [failed](/click-to-call/sdk/event-reference/failed)
* [answered](/click-to-call/sdk/event-reference/answered)
* [endCall](/click-to-call/sdk/api-reference/end-call)
* [CallEndInfo](/click-to-call/sdk/api-reference/types#callendinfo)


# error
Source: https://docs.cognigy.com/click-to-call/sdk/event-reference/error

The error event fires on asynchronous SDK errors such as SIP registration failures or audio playback issues.

Emitted when an asynchronous error occurs during the SDK lifecycle, such as a SIP registration failure, audio playback error, or call error.

<Note>
  Errors thrown by async methods like `connect()` and `startCall()` can also be caught using `try`/`catch`. The `error` event captures asynchronous errors that occur outside of direct method calls.
</Note>

## Category

Error

## Callback Parameters

| Parameter | Type    | Description                                             |
| --------- | ------- | ------------------------------------------------------- |
| `error`   | `Error` | The error object with a descriptive `message` property. |

## Example

```typescript theme={null}
client.on('error', (error) => {
  console.error('Click To Call SDK error:', error.message);
});
```

## More Information

* [Troubleshooting](/click-to-call/sdk/troubleshooting)
* [failed](/click-to-call/sdk/event-reference/failed)


# failed
Source: https://docs.cognigy.com/click-to-call/sdk/event-reference/failed

The failed event fires when a Click To Call call fails to connect or is rejected.

Emitted when a call fails to connect or is rejected.

## Category

Call

## Callback Parameters

| Parameter | Type                                                                | Description                            |
| --------- | ------------------------------------------------------------------- | -------------------------------------- |
| `session` | [`CallSession`](/click-to-call/sdk/api-reference/types#callsession) | The failed call session.               |
| `endInfo` | [`CallEndInfo`](/click-to-call/sdk/api-reference/types#callendinfo) | Information about why the call failed. |

## Example

```typescript theme={null}
client.on('failed', (session, endInfo) => {
  console.error('Call failed. Cause:', endInfo.cause);
  console.error('Originator:', endInfo.originator);
  console.error('Description:', endInfo.description);
});
```

## More Information

* [ended](/click-to-call/sdk/event-reference/ended)
* [error](/click-to-call/sdk/event-reference/error)
* [CallEndInfo](/click-to-call/sdk/api-reference/types#callendinfo)


# infoReceived
Source: https://docs.cognigy.com/click-to-call/sdk/event-reference/info-received

The infoReceived event fires when a SIP INFO message arrives during a call.

Emitted when an info message is received from the remote party during a call.

<Note>
  If the incoming info message contains a `_transcription` property, it is emitted as a [`transcription`](/click-to-call/sdk/event-reference/transcription) event instead and won't trigger `infoReceived`.
</Note>

## Category

Info Message

## Callback Parameters

| Parameter | Type     | Description                                                                                             |
| --------- | -------- | ------------------------------------------------------------------------------------------------------- |
| `data`    | `object` | The raw SIP INFO event data, containing `originator` (string) and `info` (object with `body` property). |

The `data` object has the following structure:

| Property     | Type     | Description                                                                          |
| ------------ | -------- | ------------------------------------------------------------------------------------ |
| `originator` | `string` | Who sent the info message, for example, `'remote'`.                                  |
| `info`       | `object` | The SIP INFO message object. Use `info.body` to access the message body as a string. |

## Example

```typescript theme={null}
client.on('infoReceived', (data) => {
  console.log('Info received from:', data.originator);
  try {
    const parsed = JSON.parse(data.info.body);
    console.log('Parsed body:', parsed);
  } catch {
    console.log('Raw body:', data.info.body);
  }
});
```

## More Information

* [infoSent](/click-to-call/sdk/event-reference/info-sent)
* [transcription](/click-to-call/sdk/event-reference/transcription)
* [sendInfo](/click-to-call/sdk/api-reference/send-info)


# infoSent
Source: https://docs.cognigy.com/click-to-call/sdk/event-reference/info-sent

The infoSent event fires after a SIP INFO message is successfully sent.

Emitted when an info message is successfully sent during a call via [`sendInfo()`](/click-to-call/sdk/api-reference/send-info).

## Category

Info Message

## Callback Parameters

| Parameter | Type                  | Description                                   |
| --------- | --------------------- | --------------------------------------------- |
| `text`    | `string`              | The text content of the sent message.         |
| `data`    | `Record<string, any>` | The key-value data included with the message. |

## Example

```typescript theme={null}
client.on('infoSent', (text, data) => {
  console.log('Info message sent:', text, data);
});
```

## More Information

* [infoReceived](/click-to-call/sdk/event-reference/info-received)
* [sendInfo](/click-to-call/sdk/api-reference/send-info)


# muted
Source: https://docs.cognigy.com/click-to-call/sdk/event-reference/muted

The muted event fires when the local microphone is muted during a Click To Call session.

Emitted when the local microphone is muted during a call.

## Category

Call

## Callback Parameters

| Parameter | Type                                                                | Description                                          |
| --------- | ------------------------------------------------------------------- | ---------------------------------------------------- |
| `session` | [`CallSession`](/click-to-call/sdk/api-reference/types#callsession) | The current call session with `muted` set to `true`. |

## Example

```typescript theme={null}
client.on('muted', (session) => {
  console.log('Microphone muted. Session ID:', session.id);
});
```

## More Information

* [unmuted](/click-to-call/sdk/event-reference/unmuted)
* [mute](/click-to-call/sdk/api-reference/mute)
* [CallSession](/click-to-call/sdk/api-reference/types#callsession)


# Event Reference
Source: https://docs.cognigy.com/click-to-call/sdk/event-reference/overview

Complete reference for all Click To Call SDK events.

The Cognigy Click To Call SDK uses an event-driven architecture. You can subscribe to events using [`client.on()`](/click-to-call/sdk/api-reference/on) and unsubscribe using [`client.off()`](/click-to-call/sdk/api-reference/off). All events are fully typed when using TypeScript.

## Connection Events

| Event                                                           | Description                         | Callback Parameters |
| --------------------------------------------------------------- | ----------------------------------- | ------------------- |
| [connecting](/click-to-call/sdk/event-reference/connecting)     | SIP connection is starting.         | `()`                |
| [connected](/click-to-call/sdk/event-reference/connected)       | SIP connection established.         | `()`                |
| [disconnected](/click-to-call/sdk/event-reference/disconnected) | SIP connection lost.                | `()`                |
| [registered](/click-to-call/sdk/event-reference/registered)     | SIP client registered successfully. | `()`                |
| [unregistered](/click-to-call/sdk/event-reference/unregistered) | SIP client unregistered.            | `()`                |

## Call Events

| Event                                                   | Description                             | Callback Parameters                            |
| ------------------------------------------------------- | --------------------------------------- | ---------------------------------------------- |
| [ringing](/click-to-call/sdk/event-reference/ringing)   | Call is ringing (progress).             | `(session: CallSession)`                       |
| [answered](/click-to-call/sdk/event-reference/answered) | Call was answered by the remote party.  | `(session: CallSession)`                       |
| [ended](/click-to-call/sdk/event-reference/ended)       | Call ended normally.                    | `(session: CallSession, endInfo: CallEndInfo)` |
| [failed](/click-to-call/sdk/event-reference/failed)     | Call failed to connect or was rejected. | `(session: CallSession, endInfo: CallEndInfo)` |
| [muted](/click-to-call/sdk/event-reference/muted)       | Local microphone was muted.             | `(session: CallSession)`                       |
| [unmuted](/click-to-call/sdk/event-reference/unmuted)   | Local microphone was unmuted.           | `(session: CallSession)`                       |

## Audio Events

| Event                                                            | Description                                                                                                   | Callback Parameters     |
| ---------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------- | ----------------------- |
| [captureAudio](/click-to-call/sdk/event-reference/capture-audio) | Remote audio stream is available for advanced processing. Requires `captureAudio: true` in the client config. | `(stream: MediaStream)` |
| [audioEnded](/click-to-call/sdk/event-reference/audio-ended)     | Remote audio stream ended.                                                                                    | `()`                    |

## Info Message Events

| Event                                                            | Description                                  | Callback Parameters                                      |
| ---------------------------------------------------------------- | -------------------------------------------- | -------------------------------------------------------- |
| [infoSent](/click-to-call/sdk/event-reference/info-sent)         | Info message was sent successfully.          | `(text: string, data: Record<string, any>)`              |
| [infoReceived](/click-to-call/sdk/event-reference/info-received) | Info message received from the remote party. | `(data: { originator: string, info: { body: string } })` |

## Transcription Events

| Event                                                             | Description                  | Callback Parameters                                                     |
| ----------------------------------------------------------------- | ---------------------------- | ----------------------------------------------------------------------- |
| [transcription](/click-to-call/sdk/event-reference/transcription) | Transcription data received. | `(transcription: { originator: string; messages: { text: string }[] })` |

## Error Events

| Event                                             | Description        | Callback Parameters |
| ------------------------------------------------- | ------------------ | ------------------- |
| [error](/click-to-call/sdk/event-reference/error) | An error occurred. | `(error: Error)`    |

## Unsubscribing from Events

To remove an event listener, use [`client.off()`](/click-to-call/sdk/api-reference/off) with the same event name and callback reference:

```typescript theme={null}
const onAnswered = (session) => {
  console.log('Call answered:', session.id);
};

// Subscribe
client.on('answered', onAnswered);

// Unsubscribe
client.off('answered', onAnswered);
```

<Warning>
  When using `off()`, you must pass the same function reference that was used with `on()`. Anonymous functions can't be removed.
</Warning>


# registered
Source: https://docs.cognigy.com/click-to-call/sdk/event-reference/registered

The registered event fires when the SIP client successfully registers with the server.

Emitted when the SIP client successfully registers with the SIP server. After this event is emitted, the client is ready to make calls using [`startCall()`](/click-to-call/sdk/api-reference/start-call).

## Category

Connection

## Callback Parameters

None.

## Example

```typescript theme={null}
client.on('registered', () => {
  console.log('SIP client registered — ready to make calls');
});
```

## More Information

* [unregistered](/click-to-call/sdk/event-reference/unregistered)
* [connected](/click-to-call/sdk/event-reference/connected)
* [connect](/click-to-call/sdk/api-reference/connect)


# ringing
Source: https://docs.cognigy.com/click-to-call/sdk/event-reference/ringing

The ringing event fires when an outgoing call is in progress and the remote party's phone is ringing.

Emitted when an outgoing call is ringing (progress). This event indicates that the remote party's phone is ringing and the call has not yet been answered.

## Category

Call

## Callback Parameters

| Parameter | Type                                                                | Description               |
| --------- | ------------------------------------------------------------------- | ------------------------- |
| `session` | [`CallSession`](/click-to-call/sdk/api-reference/types#callsession) | The current call session. |

## Example

```typescript theme={null}
client.on('ringing', (session) => {
  console.log('Call is ringing. Session ID:', session.id);
});
```

## More Information

* [answered](/click-to-call/sdk/event-reference/answered)
* [failed](/click-to-call/sdk/event-reference/failed)
* [startCall](/click-to-call/sdk/api-reference/start-call)
* [CallSession](/click-to-call/sdk/api-reference/types#callsession)


# transcription
Source: https://docs.cognigy.com/click-to-call/sdk/event-reference/transcription

The transcription event delivers real-time voice transcription data during a call.

Emitted when transcription data is received during a call. Transcription messages are automatically separated from regular [`infoReceived`](/click-to-call/sdk/event-reference/info-received) events.

## Category

Transcription

## Callback Parameters

| Parameter       | Type     | Description                                                                                           |
| --------------- | -------- | ----------------------------------------------------------------------------------------------------- |
| `transcription` | `object` | The transcription data extracted from the `_transcription` property of the incoming SIP INFO message. |

The transcription object typically has the following structure (depending on your Voice Gateway configuration):

| Property     | Type                 | Description                                                                  |
| ------------ | -------------------- | ---------------------------------------------------------------------------- |
| `originator` | `string`             | Indicates the source of the transcription, for example, `'user'` or `'bot'`. |
| `messages`   | `{ text: string }[]` | Array of message objects, each containing a `text` field.                    |

## Example

```typescript theme={null}
client.on('transcription', (transcription) => {
  const speaker = transcription.originator === 'user' ? 'User' : 'Bot';
  transcription.messages.forEach((message) => {
    console.log(`${speaker}: ${message.text}`);
  });
});
```

<Note>
  If an incoming info message contains a `_transcription` property, it is emitted only as a `transcription` event — not as an [`infoReceived`](/click-to-call/sdk/event-reference/info-received) event.
</Note>

## Transcription Filtering

The SDK automatically separates transcription data from regular info messages. No additional configuration is needed — the filtering is handled automatically:

* **`transcription` event**: Receives only messages that contain a `_transcription` property.
* **`infoReceived` event**: Receives all other info messages.

This means you can listen to both events independently without worrying about duplicates.

## More Information

* [infoReceived](/click-to-call/sdk/event-reference/info-received)
* [Types](/click-to-call/sdk/api-reference/types#transcription-data)


# unmuted
Source: https://docs.cognigy.com/click-to-call/sdk/event-reference/unmuted

The unmuted event fires when the local microphone is unmuted during a Click To Call session.

Emitted when the local microphone is unmuted during a call.

## Category

Call

## Callback Parameters

| Parameter | Type                                                                | Description                                           |
| --------- | ------------------------------------------------------------------- | ----------------------------------------------------- |
| `session` | [`CallSession`](/click-to-call/sdk/api-reference/types#callsession) | The current call session with `muted` set to `false`. |

## Example

```typescript theme={null}
client.on('unmuted', (session) => {
  console.log('Microphone unmuted. Session ID:', session.id);
});
```

## More Information

* [muted](/click-to-call/sdk/event-reference/muted)
* [unmute](/click-to-call/sdk/api-reference/unmute)
* [CallSession](/click-to-call/sdk/api-reference/types#callsession)


# unregistered
Source: https://docs.cognigy.com/click-to-call/sdk/event-reference/unregistered

The unregistered event fires when the SIP client is unregistered from the server.

Emitted when the SIP client is unregistered.

## Category

Connection

## Callback Parameters

None.

## Example

```typescript theme={null}
client.on('unregistered', () => {
  console.log('SIP client unregistered');
});
```

## More Information

* [registered](/click-to-call/sdk/event-reference/registered)
* [disconnect](/click-to-call/sdk/api-reference/disconnect)


# Getting Started with Click To Call SDK
Source: https://docs.cognigy.com/click-to-call/sdk/getting-started

Deploy customized, interactive voice AI Agents in minutes.

By the end of this guide, you will have a working web application that uses the Cognigy Click To Call SDK to establish a SIP-based voice call from the browser. You will install the SDK, verify browser compatibility, configure a client, and make your first call.

The code examples in this guide use TypeScript. If you use plain JavaScript, remove the type annotations.

## Prerequisites

* A modern web browser with WebRTC support. For more information, refer to [Supported Browsers](/click-to-call/sdk/supported-browsers).
* [Node.js](https://nodejs.org/) version 18 or later.
* A web application where you can install npm packages.
  * The web application must be served over HTTPS or from `localhost`. WebSocket and UDP traffic must be allowed for SIP signaling and media transport.
* A Cognigy Flow with a configured Voice Gateway Endpoint.

## Install the SDK

1. Install the SDK in your web application project using npm:

   ```bash theme={null}
   npm install @cognigy/click-to-call-sdk
   ```

2. Verify the installation:

   ```bash theme={null}
   npm list @cognigy/click-to-call-sdk
   ```

## Configure the Client

Import the SDK, verify browser support, and create a client instance:

```typescript theme={null}
import { createWebRTCClient, checkWebRTCSupport } from '@cognigy/click-to-call-sdk';

const support = checkWebRTCSupport();
if (!support.supported) {
  console.error('WebRTC not supported. Missing features:', support.missing);
  throw new Error('Browser does not support WebRTC');
}

const client = await createWebRTCClient({
  endpointUrl: 'https://your-cognigy-environment.com/endpoint-token',
  userId: 'user-123',
});
```

The `endpointUrl` parameter must point to a valid [Cognigy Voice Gateway Endpoint](/ai/agents/deploy/endpoint-reference/voice-gateway):

1. In the **Configuration** section of the Voice Gateway Endpoint, copy the URL from the **Endpoint URL** field.
2. Replace `wss://` with `https://` and remove `/voiceGateway` from the path. For example, if your Endpoint URL is:

   ```txt wrap theme={null}
   wss://endpoint-trial.cognigy.ai/d17711ad79bc73f9da03865066201d744386d78cf11651a3377ae35085902268/voiceGateway
   ```

   Then your `endpointUrl` will be:

   ```txt wrap theme={null}
   https://endpoint-trial.cognigy.ai/d17711ad79bc73f9da03865066201d744386d78cf11651a3377ae35085902268
   ```

The SDK fetches SIP credentials and server details from this URL automatically.

### (Optional) Configure Custom STUN/TURN Servers

For advanced use cases, such as specifying custom STUN or TURN servers, provide a `pcConfig` object when creating the client.

Use `pcConfig` only if your application requires custom NAT traversal or TURN servers. Otherwise, the SDK uses default servers automatically.

```typescript theme={null}
const client = await createWebRTCClient({
  endpointUrl: 'https://your-cognigy-environment.com/endpoint-token',
  userId: 'user-123',
  pcConfig: {
    iceServers: [
      { urls: 'stun:stun.l.google.com:19302' },
      { urls: 'turn:your-turn-server.com:3478', username: 'user', credential: 'password' }
    ]
  }
});
```

## Register Event Listeners

Set up event listeners before making any calls:

```typescript theme={null}
// Connection events
client.on('connecting', () => console.log('Connecting to SIP server...'));
client.on('registered', () => console.log('SIP client registered'));
client.on('disconnected', () => console.log('Disconnected from SIP server'));

// Call events
client.on('answered', (session) => console.log('Call answered. Session ID:', session.id));
client.on('ended', (session, endInfo) => console.log('Call ended. Cause:', endInfo.cause));
client.on('failed', (session, endInfo) => console.error('Call failed. Cause:', endInfo.cause));

// Error handling
client.on('error', (error) => console.error('WebRTC SDK error:', error.message));
```

<Note>
  Audio playback is automatic. You don't need to add an `<audio>` element to your HTML. For advanced use cases, such as custom audio processing or visualization, refer to [Custom Audio](/click-to-call/sdk/custom-audio).
</Note>

## Start and End a Call

With the client configured and events registered, start a voice call:

```typescript theme={null}
// Connect and start the call in one step
await client.connectAndCall();
```

To end the call and clean up:

```typescript theme={null}
await client.endCall();
await client.disconnect();
```

<Tip>
  Instead of calling `endCall()` and `disconnect()` separately, you can call [`client.destroy()`](/click-to-call/sdk/api-reference/destroy) to end the call, disconnect, and release all resources in one step.
</Tip>

## Run and Verify

1. Combine the code from the previous sections into a single function and bind it to a button click in your web application. A call must be initiated by a user gesture, for example, a button click, because browsers block audio autoplay without user interaction. Refer to the [Complete Example](#complete-example).
2. Start your application and open it in a [supported browser](/click-to-call/sdk/supported-browsers).
3. Open the browser developer console.
4. Click the button that triggers the call. You should see the following messages in the console:
   * `Connecting to SIP server...` — the SDK is establishing a connection.
   * `SIP client registered` — registration succeeded.
   * `Call answered. Session ID: ...` — the call is active and audio should be playing.

If you don't hear audio or the call fails to connect, refer to [Troubleshooting](/click-to-call/sdk/troubleshooting#call-issues).

## Complete Example

```typescript expandable theme={null}
import { createWebRTCClient, checkWebRTCSupport } from '@cognigy/click-to-call-sdk';

async function main() {
  const support = checkWebRTCSupport();
  if (!support.supported) {
    console.error('WebRTC not supported:', support.missing);
    return;
  }

  const client = await createWebRTCClient({
    endpointUrl: 'https://your-cognigy-environment.com/endpoint-token',
    userId: 'user-123',
  });

  // Connection events
  client.on('connecting', () => console.log('Connecting to SIP server...'));
  client.on('registered', () => console.log('SIP client registered'));
  client.on('disconnected', () => console.log('Disconnected from SIP server'));

  // Call events
  client.on('answered', (session) => console.log('Call answered:', session.id));
  client.on('ended', (session, endInfo) => console.log('Call ended:', endInfo.cause));
  client.on('failed', (session, endInfo) => console.error('Call failed:', endInfo.cause));
  client.on('error', (error) => console.error('Error:', error.message));

  await client.connectAndCall();

  // Clean up when the page unloads.
  // destroy() disconnects the SIP client, ends any active call, and releases all resources.
  window.addEventListener('beforeunload', () => {
    void client.destroy().catch(() => {
    });
  });
}

main();
```

## Next Steps

* [API Reference](/click-to-call/sdk/api-reference/overview)
* [Event Reference](/click-to-call/sdk/event-reference/overview)
* [Security](/click-to-call/sdk/security)
* [Troubleshooting](/click-to-call/sdk/troubleshooting)


# Initialization and Authentication
Source: https://docs.cognigy.com/click-to-call/sdk/initialization



This page explains how the Cognigy Click To Call SDK initializes a connection to the Cognigy backend, how authentication and credential exchange work, and how to configure the initialization flow for different environments.

## Initialization Steps

When you call [`createWebRTCClient()`](/click-to-call/sdk/api-reference/create-click-to-call-client), the SDK performs the following initialization sequence:

1. The SDK validates that `endpointUrl` is provided.
2. The SDK checks that the current environment supports WebRTC (via `isWebRTCSupported()`).
3. The SDK creates internal managers (`ConfigManager`, `SipManager`, `SessionManager`, `AudioManager`) and wires up event handlers.
4. The client is returned, ready to call [`connect()`](/click-to-call/sdk/api-reference/connect) or [`connectAndCall()`](/click-to-call/sdk/api-reference/connect-and-call).

<Note>
  No network requests are made during `createWebRTCClient()`. The SIP configuration is fetched later when you call `connect()` or `connectAndCall()`.
</Note>

```typescript theme={null}
import { createWebRTCClient } from '@cognigy/click-to-call-sdk';

// Step 1–3 happen inside createWebRTCClient()
const client = await createWebRTCClient({
  endpointUrl: 'https://your-cognigy-environment.com/token',
  userId: 'user-123',
});
```

## Authentication

### Endpoint-Based Authentication

The SDK authenticates with the Cognigy backend through the `endpointUrl`. When the SDK fetches the SIP configuration, the Cognigy backend:

1. Validates the endpoint configuration.
2. Generates short-lived SIP credentials scoped to the session.
3. Returns the credentials along with the SIP server address.

The SDK manages these credentials internally — they are never exposed to your application code.

### User Identification

The optional `userId` parameter is sent to the Cognigy backend as part of the configuration request. It allows the backend to:

* Associate the Click To Call session with a specific user in Cognigy.
* Correlate call data with user sessions in Cognigy Insights.
* Route calls to the correct AI Agent Flow.

```typescript theme={null}
const client = await createWebRTCClient({
  endpointUrl: 'https://your-cognigy-environment.com/token',
  userId: 'user-123', // Identifies the user in the Cognigy session
});
```

<Note>
  The `userId` is used for session identification only. Don't pass sensitive information such as passwords, tokens, or personally identifiable information (PII) as the `userId`.
</Note>

## Connection and Registration

After initialization, calling [`connect()`](/click-to-call/sdk/api-reference/connect) triggers the full connection sequence:

1. **Configuration fetch**. The SDK sends an HTTPS `GET` request to the `endpointUrl` to retrieve SIP connection parameters such as server address, credentials, and application SID. The request has a 10-second timeout.
2. **Widget validation**. The SDK verifies that the Click To Call widget is active in the endpoint configuration. If it's not active, an error is thrown.
3. **SIP User Agent creation**. The SDK creates a JsSIP User Agent using the retrieved SIP credentials and WebSocket URI.
4. **WebSocket connection**. The SDK opens a WSS (WebSocket Secure) connection to the SIP server.
5. **SIP REGISTER**. The SDK sends a SIP REGISTER request using the credentials obtained from the endpoint.
6. **Registration confirmed**. Once both the WebSocket connection is established and SIP registration succeeds, the SDK emits the [`registered`](/click-to-call/sdk/event-reference/registered) event. The entire process has a 15-second timeout.

```typescript theme={null}
client.on('connecting', () => console.log('Opening WebSocket connection...'));
client.on('connected', () => console.log('WebSocket connected'));
client.on('registered', () => console.log('SIP registration complete — ready to call'));

await client.connect();
```

The [`connectAndCall()`](/click-to-call/sdk/api-reference/connect-and-call) convenience method performs the same sequence and immediately starts a call after registration.

## Handling Initialization Errors

If initialization or authentication fails, the SDK throws an error from `createWebRTCClient()` or `connect()`. Common errors include:

| Error                                | Error Source           | Recommended Action                                                        |
| ------------------------------------ | ---------------------- | ------------------------------------------------------------------------- |
| Missing `endpointUrl`                | `createWebRTCClient()` | Provide a valid endpoint URL.                                             |
| WebRTC not supported                 | `createWebRTCClient()` | Ensure the browser supports WebRTC. Use `checkWebRTCSupport()` to verify. |
| Invalid or unreachable `endpointUrl` | `connect()`            | Verify the URL is correct and accessible over HTTPS.                      |
| Backend returns invalid SIP config   | `connect()`            | Contact your Cognigy administrator.                                       |
| Click To Call widget not active      | `connect()`            | Enable the Click To Call widget in the endpoint configuration.            |
| SIP server unreachable               | `connect()`            | Check network and firewall settings.                                      |
| SIP registration rejected            | `connect()`            | Verify SIP credentials via the Cognigy backend.                           |
| Connection timeout (15 seconds)      | `connect()`            | Check network connectivity and SIP server availability.                   |
| CORS error on config fetch           | `connect()`            | Ensure the backend allows cross-origin requests from your domain.         |

```typescript theme={null}
try {
  const client = await createWebRTCClient({
    endpointUrl: 'https://your-cognigy-environment.com/token',
    userId: 'user-123',
  });
  await client.connect();
} catch (error) {
  console.error('Initialization failed:', error.message);
}
```

## Environment Configuration

<AccordionGroup>
  <Accordion title="Production">
    In production, the `endpointUrl` should point to your production Cognigy environment. Ensure HTTPS is used and the network allows WSS and UDP traffic.
  </Accordion>

  <Accordion title="Development and Testing">
    For local development, you can use `localhost` without HTTPS. Point the `endpointUrl` to a development or staging Cognigy environment.
  </Accordion>

  <Accordion title="Custom ICE Servers">
    If your environment requires specific STUN or TURN servers for NAT traversal, configure them via the `pcConfig` option during initialization:

    ```typescript theme={null}
    const client = await createWebRTCClient({
      endpointUrl: 'https://your-cognigy-environment.com/token',
      userId: 'user-123',
      pcConfig: {
        iceServers: [
          { urls: 'stun:stun.l.google.com:19302' },
          {
            urls: 'turn:your-turn-server.com:3478',
            username: 'user',
            credential: 'password'
          }
        ]
      }
    });
    ```

    For more details on ICE or TURN configuration, refer to [Security — Network Security](/click-to-call/sdk/security#network-security).
  </Accordion>
</AccordionGroup>

## More Information

* [createWebRTCClient](/click-to-call/sdk/api-reference/create-click-to-call-client)
* [connect](/click-to-call/sdk/api-reference/connect)
* [Security](/click-to-call/sdk/security)
* [Troubleshooting](/click-to-call/sdk/troubleshooting)


# Cognigy Click To Call SDK
Source: https://docs.cognigy.com/click-to-call/sdk/overview

Cognigy Click To Call SDK for WebRTC SIP voice calling in web apps and AI integration.

*Cognigy Click To Call SDK* is a JavaScript/TypeScript SDK for adding SIP-based voice calling to web applications using WebRTC. Built on [JsSIP](https://jssip.net/), the SDK provides APIs to manage voice sessions, SIP signaling, audio streams, and transcription.

The SDK is designed for developers building custom voice interfaces on top of the Cognigy platform and Cognigy Voice Gateway.

<div>
  <video title="Cognigy Click To Call SDK demo showing call initiation and handling" aria-label="Cognigy Click To Call SDK demo showing call initiation and handling">
    <source />
  </video>
</div>

## What's in This SDK?

### SDK Purpose

The Click To Call SDK helps developers:

* Establish real-time voice calls between a user's browser and Cognigy Voice Gateway using SIP signaling and the WebRTC API.
* Manage the full call lifecycle through a simple, promise-based API.
* Handle remote audio playback automatically.
* Receive real-time transcription events during calls, with automatic separation from regular info messages.

### Requirements

| Requirement     | Details                                                                                                                     |
| --------------- | --------------------------------------------------------------------------------------------------------------------------- |
| Language        | JavaScript or TypeScript. The SDK works with any framework, including React, Angular, Vue, or vanilla JS.                   |
| Node.js         | 18 or later                                                                                                                 |
| Package manager | npm                                                                                                                         |
| Browsers        | Modern web and mobile browsers. For more information, refer to [Supported Browsers](/click-to-call/sdk/supported-browsers). |
| Network         | HTTPS or `localhost`. WebSocket and UDP traffic must be allowed.                                                            |
| Backend         | A Cognigy Flow with a configured Voice Gateway Endpoint.                                                                    |

### Use Cases

* Build multimodal browser-based voice interfaces that connect to Cognigy AI Agents.
* Embed click-to-call functionality into web applications or customer portals.
* Display real-time transcription of voice conversations in your UI.
* Integrate voice calling into existing React, Angular, Vue, or vanilla JS applications without framework lock-in.

## Key Features

| Feature                   | Description                                                                                                                                                                                                                                                                        |
| ------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Custom Audio Handling     | The automatic audio playback uses an internal `Audio` element. For advanced use cases, enable `captureAudio: true` in the client config to receive the raw `MediaStream`. The `audioEnded` event signals when playback stops. See [Custom Audio](/click-to-call/sdk/custom-audio). |
| Full Call Control         | The complete call management includes start, end, mute, unmute, and info messages as async methods. See [API Reference](/click-to-call/sdk/api-reference/overview).                                                                                                                |
| Multimodality             | The multimodal interactions combine voice, text, and structured data in real time. The `sendInfo` method allows sending text and optional structured data during calls. See [API Reference: sendInfo](/click-to-call/sdk/api-reference/send-info).                                 |
| Event-Driven Architecture | The event system provides 17 events covering connection state, call state, audio, transcription, and errors. See [Event Reference](/click-to-call/sdk/event-reference/overview).                                                                                                   |
| Real-Time Transcription   | The transcription events are received during calls and automatically filtered from regular info messages. See [Event Reference: transcription](/click-to-call/sdk/event-reference/transcription).                                                                                  |

## Changelog Reference

The latest SDK version is available on npm as `@cognigy/click-to-call-sdk`. For updates and fixes, refer to the [Cognigy Click To Call SDK GitHub repository](https://github.com/Cognigy/click-to-call-sdk).

## License

The Cognigy Click To Call SDK is licensed under the MIT License.

## What's Next?

* [Install the SDK, create a client, and make your first voice call](/click-to-call/sdk/getting-started)
* [Learn how the SDK authenticates](/click-to-call/sdk/initialization)
* [Explore all methods, configuration options, and types](/click-to-call/sdk/api-reference/overview)
* [Discover events with callback signatures and examples](/click-to-call/sdk/event-reference/overview)
* [Transport encryption, credential handling, and best practices](/click-to-call/sdk/security)
* [Find solutions for common connection, call, and browser issues](/click-to-call/sdk/troubleshooting)


# Security
Source: https://docs.cognigy.com/click-to-call/sdk/security

Learn how the Cognigy Click To Call SDK handles WebRTC security, TLS encryption, SIP credentials, and HTTPS requirements for secure browser-based calls.

This page covers security aspects of the Cognigy Click To Call SDK, including transport encryption, credential handling, and best practices.

## Transport Security

### HTTPS Requirement

The SDK requires a [secure context](https://developer.mozilla.org/en-US/docs/Web/Security/Secure_Contexts) (HTTPS or `localhost`) to function. WebRTC APIs such as `RTCPeerConnection` and `getUserMedia` are not available in insecure contexts.

### WebSocket Security

The SDK connects to the SIP server using WSS (WebSocket Secure), providing TLS encryption for all SIP signaling traffic. The WebSocket URI is retrieved from the endpoint configuration and must use the `wss://` scheme.

### Media Encryption

WebRTC media streams (audio) are encrypted using DTLS-SRTP (Datagram Transport Layer Security — Secure Real-time Transport Protocol). DTLS handles the key exchange, and SRTP encrypts the audio data. This is enforced by the browser's WebRTC implementation and requires no additional configuration.

### Configuration Fetch

During [`connect()`](/click-to-call/sdk/api-reference/connect), the SDK sends an HTTPS request to the `endpointUrl` to retrieve SIP credentials, the SIP server address, and session parameters. Because this response contains authentication credentials, the request is always made over TLS-encrypted HTTPS.

## Credential Handling

### SIP Credentials

SIP credentials are fetched from the Cognigy backend via the `endpointUrl` during [`connect()`](/click-to-call/sdk/api-reference/connect). The credentials are:

* Generated by the Cognigy backend and scoped to the session.
* Managed internally by the SDK — they are never exposed to your application code.
* Used only for SIP registration and call signaling.

### User Identification

The optional `userId` parameter is used for session identification only. It is sent as part of the SIP URI for registration.

<Warning>
  Don't pass sensitive information such as passwords, tokens, or personally identifiable information (PII) as the `userId`.
</Warning>

## Network Security

### Firewall Configuration

The SDK requires the following network access:

| Protocol | Direction | Purpose                      |
| -------- | --------- | ---------------------------- |
| HTTPS    | Outbound  | Fetch endpoint configuration |
| WSS      | Outbound  | SIP signaling via WebSocket  |
| UDP/TCP  | Outbound  | RTP media streams (audio)    |

### ICE and TURN Servers

If your network uses a restrictive firewall or NAT, configure STUN/TURN servers via the `pcConfig` option to ensure media connectivity:

```typescript theme={null}
const client = await createWebRTCClient({
  endpointUrl: 'https://your-cognigy-environment.com/token',
  pcConfig: {
    iceServers: [
      { urls: 'stun:stun.l.google.com:19302' },
      {
        urls: 'turn:your-turn-server.com:3478',
        username: 'user',
        credential: 'password'
      }
    ]
  }
});
```

TURN servers relay media traffic when direct peer-to-peer connections are blocked. Use TURN with TLS (`turns:`) for encrypted relay traffic.

## Best Practices

* Serve your application over HTTPS in production. WebRTC APIs require a secure context.
* Use TURN servers with TLS in restrictive network environments.
* Do not log SIP credentials. The SDK handles credentials internally — avoid intercepting or logging them.
* Do not pass PII as `userId`. Use an opaque session identifier instead.

## More Information

* [Initialization and Authentication](/click-to-call/sdk/initialization)
* [Supported Browsers](/click-to-call/sdk/supported-browsers)
* [Troubleshooting](/click-to-call/sdk/troubleshooting)


# Supported Browsers
Source: https://docs.cognigy.com/click-to-call/sdk/supported-browsers



The Cognigy Click To Call SDK requires a browser that supports the [WebRTC APIs](https://developer.mozilla.org/en-US/docs/Web/API/WebRTC_API): `RTCPeerConnection`, `getUserMedia`.

## Web Browsers

<Note>
  WebRTC APIs are only available in [secure contexts](https://developer.mozilla.org/en-US/docs/Web/Security/Secure_Contexts) such as HTTPS or `localhost`. Ensure your application is served over HTTPS in production.
</Note>

| Browser | Minimum Version      |
| ------- | -------------------- |
| Chrome  | 93+                  |
| Firefox | 92+                  |
| Safari  | 15.4+                |
| Edge    | 93+ (Chromium-based) |

## Mobile Browsers

<Warning>
  Mobile browsers may have stricter autoplay policies for audio. Ensure that calls are initiated by a user gesture, for example, a button tap, to avoid audio playback issues.
</Warning>

| Browser             | Platform    |
| ------------------- | ----------- |
| Chrome for Android  | Android 93+ |
| Safari for iOS      | iOS 15.4+   |
| Firefox for Android | Android 92+ |

## Runtime Detection

Use [`checkWebRTCSupport()`](/click-to-call/sdk/api-reference/check-click-to-call-support) to verify browser compatibility at runtime before creating a client:

```typescript theme={null}
import { checkWebRTCSupport } from '@cognigy/click-to-call-sdk';

const support = checkWebRTCSupport();
if (!support.supported) {
  console.error('Missing WebRTC features:', support.missing);
} else {
  console.log('WebRTC is fully supported');
}
```

## More Information

* [checkWebRTCSupport](/click-to-call/sdk/api-reference/check-click-to-call-support)
* [Security](/click-to-call/sdk/security)
* [Troubleshooting — Click To Call Not Supported](/click-to-call/sdk/troubleshooting#click-to-call-not-supported)


# Troubleshooting
Source: https://docs.cognigy.com/click-to-call/sdk/troubleshooting



This page covers common issues you may encounter when using the Cognigy Click To Call SDK and how to resolve them.

## Connection Issues

<AccordionGroup>
  <Accordion title="Client Fails to Connect">
    **Symptom**: Calling `connect()` or `connectAndCall()` throws an error or times out.

    **Possible Causes and Solutions**:

    * **Incorrect endpoint URL**. Verify that the `endpointUrl` in your configuration is correct and accessible. The URL must point to a valid Cognigy Click To Call configuration endpoint.
    * **Network restrictions**. Ensure that your network allows WebSocket connections and UDP traffic, both of which are required for SIP/WebRTC communication. Check firewall and proxy settings.
    * **CORS issues**. If the SDK can't fetch the SIP configuration, the endpoint server may not be configured to allow cross-origin requests from your domain.
  </Accordion>

  <Accordion title="SIP Registration Fails">
    **Symptom**: The `registered` event is never emitted, or the `error` event emits after connecting.

    **Possible Causes and Solutions**:

    * **Invalid SIP credentials**. The SIP configuration returned by the endpoint URL may contain invalid credentials. Contact your Cognigy administrator to verify the configuration.
    * **SIP server unreachable**. Ensure that the SIP server is reachable from the client's network. Network firewalls may block the required ports.
  </Accordion>
</AccordionGroup>

## Call Issues

<AccordionGroup>
  <Accordion title="Call Not Starting">
    **Symptom**: `startCall()` throws an error or doesn't initiate a call.

    **Possible Causes and Solutions**:

    * **Client not connected**. Ensure that `connect()` has been called and the `registered` event has been received before calling `startCall()`. Alternatively, use `connectAndCall()` to combine both steps.
    * **Microphone permission denied**. The browser requires microphone access for Click To Call calls. Ensure that the user has granted microphone permission.
  </Accordion>

  <Accordion title="No Audio">
    **Symptom**: The call connects, but no audio is heard.

    **Possible Causes and Solutions**:

    * **Autoplay policy**. Modern browsers block autoplay of media with audio. Ensure that the call is initiated by a user gesture, for example, a button click. The SDK manages audio playback internally via an `Audio` element, but browser autoplay restrictions still apply.

    * **ICE/TURN configuration**. If audio doesn't work, it may be a NAT traversal issue. Configure a TURN server in the `pcConfig` option:

      ```typescript theme={null}
      const client = await createWebRTCClient({
        endpointUrl: 'https://your-api.com/voice-connect-config',
        pcConfig: {
          iceServers: [
            { urls: 'stun:stun.l.google.com:19302' },
            {
              urls: 'turn:your-turn-server.com:3478',
              username: 'user',
              credential: 'password'
            }
          ]
        }
      });
      ```

    * **No audio tracks in stream**. The remote stream may not contain audio tracks. Check the browser developer console for warnings from the SDK such as "Stream has no audio tracks".
  </Accordion>

  <Accordion title="Call Drops Unexpectedly">
    **Symptom**: The [`ended`](/click-to-call/sdk/event-reference/ended) or [`failed`](/click-to-call/sdk/event-reference/failed) event emits shortly after the call starts.

    **Possible Causes and Solutions**:

    * **Network instability**. Check the network connection. Click To Call is sensitive to network changes, especially on mobile devices.
    * **Session timeout**. The SIP server may have a session timeout configuration. Check the `CallEndInfo` object in the `ended` or `failed` event callback for the `cause`, `originator`, and `description` fields.
  </Accordion>
</AccordionGroup>

## Transcription Issues

<AccordionGroup>
  <Accordion title="Transcription Events Not Received">
    **Symptom**: The [`transcription`](/click-to-call/sdk/event-reference/transcription) event is never emitted during a call.

    **Possible Causes and Solutions**:

    * **Transcription not enabled**. Ensure that transcription is enabled in your Voice Gateway or Cognigy Flow configuration.
    * **Info messages without `_transcription`**. Transcription events are only emitted when an incoming info message contains a `_transcription` property. Regular info messages are emitted via [`infoReceived`](/click-to-call/sdk/event-reference/info-received) instead.
  </Accordion>
</AccordionGroup>

## Browser Issues

<AccordionGroup>
  <Accordion title="Click To Call Not Supported">
    **Symptom**: [`checkWebRTCSupport()`](/click-to-call/sdk/api-reference/check-click-to-call-support) returns `{ supported: false }`.

    **Possible Causes and Solutions**:

    * **Unsupported browser**. The SDK requires a browser that supports WebRTC APIs: `RTCPeerConnection`, `getUserMedia`.
    * **Insecure context**. WebRTC APIs are only available in [secure contexts](https://developer.mozilla.org/en-US/docs/Web/Security/Secure_Contexts) such as HTTPS or `localhost`. Ensure your application is served over HTTPS.
  </Accordion>

  <Accordion title="Microphone Permission Issues">
    **Symptom**: The browser prompts for microphone access but the call fails.

    **Possible Causes and Solutions**:

    * **Permission denied**. The user must grant microphone permission. If denied, the browser won't allow access to the local audio stream.
    * **Permission previously blocked**. If the user previously blocked microphone access, they need to manually reset the permission in the browser's site settings.
  </Accordion>
</AccordionGroup>

## Error Handling

The SDK supports two complementary approaches for catching errors.

<AccordionGroup>
  <Accordion title="Promise-Based">
    All async methods return promises. Wrap them in the `try` and `catch` blocks:

    ```typescript theme={null}
    try {
      await client.connect();
      await client.startCall();
    } catch (error) {
      console.error('Operation failed:', error.message);
    }
    ```
  </Accordion>

  <Accordion title="Event-Based">
    Subscribe to the [`error`](/click-to-call/sdk/event-reference/error) event to catch asynchronous errors during the SDK lifecycle:

    ```typescript theme={null}
    client.on('error', (error) => {
      console.error('Click To Call SDK error:', error.message);
    });
    ```
  </Accordion>
</AccordionGroup>

<Tip>
  Combine both approaches. The `try` and `catch` blocks capture errors from direct calls, while the `error` event handles asynchronous errors.
</Tip>

## Debugging Tips

<AccordionGroup>
  <Accordion title="Enable Event Logging">
    Register listeners for all key events to trace the SDK lifecycle:

    ```typescript theme={null}
    client.on('connecting', () => console.log('[WebRTC] Connecting...'));
    client.on('connected', () => console.log('[WebRTC] Connected'));
    client.on('registered', () => console.log('[WebRTC] Registered'));
    client.on('disconnected', () => console.log('[WebRTC] Disconnected'));
    client.on('answered', (s) => console.log('[WebRTC] Answered:', s.id));
    client.on('ended', (s, info) => console.log('[WebRTC] Ended:', info));
    client.on('failed', (s, info) => console.error('[WebRTC] Failed:', info));
    client.on('error', (err) => console.error('[WebRTC] Error:', err.message));
    ```
  </Accordion>

  <Accordion title="Inspect WebRTC Internals">
    In Chrome, navigate to `chrome://webrtc-internals/` to view detailed information about active WebRTC connections, including ICE candidates, codec negotiation, and media statistics.
  </Accordion>

  <Accordion title="Check Network Traffic">
    Use the browser's DevTools **Network** tab to inspect WebSocket connections and verify that the SIP signaling messages are being exchanged correctly.
  </Accordion>
</AccordionGroup>


# Click to Call Tutorials
Source: https://docs.cognigy.com/click-to-call/tutorials/overview



These tutorials help you set up the [Click To Call widget](/click-to-call/overview) for specific use cases beyond the default setup.

<CardGroup>
  <Card title="Add a Text Input Field" href="/click-to-call/tutorials/text-input-field">
    Set up a Click To Call widget with a text input field. With the text input field, users can accurately type specific data, for example, numbers and names, during a voice call.
  </Card>

  <Card title="Add a Selector Field" href="/click-to-call/tutorials/selector-field">
    Set up a Click To Call widget with a selector field. With the selector field, users can select an option during a voice call.
  </Card>
</CardGroup>


# Click To Call with a Selector Field
Source: https://docs.cognigy.com/click-to-call/tutorials/selector-field



By the end of this guide, you will have learned how to set up a [Click To Call widget](/click-to-call/overview) that includes a selector field. The selector field will be displayed as an alternative to the voice input field to select predefined options. A selector field can be useful during voice conversations when the end user needs to select from a set of options, for example, when choosing a menu in a cantina setting.

## Prerequisites

* Configure an [AI Agent](/ai/agents/develop/manage-ai-agents).
* Configure a [speech provider](/ai/agents/test/voice-preview#set-up-a-speech-provider) in the [Voice Preview Settings](/ai/administer/access/project-settings#voice-preview-settings) section of the [Project settings](/ai/administer/access/project-settings).
* [Voice Gateway](/voice-gateway/getting-started#set-up-voice-gateway-for-your-organization) set up for your organization.
* Edit access to the HTML of the page where you will embed the widget.
* For testing purposes, allow microphone access in your browser and, if needed, in your device settings.

## Setup Overview

1. Configure a [Flow](/ai/agents/develop/projects-and-flows/overview) with an AI Agent Node and a Question Node. The Flow handles the conversation and opens the selector field on the Click To Call widget.
2. Configure a Voice Gateway Endpoint. The Voice Gateway Endpoint connects the Flow to Voice Gateway and activates the Click To Call widget.
3. Embed and test the widget. Add the widget script to your page and see the Click To Call widget working.

## Set Up Click To Call Widget with Selector Field

<Steps>
  <Step title="Create a Flow">
    1. Go to **Build > Flows** and click **+ New Flow**.
    2. On the **New Flow** panel, configure the following:

       * **Name** — add a unique name, for example, `Click To Call Assistant`.
       * (Optional) **Description** — add a relevant description, for example, `This Flow supports a Click To Call widget and a selector field`.
  </Step>

  <Step title="Configure an AI Agent Node">
    1. In the Flow editor, add an AI Agent Node and configure the following:

       * **AI Agent** — select the AI Agent you've previously created, for example, `Cantina Assistant`.
       * **Job Name** — `Cantina Specialist`. Save the Node.

    2. Next to the AI Agent Node, click **+** and select **Tool**.

    3. Configure the Tool Node as follows:

       * **Tool ID** — enter `select_menu`.
       * **Description** — enter `Trigger this tool as soon as the customer needs help selecting a menu`. Save the Node.
  </Step>

  <Step title="Configure a Say Node to Open the Selector">
    1. Below the `select_menu` child Node of the AI Agent Node, add a Say Node.
    2. Configure the Say Node as follows:

       * In the **Options** section, add the following code in the **Data** field:

       ```json theme={null}
       {
         "openInput": true
       }
       ```

       * In the **Settings** section, enter `Open selector` in the **Label** field. This label allows you to better identify the function of this Node.

       This Node triggers the event to open the selector field on the Click To Call widget.
    3. Add a Wait for Input Node to wait for the user to select or say an option.
  </Step>

  <Step title="Configure a Say Node to Close the Selector">
    1. Below the Wait for Input Node, add a Say Node and configure the following:

       * In the **Options** section, add the following code in the **Data** field:

       ```json theme={null}
       {
         "closeInput": true
       }
       ```

       * *(Optional)* In the **Settings** section, enter `Close selector` in the **Label** field.

       This Node closes the selector field on the widget after the user has answered.
  </Step>

  <Step title="Add a Resolve Tool Node to Return the Result to the AI Agent">
    Below the Wait for Input Node, add a Resolve Tool Node and configure the following:

    * **Answer Text** — enter `The user selected the following menu: {{ci.text}}`.

    This Node returns the user's menu to the AI Agent.
  </Step>
</Steps>

After completing these steps, your Flow should look like the following:

<Accordion title="Complete Flow">
  <Frame>
    <img alt="Complete Flow with AI Agent, Say, Wait for Input, and Resolve Tool Nodes" />
  </Frame>
</Accordion>

## Configure a Voice Gateway Endpoint

<Steps>
  <Step title="Create a Voice Gateway Endpoint">
    1. Go to **Deploy > Endpoints**.

    2. Click **+ New Endpoint**, select **Voice Gateway** and configure the following:

       * **Name** — for example, `Click To Call Cantina Support`.
       * **Flow** — Select the Flow you previously configured, in this example, `Click To Call Assistant`.

    3. Click **Save**.
  </Step>

  <Step title="Generate the Click To Call Embedding Code">
    1. In the Endpoint settings, find **Click To Call Embedding HTML**.
    2. Click the **Set Up Click To Call Integration** button in that field to generate the embedding code in the code editor. Hover over the code editor and click the **Copy to clipboard** button. You will use this code to embed the Click To Call widget in your website.
  </Step>

  <Step title="Enable and Configure the Click To Call Widget">
    1. In **Click To Call Widget Settings**:

       * **Enable Click To Call Widget** — activate the toggle.
       * **AI Agent Name** — Name shown on the widget, for example, `Cantina Support`.
       * **Tagline** — Short text under the name, for example, `Let me know what you want to eat!`.
       * *(Optional)* **AI Agent Avatar Logo URL** — Add a URL of the image shown on the widget, for example, `https://www.<domain>.com/logo.png`. If empty, the Cognigy logo is used.
       * *(Optional)* **Theme** — select the color theme of the widget.

    2. Save the Endpoint.
  </Step>
</Steps>

## Embed and Test Widget

<Steps>
  <Step title="Embed the Widget in your Website">
    1. Paste the code you copied previously from the Endpoint settings into your page's `<body>` or before `</body>`. The code looks similar to the following:

    ```html expandable theme={null}
    <script src="https://github.com/Cognigy/click-to-call-widget/releases/latest/download/webRTCWidget.js"></script>
    <script>
        addEventListener("load", (event) => {
            if (window.initWebRTCWidget) {
                window.initWebRTCWidget(
                    "<ENDPOINT_URL>/<YOUR_ENDPOINT_CONFIG_TOKEN>"
                )
            }
        });
    </script>
    ```

    Where `<ENDPOINT_URL>` and `<YOUR_ENDPOINT_CONFIG_TOKEN>` are the URL and the Endpoint configuration token, respectively. For example, `https://endpoint-dev.cognigy.ai/ab8b929039b427fed7ee84bb799acd7f254fc9254be27c87a78fc8a70fb048ec` for the development environment or `https://endpoint.cognigy.ai/ab8b929039b427fed7ee84bb799acd7f254fc9254be27c87a78fc8a70fb048ec` for the production environment.

    2. Make sure the actual Endpoint configuration URL is the same as the one you copied from **Click To Call Embedding HTML** in the Voice Gateway Endpoint.
    3. Edit the widget script to add the dropdown menu and events to open and close the selector when the Flow sends `openInput: true` or `closeInput: true` in the Say Nodes. The following code is an example of how to do this:

    ```html expandable theme={null}
    <script src="https://github.com/Cognigy/click-to-call-widget/releases/latest/download/webRTCWidget.js"></script>
    <script>
        addEventListener("load", function () {
            if (typeof window.initWebRTCWidget !== "function") return;
            var token = "<ENDPOINT_URL>/<YOUR_ENDPOINT_CONFIG_TOKEN>";
            window.initWebRTCWidget(token).then(function (widget) {

                var currentSession = null;

                // Returns the widget container element (tries vc_, voice_connect_, and webrtc_ class names).
                function getWidgetContainer() {
                    return document.querySelector(".vc_widget_container") ||
                        document.querySelector(".voice_connect_widget_container") ||
                        document.querySelector(".webrtc_widget_container") ||
                        document.querySelector("[class*='widget_container']");
                }

                // Creates and shows a selector (dropdown) inside the transcript container or below the widget. Options: Mediterranean, Light, Comfort, Vegetarian. On Submit, sends the value via session.sendInfo() and hides the selector.
                function showSelectorField(session) {
                    currentSession = session;
                    var wrap = document.getElementById("widget-selector-wrap");
                    if (wrap) {
                        wrap.style.display = "block";
                        return;
                    }
                    var widgetEl = getWidgetContainer();
                    if (!widgetEl) widgetEl = document.body;
                    wrap = document.createElement("div");
                    wrap.id = "widget-selector-wrap";
                    wrap.style.cssText = "width:100%;max-width:460px;margin-top:8px;margin-bottom:8px;box-sizing:border-box;display:block;";
                    var select = document.createElement("select");
                    select.id = "widget-menu-select";
                    select.style.cssText = "width:100%;padding:8px 12px;font-size:14px;border:1px solid #ccc;border-radius:6px;box-sizing:border-box;margin-bottom:8px;";
                    select.setAttribute("aria-label", "Choose a menu");
                    var placeholder = document.createElement("option");
                    placeholder.value = "";
                    placeholder.textContent = "Select your menu";
                    select.appendChild(placeholder);
                    [{ value: "Mediterranean", label: "Mediterranean" }, { value: "Light", label: "Light" }, { value: "Comfort", label: "Comfort" }, { value: "Vegetarian", label: "Vegetarian" }].forEach(function (opt) {
                        var option = document.createElement("option");
                        option.value = opt.value;
                        option.textContent = opt.label;
                        select.appendChild(option);
                    });
                    var btn = document.createElement("button");
                    btn.type = "button";
                    btn.textContent = "Submit";
                    btn.style.cssText = "width:100%;padding:8px 12px;font-size:14px;border:1px solid #ccc;border-radius:6px;box-sizing:border-box;cursor:pointer;";
                    btn.setAttribute("aria-label", "Submit");
                    function submitSelection() {
                        var value = select.value ? select.value.trim() : "";
                        if (!value) return;
                        if (currentSession && typeof currentSession.sendInfo === "function") currentSession.sendInfo(value);
                        hideSelectorField();
                    }
                    select.addEventListener("change", function () { btn.disabled = !select.value; });
                    btn.disabled = true;
                    btn.addEventListener("click", submitSelection);
                    wrap.appendChild(select);
                    wrap.appendChild(btn);
                    if (widgetEl === document.body) {
                        widgetEl.appendChild(wrap);
                    } else {
                        var transcriptContainer = document.querySelector(".vc_transcript_container") ||
                            document.querySelector(".voice_connect_transcript_container") ||
                            document.querySelector(".webrtc_transcript_container") ||
                            document.querySelector("[class*='transcript_container']");
                        if (transcriptContainer) {
                            transcriptContainer.appendChild(wrap);
                        } else {
                            var stack = widgetEl.parentElement;
                            var transcriptWrapper = stack && (stack.querySelector(".vc_widget_transcript_wrapper") ||
                                stack.querySelector(".voice_connect_widget_transcript_wrapper") ||
                                stack.querySelector(".webrtc_widget_transcript_wrapper"));
                            if (transcriptWrapper) {
                                transcriptWrapper.insertAdjacentElement("afterend", wrap);
                            } else {
                                var contentContainer = widgetEl.querySelector(".vc_widget_content_container") ||
                                    widgetEl.querySelector(".voice_connect_widget_content_container") ||
                                    widgetEl.querySelector(".webrtc_widget_content_container");
                                if (contentContainer) {
                                    contentContainer.insertAdjacentElement("afterend", wrap);
                                } else {
                                    widgetEl.appendChild(wrap);
                                }
                            }
                        }
                    }
                }

                // Hides the selector and clears the current session reference.
                function hideSelectorField() {
                    currentSession = null;
                    var wrap = document.getElementById("widget-selector-wrap");
                    if (wrap) wrap.style.display = "none";
                }

                // Returns true if the INFO body contains openInput: true (Flow asks to show the selector).
                function hasOpenInput(body) {
                    if (body == null) return false;
                    if (typeof body === "string") {
                        try { body = JSON.parse(body); } catch (e) { return false; }
                    }
                    var infoBody = typeof body === "object" ? body : null;
                    if (!infoBody) return false;
                    if (infoBody.openInput === true) return true;
                    if (infoBody.data && infoBody.data.openInput === true) return true;
                    if (infoBody.data && infoBody.data.data && infoBody.data.data.openInput === true) return true;
                    return false;
                }

                // Returns true if the INFO body contains closeInput: true (Flow asks to close the selector).
                function hasCloseInput(body) {
                    if (body == null) return false;
                    if (typeof body === "string") {
                        try { body = JSON.parse(body); } catch (e) { return false; }
                    }
                    var infoBody = typeof body === "object" ? body : null;
                    if (!infoBody) return false;
                    if (infoBody.closeInput === true) return true;
                    if (infoBody.data && infoBody.data.closeInput === true) return true;
                    if (infoBody.data && infoBody.data.data && infoBody.data.data.closeInput === true) return true;
                    return false;
                }

                // Runs when a new call (RTC session) starts.
                widget.on("newRTCSession", function (session) {

                    // Runs when the Flow sends a SIP INFO: show the selector if openInput: true, hide it if closeInput: true.
                    session.on("newInfo", function (ev) {
                        if (ev.originator !== "remote") return;
                        var info = ev.info || ev;
                        var body = info && (info.body !== undefined ? info.body : info.request && info.request.body);
                        if (body === undefined && typeof ev.body !== "undefined") body = ev.body;
                        if (body == null) return;
                        try {
                            if (hasCloseInput(body)) { hideSelectorField(); return; }
                            if (hasOpenInput(body)) showSelectorField(session);
                        } catch (e) {}
                    });

                    // Runs when the call ends normally; hide the selector.
                    session.on("ended", hideSelectorField);

                    // Runs when the session is terminated; hide the selector.
                    session.on("terminated", hideSelectorField);
                });
            });
        });
    </script>
    ```
  </Step>

  <Step title="Test the Widget">
    1. Open the page in a browser, allow microphone access when prompted, and click <img alt="start voice call" /> on the widget to start a voice conversation with your AI Agent.
    2. After the AI Agent greets you, say you want to select a menu. The selector field opens on the widget.
    3. Try answering the question using the dropdown (Mediterranean, Light, Comfort, Vegetarian) and by voice.
  </Step>
</Steps>

## See Also

* [Click To Call](/click-to-call/overview)
* [Voice Gateway Endpoint](/ai/agents/deploy/endpoint-reference/voice-gateway)
* [Flows](/ai/agents/develop/projects-and-flows/overview)
* [AI Agent Node](/ai/agents/develop/node-reference/ai/ai-agent)
* [Manage AI Agents](/ai/agents/develop/manage-ai-agents)


# Click To Call with a Text Input Field
Source: https://docs.cognigy.com/click-to-call/tutorials/text-input-field



By the end of this guide, you will have learned how to set up a [Click To Call widget](/click-to-call/overview) that includes a text input field. The text input field will be displayed as an alternative to the voice input field. A text field can be useful during voice conversations when the end user needs to provide specific data precisely, for example, for copying and pasting a long tracking number or making sure a name is spelled correctly.

<div>
  <video title="Click To Call widget with text input field demonstration" aria-label="Click To Call widget showing user interaction and including a text input field">
    <source />
  </video>
</div>

## Prerequisites

* Configure an [AI Agent](/ai/agents/develop/manage-ai-agents).
* Configure a [speech provider](/ai/agents/test/voice-preview#set-up-a-speech-provider) in the [Voice Preview Settings](/ai/administer/access/project-settings#voice-preview-settings) section of the [Project settings](/ai/administer/access/project-settings).
* [Voice Gateway](/voice-gateway/getting-started#set-up-voice-gateway-for-your-organization) set up for your organization.
* Edit access to the HTML of the page where you will embed the widget.
* For testing purposes, allow microphone access in your browser and, if needed, in your device settings.

## Setup Overview

1. Configure a [Flow](/ai/agents/develop/projects-and-flows/overview) with an AI Agent Node and a tool to find packages. The Flow handles the conversation and allows the user to find their package by typing in the text input field.
2. Configure a Voice Gateway Endpoint. The Voice Gateway Endpoint connects the Flow to Voice Gateway and activates the Click To Call widget.
3. Embed and test the widget. Add the widget script to your page and see the Click To Call widget working.

## Set Up a Click To Call Widget with a Text Input Field

<Steps>
  <Step title="Create a Flow">
    1. Go to **Build > Flows** and click **+ New Flow**.
    2. On the **New Flow** panel, configure the following:

       * **Name** — add a unique name, for example, `Click To Call Assistant`.
       * (Optional) **Description** — add a relevant description, for example, `This Flow supports a Click To Call widget and an input text field`.
  </Step>

  <Step title="Configure an AI Agent Node">
    1. In the Flow editor, add an AI Agent Node and configure the following:

       * **AI Agent** — select the AI Agent you've previously created, for example, `Package Support Assistant`.
       * **Job Name** — `Package Support Specialist`. Save the Node.

    2. Next to the AI Agent Node, click **+** and select **Tool**.

    3. Configure the Tool Node as follows:

       * **Tool ID** — enter `find_package`.
       * **Description** — enter `Trigger this tool as soon as the customer needs help finding their packages.`. Save the Node.
  </Step>

  <Step title="Configure a Say Node to Open the Input Field">
    1. Below the Default child Node of the AI Agent Node, add a Say Node.
    2. Configure the Say Node as follows:

       * In the **Options** section, add the following code in the **Data** field:

       ```json theme={null}
       {
         "openInput": true
       }
       ```

       * In the **Settings** section, enter `Open input field` in the **Label** field. This label allows you to better identify the function of this Node.

       This Node triggers the event to open the input field on the Click To Call widget.
  </Step>

  <Step title="Configure a Question Node">
    1. Below the `find_package` child Node, add a Question Node and configure the following:

       * **Question Type** — select **Number**.
       * **Text** — enter the following text: `What is your tracking number?`.
  </Step>

  <Step title="(Optional) Store the Result in the Context Object">
    1. Below the Question Node, add an Add To Context Node and configure the following:

       * **Key** — enter `trackingNumber`.
       * **Value** — enter `{{ci.slots?.NUMBER?}}`.

       With this CognigyScript expression, the Node stores the user's answer from the input field under `trackingNumber` in the Context object. You can use this value for further processing. For example, you can send this to your package management system and track the delivery status.
  </Step>

  <Step title="Add a Resolve Tool Node to Return the Result to the AI Agent">
    Below the Add To Context Node, add a Resolve Tool Node and configure the following:

    * **Answer Text** — enter `The user's package number is {{cc.trackingNumber? || ci.text}}`.

    This Node returns the user's tracking number to the AI Agent.
  </Step>
</Steps>

After completing these steps, your Flow should look like the following:

<Accordion title="Complete Flow">
  <Frame>
    <img alt="Complete Flow with AI Agent, Say, Question, Add To Context, and Resolve Tool Nodes" />
  </Frame>
</Accordion>

## Configure a Voice Gateway Endpoint

<Steps>
  <Step title="Create a Voice Gateway Endpoint">
    1. Go to **Deploy > Endpoints**.

    2. Click **+ New Endpoint**, select **Voice Gateway** and configure the following:

       * **Name** — for example, `Click To Call Support`.
       * **Flow** — Select the Flow you previously configured, in this example, `Click To Call Assistant`.

    3. Click **Save**.
  </Step>

  <Step title="Generate the Click To Call Embedding Code">
    1. In the Endpoint settings, find **Click To Call Embedding HTML**.
    2. Click the **Set Up Click To Call Integration** button in that field to generate the embedding code in the code editor. Hover over the code editor and click the **Copy to clipboard** button. You will use this code to embed the Click To Call widget in your website.
  </Step>

  <Step title="Enable and Configure the Click To Call Widget">
    1. In **Click To Call Widget Settings**:

    * **Enable Click To Call Widget** — activate the toggle.
    * **AI Agent Name** — Name shown on the widget, for example, `Package Support`.
    * **Tagline** — Short text under the name, for example, `Find your package!`.
    * *(Optional)* **AI Agent Avatar Logo URL** — Add a URL of the image shown on the widget, for example, `https://www.<domain>.com/logo.png`. If empty, the Cognigy logo is used.
    * *(Optional)* **Theme** — select the color theme of the widget.

    2. Save the Endpoint.
  </Step>
</Steps>

## Embed and Test Widget

<Steps>
  <Step title="Embed the Widget in your Website">
    1. Paste the code you copied previously from the Endpoint settings into your page's `<body>` or before `</body>`. The code looks similar to the following:

    ```html expandable theme={null}
    <script src="https://github.com/Cognigy/click-to-call-widget/releases/latest/download/webRTCWidget.js"></script>
    <script>
        addEventListener("load", (event) => {
            if (window.initWebRTCWidget) {
                window.initWebRTCWidget(
                    "https://endpoint-dev.cognigy.ai/<YOUR_ENDPOINT_CONFIG_TOKEN>"
                )
            }
        });
    </script>
    ```

    2. Make sure the actual Endpoint configuration URL is the same as the one you copied from **Click To Call Embedding HTML** in the Voice Gateway Endpoint.
    3. Edit the script to add the text input field and events to open and close the input field when the Flow sends `openInput: true` in the Say Node. The following code is an example of how to do this:

    ```html expandable theme={null}
    <script src="https://github.com/Cognigy/click-to-call-widget/releases/latest/download/webRTCWidget.js"></script>
    <script>
        addEventListener("load", function () {
            if (typeof window.initWebRTCWidget !== "function") return;
            var token = "<ENDPOINT_URL>/<YOUR_ENDPOINT_CONFIG_TOKEN>";
            window.initWebRTCWidget(token).then(function (widget) {

                var currentSession = null;

                // Returns the widget container element (tries vc_, voice_connect_, and webrtc_ class names).
                function getWidgetContainer() {
                    return document.querySelector(".vc_widget_container") ||
                        document.querySelector(".voice_connect_widget_container") ||
                        document.querySelector(".webrtc_widget_container") ||
                        document.querySelector("[class*='widget_container']");
                }

                // Creates and shows a text input below the transcript (or below the widget content). On Enter, sends the value to the Flow via session.sendInfo().
                function showInputField(session) {
                    currentSession = session;
                    var wrap = document.getElementById("widget-input-wrap");
                    if (wrap) {
                        wrap.style.display = "block";
                        var input = document.getElementById("widget-text-input");
                        if (input) input.focus();
                        return;
                    }
                    var widgetEl = getWidgetContainer();
                    if (!widgetEl) widgetEl = document.body;
                    wrap = document.createElement("div");
                    wrap.id = "widget-input-wrap";
                    wrap.style.cssText = "width:100%;max-width:460px;margin-top:8px;margin-bottom:8px;box-sizing:border-box;display:block;";
                    var input = document.createElement("input");
                    input.type = "text";
                    input.id = "widget-text-input";
                    input.placeholder = "Type a message and press Enter...";
                    input.style.cssText = "width:100%;padding:8px 12px;font-size:14px;border:1px solid #ccc;border-radius:6px;box-sizing:border-box;";
                    input.addEventListener("keydown", function (event) {
                        if (event.key !== "Enter") return;
                        event.preventDefault();
                        var value = (event.target.value || "").trim();
                        if (!value) return;
                        if (currentSession && typeof currentSession.sendInfo === "function") currentSession.sendInfo(value);
                        event.target.value = "";
                    });
                    wrap.appendChild(input);
                    if (widgetEl === document.body) {
                        widgetEl.appendChild(wrap);
                    } else {
                        var stack = widgetEl.parentElement;
                        var transcriptWrapper = stack && (stack.querySelector(".vc_widget_transcript_wrapper") ||
                            stack.querySelector(".voice_connect_widget_transcript_wrapper") ||
                            stack.querySelector(".webrtc_widget_transcript_wrapper"));
                        if (transcriptWrapper) {
                            transcriptWrapper.insertAdjacentElement("afterend", wrap);
                        } else {
                            var contentContainer = widgetEl.querySelector(".vc_widget_content_container") ||
                                widgetEl.querySelector(".voice_connect_widget_content_container") ||
                                widgetEl.querySelector(".webrtc_widget_content_container");
                            if (contentContainer) {
                                contentContainer.insertAdjacentElement("afterend", wrap);
                            } else {
                                widgetEl.appendChild(wrap);
                            }
                        }
                    }
                    input.focus();
                }

                // Hides the text input and clears the current session reference.
                function hideInputField() {
                    currentSession = null;
                    var wrap = document.getElementById("widget-input-wrap");
                    if (wrap) wrap.style.display = "none";
                }

                // Returns true if the INFO body contains openInput: true (Flow asks to show the text input).
                function hasOpenInput(body) {
                    if (body == null) return false;
                    if (typeof body === "string") {
                        try { body = JSON.parse(body); } catch (e) { return false; }
                    }
                    var infoBody = typeof body === "object" ? body : null;
                    if (!infoBody) return false;
                    if (infoBody.openInput === true) return true;
                    if (infoBody.data && infoBody.data.openInput === true) return true;
                    if (infoBody.data && infoBody.data.data && infoBody.data.data.openInput === true) return true;
                    return false;
                }

                // Runs when a new call (RTC session) starts.
                widget.on("newRTCSession", function (session) {

                    // Runs when the Flow sends a SIP INFO: show the input if openInput: true.
                    session.on("newInfo", function (ev) {
                        if (ev.originator !== "remote") return;
                        var info = ev.info || ev;
                        var body = info && (info.body !== undefined ? info.body : info.request && info.request.body);
                        if (body === undefined && typeof ev.body !== "undefined") body = ev.body;
                        if (body == null) return;
                        try {
                            if (hasOpenInput(body)) showInputField(session);
                        } catch (e) {}
                    });

                    // Runs when the call ends normally; hide the input.
                    session.on("ended", hideInputField);

                    // Runs when the session is terminated; hide the input.
                    session.on("terminated", hideInputField);
                });
            });
        });
    </script>
    ```
  </Step>

  <Step title="Test the Widget">
    1. Open the page in a browser, allow microphone access when prompted, and click <img alt="start voice call" /> on the widget to start a voice conversation with your AI Agent.
    2. After the AI Agent greets you, the input field opens.
    3. Say you want to find your package.
    4. Try answering the question in the text input field and by voice.
  </Step>
</Steps>

## See Also

* [Click To Call](/click-to-call/overview)
* [Voice Gateway Endpoint](/ai/agents/deploy/endpoint-reference/voice-gateway)
* [Flows](/ai/agents/develop/projects-and-flows/overview)
* [AI Agent Node](/ai/agents/develop/node-reference/ai/ai-agent)
* [Manage AI Agents](/ai/agents/develop/manage-ai-agents)


# Welcome to Cognigy Academy
Source: https://docs.cognigy.com/help/cognigy-academy

Explore Cognigy Academy and Partner Academy to access AI Agent training courses, certifications, and hands-on learning for conversational AI solutions.

Cognigy offers two dedicated learning platforms designed to help you master Conversational AI — whether you're implementing solutions for clients or managing your own customer experience.

<Card title="🤝 Cognigy Partner Academy" href="https://partneracademy.cognigy.com">
  <p>For our valued partners, including:</p>

  <ul>
    <li>Implementation partners</li>
    <li>Technology partners</li>
    <li>System integrators</li>
    <li>Resellers</li>
  </ul>

  <p>Access specialized enablement programs, technical certifications, and best practices to help you deliver exceptional conversational AI solutions for your clients.</p>
</Card>

<Card title="🎓 Cognigy Academy" href="https://academy.cognigy.com">
  <p><strong>For customers and users of Cognigy.</strong></p>
  <p>Learn how to design, build, and manage AI Agents that transform your customer experience.</p>
  <p>Discover courses on building AI Agents, accessing APIs, analytics, and more — all tailored to your success with Cognigy.</p>
</Card>


# Get Help & Report Problems
Source: https://docs.cognigy.com/help/get-help

Learn how to get help with Cognigy by submitting support tickets, using in-product help, and accessing other resources to resolve issues quickly.

Whenever you feel stuck or encounter a problem on our platform, don't hesitate to ask for help. There are many ways to get help, some of which are described on this page.

## In-Product Contextual Help

Inside Cognigy, you can find contextual in-product help by clicking the Help icon in the App Bar on the upper right of the screen.

<Frame>
  <img alt="Help button in the Cognigy app bar" />
</Frame>

When clicking on this menu, the help widget will open, and you can search for help directly within the product.

<Frame>
  <img alt="In-product help widget with search functionality" />
</Frame>

## Help Center

Every article in the Help Widget comes from the Cognigy Help Center, which can be reached by clicking the "View original article" link.

<Frame>
  <img alt="View original article link in the help widget" />
</Frame>

Alternatively, you can always reach the [Cognigy Help Center](https://support.cognigy.com/hc/en-us).

## Submit Support Tickets

In case you've found a bug or need technical assistance, you can file a support ticket. You can do this in three ways:

1. Directly from the Help Widget inside Cognigy by clicking **Contact US** at the bottom.
2. By registering at [https://support.cognigy.com](https://support.cognigy.com) and filing a ticket from there.
3. By sending an email to [support@cognigy.com](mailto:support@cognigy.com).

## Ticket Content Recommendations

For a better and faster handling of your ticket, you should provide as much information as you can give us.

The following sections describe recommended attachments and information you could send us to accelerate the ticket handling process.

## General Problems

For an issue which you can't really associate with a specific resource (flow, endpoint etc.) add at least a screenshot of the console (**F12**) as an attachment to the ticket.

## Problems with Flows

The ticket for a problem with a flow should have the Flow and all its dependencies attached.

You can also attach a screenshot of the [logs](/ai/agents/test/logs) to the ticket.

## Problems with the UI

In addition to your description, add a screenshot of the bug to your ticket.


# Cognigy Support
Source: https://docs.cognigy.com/help/index



Use the Help Center and other resources from Cognigy. If they don't help, contact Cognigy technical support for assistance with troubleshooting.

<CardGroup>
  <Card title="Technical Support" icon="fire-extinguisher" href="/help/get-help">
    Contact Cognigy Support for help troubleshooting any issues you encounter while using Cognigy products.

    <br />
  </Card>

  <Card title="Help Center" icon="book" href="https://support.cognigy.com/hc/en-us">
    Explore solutions, best practices, and tutorials based on real user questions. Find answers and resolve issues quickly.

    <br />
  </Card>

  <Card title="Trust and Compliance" icon="shield" href="https://trust.cognigy.com/">
    Learn about Cognigy's approach to security, ethics, and compliance processes.

    <br />

    <a href="https://trust.cognigy.com/">Trust Center</a> • <a href="https://www.digital-dsb.eu/home_en/">Compliance Reporting</a>
  </Card>

  <Card title="Cognigy Academy" icon="book-a" href="/help/cognigy-academy">
    Master Conversational AI with Cognigy Academy — learn to implement solutions for clients or enhance your own customer experience.
  </Card>

  <Card title="Other Resources" icon="info" href="https://www.cognigy.com/">
    Discover more about Cognigy — read our [blog](https://www.cognigy.com/blog), watch [videos](https://www.youtube.com/c/Cognigy), and check out our open-source projects.

    <br />

    <a href="https://www.cognigy.com/blog">Cognigy Blog</a> • <a href="https://www.youtube.com/c/Cognigy">YouTube Channel</a> • <a href="https://www.cognigy.com/">Website</a> • <a href="https://github.com/Cognigy">GitHub</a>
  </Card>
</CardGroup>


# Cognigy
Source: https://docs.cognigy.com/index

Welcome to the Cognigy documentation

<div>
  <div>
    Cognigy is a platform designed for large-scale enterprises to create, deploy, and manage chat and voice AI Agents.

    <p>
      The platform leverages advanced NLU technology and LLMs to develop conversational interfaces that can accurately understand and interpret user queries.
    </p>

    <p>
      Cognigy provides a seamless and intuitive user experience that can be deployed across various communication channels such as websites, mobile apps, and messaging platforms.
    </p>
  </div>

  <div>
    <video aria-label="Cognigy hero animation showing AI Agent capabilities: memory, knowledge, tools, and progressive reasoning.">
      <source type="video/mp4" />
    </video>

    <video aria-label="Cognigy hero animation showing AI Agent capabilities: memory, knowledge, tools, and progressive reasoning.">
      <source type="video/mp4" />
    </video>

    <button type="button" aria-label="Pause video">
      ⏸
    </button>
  </div>
</div>

<CardGroup>
  <Card title="Overview" icon="book-open" href="/ai/overview/about-cognigy-ai">
    Introduction to Cognigy AI Agents and their capabilities.
  </Card>

  <Card title="Quickstart Guide" icon="rocket" href="/ai/agents/quickstart">
    Step-by-step instructions to set up your AI Agents swiftly.
  </Card>

  <Card title="Build" icon="hammer" href="/ai/agents/develop/projects-and-flows/projects">
    Learn how to construct and customize your AI Agents effectively.
  </Card>

  <Card title="Deploy" icon="cloud-upload" href="/ai/agents/deploy/deploying-your-ai-agents">
    Guidelines for deploying your AI Agents across various platforms.
  </Card>

  <Card title="Test" icon="flask-conical" href="/ai/agents/test/testing-your-ai-agents">
    Methods to test and validate the performance of your AI Agents.
  </Card>

  <Card title="Analyze" icon="chart-line" href="/ai/agents/analyze/overview">
    Tools and techniques to analyze AI Agent interactions and data.
  </Card>

  <Card title="Handover Providers" icon="headset" href="/ai/escalate/handovers">
    Information on integrating human handover providers with AI Agents.
  </Card>

  <Card title="Platform Features" icon="star" href="/ai/platform-features/overview">
    Explore the comprehensive features available on the Cognigy platform.
  </Card>

  <Card title="Administration and Installation" icon="settings" href="/ai/administer/overview">
    Install Cognigy and administer your organizations
  </Card>
</CardGroup>


# Basic Analysis
Source: https://docs.cognigy.com/insights/conversation-analyzer/basic-analysis

Learn how the Basic Analysis dashboard in Conversation Analyzer scores conversations against predefined criteria.

<a href="/release-notes/2026.14"><Badge>Updated in 2026.14 (beta)</Badge></a>

<Note>
  This feature is in beta. We encourage you to try it out and provide us with feedback.
</Note>

*Basic Analysis* scores conversations against predefined criteria, grouped into four sets. Use this dashboard to monitor customer sentiment, containment rates, AI behavior quality, and conversational experience quality across your conversations.

<Frame>
  <img alt="Basic Analysis Dashboard Overview" />
</Frame>

## When to Use Basic Analysis

* **After each analysis run** — review how your AI Agent performs against quality and containment benchmarks.
* **When escalation rates increase** — identify the top reasons customers are handed over to human agents.
* **After updating your AI Agent** — verify that metrics such as Instruction Following and Tool Usage have improved.
* **During periodic QA reviews** — spot drops in sentiment, empathy, or resolution rate that require attention.

## Restrictions

* Basic Analysis uses predefined criteria. To evaluate custom business-specific behaviors, use [Custom Evaluation](/insights/conversation-analyzer/custom-evaluation).

## Configuration

To deactivate specific criteria:

1. In the left-side menu of the Insights interface, go to **Configuration**.
2. Go to the **Basic Analysis Criteria** section.
3. Deactivate criteria sets you don't want to include. Save changes.

<Frame>
  <img alt="Basic Analysis Criteria Config" />
</Frame>

## Criteria and Charts

### Customer Sentiment

Analyzes the emotional state of the customer across the conversation.

| Criterion           | Description                                                                                                                                |
| ------------------- | ------------------------------------------------------------------------------------------------------------------------------------------ |
| Overall Sentiment   | Shows the overall sentiment of customer conversations.                                                                                     |
| Sentiment Trend     | Shows how customer sentiment changes over time.                                                                                            |
| Emotional Intensity | Shows the emotional intensity of customer sentiment. Helps identify conversations with strong emotions that may require special attention. |
| Escalation Risk     | Shows the escalation risk of customer conversations. Helps identify conversations that may require proactive intervention.                 |

**Example:** Analysis of a credit card dispute AI Agent shows 38% of conversations have high Emotional Intensity, and Escalation Risk is elevated where customers have contacted support more than twice in the same week. The team uses this to prioritize proactive callbacks.

### Containment & Success

Measures how often the AI Agent resolves customer issues without handover to a human agent.

| Criterion             | Description                                                                                     |
| --------------------- | ----------------------------------------------------------------------------------------------- |
| Resolution Rate       | Shows the percentage of conversations successfully resolved by the AI Agent without escalation. |
| Escalations Triggered | Shows the percentage of conversations escalated to human agents.                                |
| Escalation Reasons    | Shows the percentage of the top reasons for escalation.                                         |

**Example:** A loan inquiry AI Agent shows a 71% resolution rate. Escalation Reasons reveal that 48% of escalations are triggered by human requests, suggesting customers frequently prefer agent assistance for complex financial topics.

### AI Behaviour Quality

Evaluates how well the AI Agent followed its instructions and used tools correctly.

| Criterion              | Description                                                                                                                                                                             |
| ---------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Instruction Following  | Shows how well the AI Agent follows the provided instructions in conversations.                                                                                                         |
| Tool Usage Correctness | Shows how correctly the AI Agent uses tools in conversations, based on the transcript. The Conversation Analyzer doesn't access underlying tool specifications or actual API responses. |

**Example:** Instruction Following drops to 6.1 out of 10 after a policy update. The team identifies that the AI Agent is still citing outdated interest rates and updates the knowledge base.

### AI Agent Experience Quality

Assesses the conversational quality of the AI Agent. Politeness, Empathy, Professional Tone, and Response Clarity metrics are displayed in a radar chart.

**Example:** A mortgage advisor AI Agent scores high on Politeness (8.8) and Professional Tone (9.1), but low on Empathy (6.1). Flagged conversations show the AI Agent rarely acknowledges customer frustration before moving to problem-solving.


# Custom Evaluation
Source: https://docs.cognigy.com/insights/conversation-analyzer/custom-evaluation

Learn how to define your own scoring criteria in Conversation Analyzer using widget types such as Yes/No, 3-option scale, 5-option scale, percentage, and numeric score.

<a href="/release-notes/2026.12"><Badge>Added in 2026.12 (beta)</Badge></a>

<Note>
  This feature is in beta. We encourage you to try it out and provide us with feedback.
</Note>

*Custom Evaluation* lets you define your own scoring criteria in addition to Basic Analysis. The custom evaluation allows you to evaluate behavior specific to your business domain, compliance requirements, or product context.

<Frame>
  <img alt="Custom Evaluation Dashboard Overview" />
</Frame>

## When to Use Custom Evaluation

* **When predefined criteria aren't enough** — evaluate business-specific behaviors not covered by Basic Analysis.
* **For compliance monitoring** — create criteria to track whether AI Agents follow regulatory or policy requirements specific to your industry.
* **After launching new products or services** — create criteria to measure how well the AI Agent handles new offerings.
* **During A/B testing** — compare AI Agent behavior across different prompt or instruction variants.

## Restrictions

* You can add up to 10 custom evaluation criteria.

## Configuration

1. In the left-side menu of the Insights interface, go to **Configuration**.

2. In the **Custom Criteria** section, click **+ Add Custom Criterion** and select a widget from the **Widget Type** list:

   <Tabs>
     <Tab title="Yes / No">
       A binary criterion with a pass or fail outcome. Use this for simple checks where you want to know whether a specific behavior occurred.

       * **Title** — a short display name for the criterion. For example, `Accurate Product Information`.
       * **Instructions for the LLM** — enter a statement that the LLM must evaluate and return as pass or fail. For example, `The agent provides accurate information about financial products, rates, fees, or terms without contradicting known product documentation`.

       The widget will be rendered as a ring chart on the Custom Evaluation dashboard, showing the percentage of conversations that passed versus failed this criterion.
     </Tab>

     <Tab title="3-Option Scale">
       A scored criterion with three labeled outcome levels mapped to scores 1, 2, and 3. Score 1 is the lowest outcome, score 3 is the highest.

       * **Title** — a short display name for the criterion. For example, `Regulatory Compliance`.
       * **Instructions for the LLM** — enter a statement that the LLM must evaluate and return as a score from 1 to 3. For example, `The agent includes required regulatory disclosures appropriate to the financial topic discussed. Score 1 if no disclosures were made, 2 if disclosures were incomplete or misplaced, 3 if all required disclosures were provided correctly`.
       * **Option Labels** — three labels for scores 1, 2, and 3. For example, `Not compliant / Partially compliant / Fully compliant`.

       The widget will be rendered as a bar chart on the Custom Evaluation dashboard, showing the distribution of conversations across the three outcome levels.
     </Tab>

     <Tab title="5-Option Scale">
       A scored criterion with five labeled outcome levels mapped to scores 1 through 5. Use this when you need finer-grained scoring than the 3-option scale.

       * **Title** — a short display name for the criterion. For example, `Accurate Product Information`.
       * **Instructions for the LLM** — enter a statement that the LLM must evaluate and return as a score from 1 to 5. For example, `How accurately the agent describes financial products, rates, fees, or terms compared to known product documentation? Score 1 if completely wrong, 2 if mostly incorrect with some accurate details, 3 if roughly half accurate, 4 if mostly correct with minor errors, 5 if fully accurate evaluation`.
       * **Option Labels** — five labels for scores 1 through 5. For example, `Completely inaccurate / Mostly inaccurate / Partially accurate / Mostly accurate / Fully accurate`.

       The widget will be rendered as a bar chart on the Custom Evaluation dashboard, showing the distribution of conversations across the five outcome levels.
     </Tab>

     <Tab title="Percentage">
       A criterion scored as a value between 0 and 100%. Use this for proportional measures across multiple instances within a conversation.

       * **Title** — a short display name for the criterion. For example, `Escalation Appropriateness`.
       * **Instructions for the LLM** — enter a statement that the LLM must evaluate and return as a percentage. For example, `What percentage of escalation decisions in this conversation were appropriate? Consider whether escalations to a human advisor were justified and whether cases that could have been self-served were incorrectly escalated. Return a value between 0 and 100`.

       The widget will be rendered as an indicator chart on the Custom Evaluation dashboard, showing the average percentage score across conversations.
     </Tab>

     <Tab title="Numeric Score">
       A criterion scored as a number on a scale you define. Use this when you want a continuous score with maximum flexibility.

       * **Title** — a short display name for the criterion. For example, `Sensitive Data Handling`.
       * **Instructions for the LLM** — enter a statement that the LLM must evaluate and return as a numeric score. For example, `Rate how well the agent handled sensitive data on a scale of 0 to 10. Score 0 if sensitive data was mishandled, 5 if handling was adequate but inconsistent, 10 if sensitive data was managed correctly and securely throughout the entire conversation`.

       The widget will be rendered as an indicator chart on the Custom Evaluation dashboard, showing the average numeric score across conversations.
     </Tab>
   </Tabs>

3. Fill in the required fields for the selected widget type. Save changes.

<Frame>
  <img alt="Custom Criteria Config" />
</Frame>

## Tips for Writing LLM Instructions

The quality of your custom criteria results depends directly on how clearly you write the LLM instructions.

* **Be specific.** Vague instructions produce inconsistent scores. Instead of "Was the agent helpful?", describe exactly what behavior to look for and what each score level means.
* **Define the scale explicitly.** For numeric and scale-based criteria, always describe what the lowest and highest scores represent.
* **Use plain language.** Write instructions the way you would explain the task to a colleague, not in technical jargon.
* **One criterion, one thing.** Each criterion should evaluate a single, well-defined behavior. Combining multiple checks in one criterion makes scores harder to interpret.
* **Test with known conversations.** After adding a custom criterion, run a manual analysis on conversations where you already know the expected outcome to verify the scoring behaves as intended.


# Conversation Analyzer
Source: https://docs.cognigy.com/insights/conversation-analyzer/overview

Learn how Conversation Analyzer automatically scores your AI Agent's conversations based on configurable criteria across Topic Discovery, Basic Analysis, and Custom Evaluation dashboards.

<a href="/release-notes/2026.18"><Badge>Added in 2026.18 (beta)</Badge></a>

<Note>
  This feature is in beta. We encourage you to try it out and provide us with feedback.
</Note>

*Conversation Analyzer* is a quality evaluation tool in Cognigy Insights that automatically scores your AI Agent's conversations based on predefined and custom criteria. Instead of reviewing [conversation transcripts](/insights/explorers/conversations) manually, you define an evaluation topic and a set of metrics, and the analyzer processes your conversations and surfaces results across dashboards.

<div>
  <video title="Conversation Analyzer Overview" aria-label="The video shows an overview of Conversation Analyzer's three dashboards: Topic Discovery, Basic Analysis, and Custom Evaluation. The video demonstrates how to configure evaluation settings, run analysis on conversations, and check results across the dashboards.">
    <source />
  </video>
</div>

## When to Use it

* Understand what topics your AI Agent handles most frequently and how that changes over time.
* Measure whether your AI Agent is resolving issues, following instructions, and communicating with quality.
* Define custom evaluation criteria specific to your business, compliance requirements, or product domain.
* Replace manual QA spot-checks with consistent, automated scoring at scale.

## Key Benefits

* **Reduce manual QA effort**. Define criteria once and the analyzer scores conversations automatically at scale, replacing time-consuming manual transcript reviews.
* **Catch issues earlier**. Scheduled runs surface declining sentiment trends, rising escalation risk, or drops in instruction-following before they become visible in support volumes.
* **Consistent evaluation**. Unlike manual reviews, the analyzer applies the same criteria to every conversation and every run, helping reduce reviewer bias and inconsistency.
* **Identify coverage gaps**. Find topics your AI Agent doesn't cover and expand its scope accordingly.

## Prerequisites

<Accordion title="Set Up your Model">
  Conversation Analyzer uses a large language model to evaluate conversations. Before running your first analysis, configure which model to use in your Project settings.

  <Tip>
    Reasoning models can provide more accurate results, but they typically come with higher costs and longer processing times. For brief analyses, non-reasoning models may be sufficient.
  </Tip>

  1. In Cognigy, open your Project.
  2. Go to **Manage > Settings > Generative AI Settings**.
  3. Under **Conversation Analyzer**, select a model from the list. Save changes.

  If you don't select a model, the default model is used.
</Accordion>

## Restrictions

* Conversation Analyzer doesn't support the following models:
  * OpenAI — `gpt-5-6-sol`, `gpt-5-6-terra`, `gpt-5-6-luna`
  * Microsoft Azure OpenAI — `gpt-5-6-sol`, `gpt-5-6-terra`, `gpt-5-6-luna`
  * Anthropic — `claude-sonnet-4-6`
  * Google GenAI — `gemini-3.5-flash`, `gemini-3.6-flash`, `gemini-3.7-flash`, `gemini-3.1-flash-lite`, `gemini-3.1-pro-preview`, `gemini-3-flash-preview`

## Working with Conversation Analyzer

<AccordionGroup>
  <Accordion title="1. Configure Evaluation Settings">
    In the left-side menu of the Insights interface, go to **Configuration** and set up the evaluation topic, schedule, and criteria:

    * [Topic Discovery](/insights/conversation-analyzer/topic-discovery#configuration) — select an evaluation topic that reflects the main subject area of your AI Agent's conversations.
    * [Basic Analysis](/insights/conversation-analyzer/basic-analysis#configuration) — activate or deactivate criteria sets in Basic Analysis based on what you want to measure.
    * [Custom Evaluation](/insights/conversation-analyzer/custom-evaluation#configuration) — define custom evaluation criteria specific to your business needs.
  </Accordion>

  <Accordion title="2. Run Analysis">
    You can run analysis on conversations in the following ways:

    <Tabs>
      <Tab title="On a Schedule">
        To schedule an analysis, follow these steps:

        1. In the left-side menu of the Insights interface, go to **Configuration**.
        2. In the **Topic & Scheduler** section, go to **Analysis Scheduler**.
        3. By default, the scheduler is off. Select how often you want the analysis to run:
           * **Daily** — runs once a day, analyzing conversations from the previous 24 hours.
           * **Weekly** — runs once a week, analyzing conversations from the previous 7 days.
             Each run includes only conversations that haven't yet been analyzed. The time window is calculated from when the scheduled run executes, not from when you save the scheduler settings.
        4. In the **Filter only Analyze Transcripts Matching** section, configure which conversations to include:
           * **Endpoint** — limit analysis to a specific Endpoint.
           * **Snapshot** — limit to a specific Snapshot.
           * **Channel** — limit to a specific channel.
        5. Save changes.

        The **Analysis Scheduler** section shows when the next run is scheduled. For example, `Next run: May 15, 2026, 2:00 AM`.
      </Tab>

      <Tab title="Manually">
        To run an analysis manually, follow these steps:

        1. In the left-side menu of the Insights interface, go to **Conversations**.
        2. In the filter bar, select the scope of your analysis. You can filter by time span, Endpoint, and more.
        3. Click **Run Analysis**. Select one of the following options:

           | Option                    | Description                                                                                                                       |
           | ------------------------- | --------------------------------------------------------------------------------------------------------------------------------- |
           | **Reanalyze all**         | Analyzes all conversations, including those already analyzed. Overwrites previous LLM analysis results.                           |
           | **Skip already analyzed** | Analyzes only conversations that haven't been analyzed yet. Adds new data to the dashboards without overwriting existing results. |

        The **Analyzing conversations** progress bar appears at the top of the Conversations page.
        4\. When the analysis is complete, click **View Analysis** in the upper-right corner. The **Analysis** page opens.
      </Tab>
    </Tabs>
  </Accordion>

  <Accordion title="3. Check Results">
    You can check the results for:

    <Tabs>
      <Tab title="All Conversations">
        In the left-side menu of the Insights interface, go to **Analysis** to view topics, performance metrics, and evaluation results.
      </Tab>

      <Tab title="A Specific Conversation">
        In the left-side menu of the Insights interface, go to **Conversations**. Select the entry from the table you want to review, then click **>** to open **Evaluation Details**.
      </Tab>
    </Tabs>
  </Accordion>
</AccordionGroup>


# Topic Discovery
Source: https://docs.cognigy.com/insights/conversation-analyzer/topic-discovery

Learn how to use the Topic Discovery dashboard in Conversation Analyzer to understand how conversations are distributed across topics, identify emerging patterns, and keep your topic hierarchy up to date.

<a href="/release-notes/2026.14"><Badge>Updated in 2026.14 (beta)</Badge></a>

<Note>
  This feature is in beta. We encourage you to try it out and provide us with feedback.
</Note>

*Topic Discovery* shows how conversations are distributed across topics after each analysis run. Use it to understand what your AI Agent handles, monitor emerging patterns, and keep your topic hierarchy current.

<Frame>
  <img alt="Topic Discovery Dashboard Overview" />
</Frame>

## When to Use Topic Discovery

* **After each analysis run** — check whether your topic hierarchy still reflects what customers are talking about. If not, consider selecting a broader or more specific topic.
* **When mapped coverage drops** — identify conversations falling outside your current topic set.
* **When a new product or feature launches** — monitor whether it generates new conversation topics that aren't covered yet.
* **During periodic QA reviews** — spot topic overlap and consolidate redundant topics. Then update your AI Agent logic for more structured conversation handling. The next analysis run will reflect the changes in the topic hierarchy.
* **When AI suggestions appear** — evaluate whether the analyzer has detected emerging subtopics worth adding.

## Restrictions

* A conversation can fall under multiple subtopics simultaneously, so subtopic counts may exceed the parent topic count.

## Configuration

1. In the left-side menu of the Insights interface, go to **Configuration**.
2. In the **Topic & Scheduler** section, open the **Evaluation Topic** list.
3. Select the topic you want to evaluate, for example, Customer Service. Save changes. The LLM automatically detects subtopics such as **Account Management** or **Billing Disputes** based on conversation content. You don't need to define subtopics manually.

   <Frame>
     <img alt="Topic Discovery Config" />
   </Frame>

## Charts

### Top Rising Topic

Shows the topic with the highest increase in conversation volume between the last two analysis runs. The number displayed reflects how many conversations matched this topic in the most recent run.

Use this to catch sudden spikes in a specific topic before they become a support issue.

**Example:** Between Monday and Tuesday's runs, conversations about fraud alerts jumped from 12 to 89. Top Rising Topic surfaces this immediately, prompting the team to investigate whether a phishing campaign is driving the spike.

### Mapped Coverage

Shows the percentage of conversations covered by the topics the LLM identified within the [selected evaluation topic](#configuration). A high percentage means most conversations are being categorized. A low percentage suggests your topics may not reflect what customers are actually discussing.

**Example:** Mapped Coverage shows 74%. The remaining 26% represents conversations that matched none of the topics in the current hierarchy, suggesting other topics may need to be analyzed.

### Topic Overlap

Shows the overlap between different topics based on conversation content. High overlap between two topics may indicate they are too similar, or that a parent-child relationship between them would be more accurate.

**Example:** The card disputes and fraud alerts topics show 61% overlap, suggesting fraud alerts could be restructured as a subtopic under card disputes.

### New Topics Found

Shows the number of conversations that didn't match any of your selected topics. Reviewing this after each run helps you identify subject areas your topic hierarchy doesn't cover yet. Adding these topics improves mapped coverage over time.

**Example:** After a product update introducing a new savings tier, New Topics Found increases by 43. Reviewing these conversations reveals customers are asking about premium savings. This topic isn't in the hierarchy yet.

### Topic Hierarchy

Displays all topics found in your conversations, organized by topic, subtopic group, and subtopic. Each topic shows how many conversations it covers.

* A single conversation can match multiple subtopics, so subtopic counts may exceed the topic count. This behavior is expected, not an error.
* AI-suggested topics appear beneath their parent topic. Click **+** next to a suggestion to add it to the hierarchy. The **+** button appears only when the LLM has suggestions for the [selected topic](#configuration).

  <Frame>
    <img alt="An AI-suggested topic appeared beneath the parent topic" />
  </Frame>

**Example:** The Loans topic shows 320 conversations. Its subtopics show: Personal Loans (180), Mortgage Inquiries (95), Early Repayment (74). The total exceeds 320 because some conversations covered more than one subtopic.


# Business Dashboard
Source: https://docs.cognigy.com/insights/dashboards/business

Monitor real-time session metrics, location activity, active AI and human agent data, and business outcomes.

<a href="/release-notes/2026.1"><Badge>Updated in 2026.1</Badge></a>

<Note>
  Note that the Goals-related charts are available only if you have migrated to the new [Insights framework](/insights/data-management/pre-aggregation).
</Note>

The *Business Dashboard* gives you a quick overview of your key session metrics through interactive charts and tables.
It includes location-based activity, active sessions by AI or human agents, and business outcomes, such as time and cost savings.

The summary of key metrics is represented in the following charts:

* [Interactive Map](#interactive-map)
* [Tables](#tables)
* [Bar Charts](#bar-charts)
* [Indicators](#indicators)

All charts display data based on the filter settings you've selected. The charts display data recorded within the last 15 minutes. Some charts refresh every 30 seconds.

<div>
  <video title="Business dashboard overview" aria-label="The video shows the Business dashboard: a rotating globe displaying session activity by city and country, a list of top session locations, an active sessions indicator, and metrics for time saved, cost savings, revenue earned, and top goals.">
    <source />
  </video>
</div>

## Prerequisites

If you have an on-premises installation, add the following feature flags to the `values.yaml` file:

* `FEATURE_INSIGHTS_USE_BUSINESS_DASHBOARD: "true"`
* `FEATURE_INSIGHTS_BUSINESS_DASHBOARD_DIRECT_POSTGRESQL: "true"`. Add this feature flag if you've migrated to the new Insights framework.

## Filters

In the top bar, unlike other reports and explorers, you can select multiple Projects to view their data on the dashboard. However, if you switch between Projects and then open reports or explorers, the **Time Span** filter will reset for them. This reset happens because of the 15-minute time window limitation for data display on the dashboard.

Additionally, you can use the following filters:

| Filter   | Description                                                                                                                                                                                                                                                            |
| -------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Location | You can select one of the following options: <ul><li>**City** – the default option, filters data by city.</li><li>**Country** – filters data by country.</li></ul>                                                                                                     |
| Channel  | You can select one of the following filters: <ul><li>**All** – the default filter, includes all sessions.</li><li>**Voice** – filters sessions created only via the Voice Gateway Endpoint.</li><li>**Chat** – filters sessions created via other Endpoints.</li></ul> |

### Reset Filters

To reset all filters, at the top of the page, click the **Reset** button on the right side of the filter panel.

If the **Reset** button is not displayed,
click <img alt="filter" /> in the upper-right corner of the page.

## Interactive Map

Displays the number of sessions by country or city.

To view session data for a specific location, use the **Location** filter to select a country or city. Then, hover over a location on the globe to see the number of sessions.

<div>
  <video title="Interactive session map showing geographical distribution of conversations" aria-label="The video shows a rotating 3D globe with bars marking session volume at each location, updating in real time as conversations occur.">
    <source />
  </video>
</div>

You can navigate the globe using the following controls:

<Accordion title="Controls">
  | Icon                     | Name       | Description                       |
  | ------------------------ | ---------- | --------------------------------- |
  | <img alt="play" /> ️     | Play       | Starts the animation.             |
  | <img alt="stop" />       | Stop       | Stops the animation.              |
  | <img alt="fullscreen" /> | Fullscreen | Expands the globe to full screen. |
  | <img alt="zoom out" />   | Zoom out   | Zooms out to show a wider area.   |
  | <img alt="zoom in" />    | Zoom in    | Zooms in to show more detail.     |

  The chart shows changes within the last 15 minutes and updates every 30 seconds.
</Accordion>

## Tables

### Top Sessions

Displays the total number of current sessions by country or city, including both ongoing sessions and those started in the last 15 minutes.

To view session data for a specific location, use the **Location** filter to switch between country and city views.

#### Unknown Locations

If a city or country is marked as `Unknown`, it means Insights couldn't determine where the session occurred.

In the city view:

* If the entry has a flag, the country is known, but the specific city isn't.
* If there's a question mark instead of a flag, both the city and country are unknown.

In the country view, all known countries will be listed. There will be only one `Unknown` entry to represent all sessions with unidentified locations.

### Top 5 Goals

Displays the top 5 goals completed in the last 15 minutes, allowing you to track recent performance quickly.

## Bar Charts

### Active Sessions

Displays the total number of sessions started within the last 15 minutes, refreshed every 30 seconds.

The chart breaks down sessions handled by AI Agents versus those escalated to human agents, enabling you to track how many sessions AI Agents resolved independently.

<Frame>
  <img alt="Bar chart comparing sessions handled by AI Agents versus those escalated to human agents" />
</Frame>

## Indicators

### Time Saved

Displays the total time saved by summing the time saved from completed and partially completed goals in the last 15 minutes.

### Cost Savings

Displays the total cost savings for completed and partially completed goals in the last 15 minutes.

### Revenue Earned

Displays the total revenue earned for completed and partially completed goals in the last 15 minutes.

## More Information

* [Time Zone with Pre-aggregation](/insights/data-management/time-zone-with-pre-aggregation)


# Data Expiration in Insights
Source: https://docs.cognigy.com/insights/data-management/data-expiration

Configure analytics data retention in Cognigy Insights. Set data expiration globally or per organization to meet compliance and storage requirements.

<a href="/release-notes/4.59"><Badge>Updated in 4.59</Badge></a>

The *Data Expiration* feature helps organizations manage data retention by allowing you to view and change the remaining time until stored data is deleted.

You can configure this feature globally or per organization to meet compliance and storage requirements.

Daily use of Cognigy products resets the data retention countdown, keeping your data from expiring.
For example, if the data expiration is set to 30 days and a user sends a message on May 1, the data will be retained until May 31; if they send another message on May 15, the expiration shifts to June 14.

## Key Benefits

* **Improved Data Compliance**. Helps organizations follow laws and regulations by automatically deleting data after a set retention period.
* **Optimized Storage Management**. Prevents unnecessary data buildup by removing expired data, which lowers storage costs and keeps the system running smoothly.
* **Flexible Configuration for SaaS and On-Premises**. Allows you to set data expiration rules globally or individually for each organization, whether you use SaaS or on-premises installations.

## How to Use

### Activate Data Expiration

By default, Cognigy keeps analytics data to a maximum of 1 year. You can reduce this data expiration period as follows:

* For on-premises installations, [set data expiration values](/ai/administer/access/management-ui#change-expiration-values) in the [Management UI](/ai/administer/access/management-ui).
* For SaaS installations, contact [Cognigy technical support](/help/get-help#help-center) and specify the limits in your request.

The data expiration feature applies only to new data obtained by Cognigy after activating data expiration in the Management UI. Any datasets obtained before setting the expiration value remain unaffected from this change.

<Accordion title="Example">
  You set a 1-minute expiration value at 13:00 for Sessions in the Management UI.

  All Sessions created after 13:00 will automatically expire after 1 minute.

  Sessions created before 13:00 will remain unchanged.
</Accordion>

### Check when Data Expires

To check when data expires, click <img alt="check data expiration" /> in the upper-right corner of the Insights interface.
To view the exact time in minutes, hover over the displayed value.

The table shows the [data from the Analytics database](/ai/agents/analyze/collecting-data) covered by data expiration and illustrates how they map to the data displayed in reports and explorers.

<Accordion title="View the table">
  | Data                | Description                                                                                            | Reports and Explorers                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                |
  | ------------------- | ------------------------------------------------------------------------------------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
  | Analytics           | The time span between the creation of all Analytics data and its deletion from the Analytics database. | [Overview](/insights/reports/overview): <br />- Number of Inputs <br />- Top Intents <br /> [Engagement](/insights/reports/engagement): <br />- Avg. Session Length<br /> - Max. Session Length <br /> [NLU Performance](/insights/reports/nlu-performance): <br /> - Avg Intent Score <br /> - Average Execution Time <br /> - Intents by Score Range <br /> - Max Intent Score <br /> - Top Intents <br /> - Top Intent Scores <br />                                                                                                              |
  | Conversations       | The time span between the creation of chat history analytics and its deletion from the database.       | [Conversations explorer](/insights/explorers/conversations)                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          |
  | Sessions            | The time span between the creation of Insights session analytics and its deletion from the database.   | [Overview](/insights/reports/overview): <br /> - Active Sessions<br /> - Positive Ratings<br /> - Sessions<br /> - Session by Channel<br /> - Total Sessions<br /> - Top Tasks<br />- Top Slots<br /> - Unique Contacts<br /> - Users by Locale<br /> [Engagement](/insights/reports/engagement): <br />- Contacts <br />- Goals Reached <br />- Handovers by time of the day <br />- Live Agent Escalations <br />- Positive Ratings <br />- Positive Ratings over Time<br />[NLU Performance](/insights/reports/nlu-performance):<br />- Top Slots |
  | Step Events         | The time span between the creation of Step data and its deletion from the database.                    | -                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
  | LiveAgentEscalation | The time span between the creation of Live Agent data and its deletion from the database.              | [Live Agent](/insights/reports/live-agent)                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
  | voice-gatewayEvents | The time span between the creation of Voice Gateway data and its deletion from the database.           | -                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
  | GoalsV2Events       | The time span between the creation of Goals data and its deletion from the database.                   | [Goals](/insights/reports/goals)                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
</Accordion>

### Change Data Expiration

You can change data expiration values as follows:

* For on-premises Cognigy installations, you can adjust data expiration values via the [Management UI](/ai/administer/access/management-ui).
* For SaaS installations, contact [technical support](/help/get-help#help-center).

When you adjust the data expiration value, it applies only to new data obtained by Cognigy. The data expiration value for previously obtained data remains unchanged.

### Customize Data Expiration for Multiple Organizations

If you manage multiple organizations, you can customize the data expiration settings for each organization independently through the Management UI.
The default data expiration value is set in the environment variable in the `values.yaml` file.

<Note>
  The Management UI value can override the environment variable for an organization.
  This override happens only if the Management UI value is positive and less than the environment variable.

  This setup balances global defaults with organization-specific requirements.
</Note>

## More Information

* [Data Expiration Cognigy](/ai/agents/analyze/data-expiration)


# Data Pre-Aggregation
Source: https://docs.cognigy.com/insights/data-management/pre-aggregation

Data pre-aggregation in Cognigy Insights pre-computes tables for faster report visualization, reducing query times and enabling instant filter application.

<Note>
  We are introducing a new architecture framework for Cognigy Insights, which significantly enhances the application's data processing capacity. This upgraded framework not only improves overall performance but also minimizes loading times for report data. To use this framework, migration is required.

  The migration to the new framework isn't linked to a specific release. We will start with a few SaaS customers, inform them, and provide assistance for migration. For on-premises customers, the migration guide will be provided soon.

  Afterward, we will migrate the remaining SaaS customers.
</Note>

*Data pre-aggregation* is a specific type of data aggregation that optimizes query performance and reduces computational resources by aggregating data before it's stored.

After migrating to the new framework, Insights uses pre-aggregation,
which means that it generates pre-aggregated tables based on raw data.
This approach enables instant visualization of charts on reports, even when applying filters.
Updating pre-aggregated tables takes some time, so despite the fact that data on reports is always displayed,
data refreshing may take some time depending on the source data.
An estimated data refresh time will be indicated on your report in the Insights UI and can take up to 2 hours.

## More Information

* [Time Zone with Pre-aggregation](/insights/data-management/time-zone-with-pre-aggregation)


# Time Zone
Source: https://docs.cognigy.com/insights/data-management/time-zone

Configure time zone preferences in Cognigy Insights to view reports in your local time.

<Warning>
  This page explains how the Time Zone works before we start migrating to the [new Insights framework](/insights/data-management/time-zone-with-pre-aggregation). If the migration process has not yet started for your organization, this page remains relevant for you.
</Warning>

In Insights, *Time zone* settings play a crucial role in presenting time-based data accurately to users based on their geographical location.
By selecting the appropriate time zone, you can view data in your local time, which is essential for accurate data analysis and interpretation.

Insights supports different time zones with various time offsets. The complete list of time zones is available in your personal profile in the Cognigy interface. Report data is generated on the fly. However, if you choose a rarely used time zone and have a large amount of processed data, report information may not load promptly and could experience delays.

By default, Insights relies on the time zone preferences of your device according to supported time zones.

## Change Time Zone Settings

By default, Insights uses time zone preferences of your device. You can override this in **My Profile**.

To change time zone preferences, follow these steps:

1. In the upper-right corner of the **Main** page, select **User Menu > My Profile**.
2. In the **Preferences** section, select time zone from the **Select Insights time zone** list.

New time zone will be applied to all existing data.

<Note>
  Custom time zone preferences are stored in the browser storage, which means you will need to select your custom time zone preference again if you change the device or the browser.
</Note>

## Time Zone Considerations: OData and Insights

In the OData feed, all times are consistently presented in UTC format.
When comparing time zones from OData to Insights Reports, consider this to ensure accurate data interpretation.

## More Information

* [Pre-aggregation](/insights/data-management/pre-aggregation)
* [Time Zone with Pre-aggregation](/insights/data-management/time-zone-with-pre-aggregation)


# Time Zone with Pre-Aggregation
Source: https://docs.cognigy.com/insights/data-management/time-zone-with-pre-aggregation

Learn how Cognigy Insights manages time zone settings with pre-aggregation for accurate data reporting.

<Note>
  We are introducing a new architecture framework for Cognigy Insights, which significantly enhances the application's data processing capacity. This upgraded framework not only improves overall performance but also minimizes loading times for report data. To use this framework, migration is required.

  The migration to the new framework isn't linked to a specific release. We will start with a few SaaS customers, inform them, and provide assistance for migration. For on-premises customers, the migration guide will be provided soon.

  Afterward, we will migrate the remaining SaaS customers.
</Note>

In Insights, *Time zone* settings play a crucial role in presenting time-based data accurately to users based on their geographical location.
By selecting the appropriate time zone, you can view data in your local time, which is essential for accurate data analysis and interpretation.

By default, Insights relies on the time zone preferences of your device according to supported time zones.

## Supported Time Zones

With migrating to the new framework, a significant change was made to the data-fetching process from the database.
Before the migration, the system fetched the entire dataset from the database based on the applied filters and time zone,
which often resulted in slow queries and timeouts due to the large amount of data being processed.

After migrating to the new framework, Insights introduced [pre-aggregation](/insights/data-management/pre-aggregation) to improve data retrieval performance.
Now, Insights groups data by time zones on an hourly basis, enabling quicker queries and on-the-fly data generation. However, the update has reduced the number of supported time zones to optimize efficiency and prevent timeouts.

By default, the `(UTC+00:00)` UTC time zone supports pre-aggregation.
You can include 2 extra custom time zones (recommended), with a maximum of 5.

If a previously set time zone is not supported, it will be reset to the default time zone or the nearest supported time zone.
In the Insights interface, you will see the corresponding warning.

## Custom Time Zone

The configuration of additional time zones depends on the type of your installation:

<Tabs>
  <Tab title="SaaS">
    Contact your customer success manager to add preferable custom time zones.
  </Tab>

  <Tab title="On-Premises">
    If you are using an on-premises environment, you can specify a custom time zone.
    Note that each time zone selection necessitates a separate set of pre-aggregations, which can increase the system load.

    We recommend selecting 2–3 time zones, with UTC included, as the ideal configuration. The maximum allowed limit is 5 time zones, and UTC must always be included in your value.

    Before installation or an upgrade, we strongly advise aligning the list of supported time zones with the customer's preferences. To do this, specify `CUBEJS_SCHEDULED_REFRESH_TIMEZONES=<time-zone-1>,<time-zone-2>,<time-zone-3>` in the `values.yaml` file. For example, `CUBEJS_SCHEDULED_REFRESH_TIMEZONES=UTC,CET,ACST`.
  </Tab>
</Tabs>

## Change Time Zone Settings

<Note>
  If you decide to set a [custom time zone](#custom-time-zone) preference, it will be saved in your browser storage.
  However, keep in mind that if you switch to a different device or browser, you will need to reselect your desired custom time zone preference.
</Note>

To change time zone preferences, follow these steps:

1. In the upper-right corner of the **Main** page, select **User Menu > My Profile**.
2. In the **Preferences** section, select time zone from the **Select Insights time zone** list.

New time zone will be applied to all existing data.

## Time Zone Considerations: OData and Insights

In the OData feed, all times are consistently presented in UTC format.
When comparing time zones from OData to Insights Reports, consider this to ensure accurate data interpretation.

## More Information

* [Data pre-aggregation](/insights/data-management/pre-aggregation)
* [Time Zone](/insights/data-management/time-zone)


# Download Reports in CSV
Source: https://docs.cognigy.com/insights/download-reports

Learn how to export Cognigy Insights data as CSV files for analysis in Excel or databases, covering overview, engagement, NLU, and live agent charts.

You can download detailed reports in CSV format from the charts on the Insights reports:

* [Overview report](/insights/reports/overview)
* [Engagement report](/insights/reports/engagement)
* [NLU Performance](/insights/reports/nlu-performance)
* [Live Agent](/insights/reports/live-agent)

These CSV reports contain crucial information, including intent scores, transcripts, and conversation details, enabling a better understanding of user behavior.

The CSV format ensures compatibility with various tools, making analysis seamless.
You can easily import this data into Microsoft Excel or databases like Microsoft Access.

## Download Reports via Insights UI

To download a report, follow these steps:

1. Go to Insights and select one of the previously mentioned reports in the side menu bar on the left side.
2. Click <img alt="menu" /> in the top right corner of a report chart you want to get as a report.
3. Select **Download report as .csv file**. The report will now be stored in a CSV file.

The downloaded report file contains the complete chart dataset based on the filter settings. You can open the CSV file using applications such as Microsoft Excel or similar software.

## More Information

* [Cognigy Insights](/insights/overview)


# Conversations (former Transcript Explorer)
Source: https://docs.cognigy.com/insights/explorers/conversations

Analyze sessions and their transcripts in the Conversations explorer to help you improve your AI Agent's performance.

<a href="/release-notes/2026.14"><Badge>Updated in 2026.14 (beta)</Badge></a>

The *Conversations Explorer* is an analytics tool in Cognigy Insights that helps you improve your agent's performance,
provide better customer service, and enhance the overall user experience.
It allows you to analyze transcripts of conversations between end users and agents in detail.
The Conversations explorer has a search feature that lets you look for specific phrases within a list of transcripts.
For example, you can search for keywords such as booking, canceling, or messaging to find relevant messages quickly.

<Note>
  Sessions with personal data masked using the [Delete Profile](/ai/agents/develop/node-reference/analytics/delete-profile) or [Deactivate Profile](/ai/agents/develop/node-reference/analytics/deactivate-profile) Nodes are excluded from the Conversations explorer.
</Note>

To open the Conversations explorer, use one of the following options:

<Tabs>
  <Tab title="Via Cognigy">
    1. In the left-side menu of the Cognigy interface, select a Project.
    2. Within the Project interface, navigate to Insights using one of the following options:
       * In the left-side menu, select **Insights**.
       * In the bottom-left corner, select **Insights** from the User Menu.<br />
         The Cognigy Insights Overview report will be displayed by default.
    3. In the left-side menu of the Insights interface, select **Conversations**.
  </Tab>

  <Tab title="Via Insights">
    In the left-side menu of the Insights interface, select **Conversations**. The Conversations explorer interface will be displayed.

    <Frame>
      <img alt="Conversations interface showing a filtered list of sessions" />
    </Frame>
  </Tab>
</Tabs>

## List of Sessions

By default, the Conversations explorer displays a filtered list of sessions for the last month.

The list of sessions has the following parameters:

| Title        | Description                                                                                                                                                   |
| ------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Endpoint     | The Endpoint handling the session. For example, Cognigy Webchat or the Interaction Panel.                                                                     |
| Session ID   | The unique ID for each session. <br /> To copy the Session ID, on the right side of the selected session row, click <img alt="menu" /> **> Copy session ID**. |
| User ID      | The unique ID for the end user. To copy the User ID, on the right side of the selected session row, click <img alt="menu" /> **> Copy user ID**.              |
| Messages     | The total number of messages within the specific session.                                                                                                     |
| Last Message | The date and time when the last message was received within the session.                                                                                      |
| Analysis     | Reflects whether the session has been analyzed by [Conversation Analyzer](/insights/conversation-analyzer/overview).                                          |

### Sort Sessions

Sessions can be sorted by their `Session ID`, `Messages`, or `Last Message` parameters. To select a sorting parameter, hover over a column title and select the option you prefer:

* Click the up arrow <img alt="sort ascending" /> to sort in ascending order.
* Click the down arrow <img alt="sort descending" /> to sort in descending order.

## Search for Sessions

To find specific sessions, do the following:

1. In the upper-right corner of the **Conversations** page, locate the search field.
2. In the search field, specify text messages (for more precise results, use keywords or phrases), Session ID, or User ID.

The Conversations explorer immediately updates the list of sessions according to the search results.

## Filter Sessions

There are several filters available to filter sessions. The filter bar, featuring global and local filters, is visible at the top of the page. Some filters are not immediately visible and require clicking **more filters** to be viewed and selected.
You can filter sessions by using the following options:

* [Global Filters](#global-filters)
* [Local Filters](#local-filters)

To hide the **Filter Bar**, click <img alt="filter" /> in the upper-right corner.

### Global Filters

For more information, read the [Global Filter](/insights/global-filters) documentation.

### Local Filters

Local filters are unique to a specific explorer/report. You can use them to refine your research.

To open the **Local Filters** selection,
click **more filters <img alt="filter" />** in the filter bar at the top of the **Conversations** page.

| Filter              | Description                                                                                                                                                                                                                                                                                                                        |
| ------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Message Rating      | Select the rating criteria: <br /> - **Positive** — messages perceived positively. <br />- **None** — messages with no sentiment. <br />- **Negative** — messages perceived negatively.                                                                                                                                            |
| Source              | Select one or more message sources: <br /> - **AI Agent Messages** — messages sent by AI Agents. <br /> - **User Messages** — messages sent by users. <br /> - **Agent Messages** — messages sent by human agents.                                                                                                                 |
| Goals Completed     | Select from predefined goals to narrow down results according to specific objectives or targets configured with [Tasks](/ai/agents/develop/node-reference/overview).                                                                                                                                                               |
| Flow                | Specify one or more Flows to filter messages generated during the selected Flow passing.                                                                                                                                                                                                                                           |
| Contains Step       | Refine your search results by selecting a predefined analytic step. To access analytic steps, ensure you define descriptive names in the Node settings of your Flow within Cognigy. These steps are crucial for identification and analysis within Cognigy Insights, as they are logged once Nodes are passed during interactions. |
| Message count: From | Set a minimum threshold for the number of messages in a session to be considered. Use the up and down arrow icons to specify the exact number. Sessions with fewer messages than this threshold will not be included in the results.                                                                                               |
| Message count: To   | Set a maximum threshold for the number of messages in a session to be considered. Use the up and down arrow icons to specify the exact number. Sessions with more messages than this threshold will not be included in the results.                                                                                                |

### Exclude Filters

When you toggle the **Negate** option, all filters except the selected one will be analyzed.
For example, if you select only `Webchat`
in the **Endpoint** filter and **Negate**, all data except the Webchat Endpoint will be analyzed.

### Reset Filters

To reset all filters, at the top of the page, click the **Reset** button on the right side of the filter panel.

If the **Reset** button is not displayed,
click <img alt="filter" /> in the upper-right corner of the page.

## View Session Transcripts

Session transcripts are logged conversations between end users and agents.
The Conversations explorer enables you to view session transcripts and analyze them to understand interactions between end users and agents better.
To view a session transcript, select a session from the list on the **Conversations** page.

### Session Transcript

Session transcripts display all individual messages between the end user and agent during their conversation. By selecting a session, the corresponding transcript will be displayed on the right side of the page.
You can use the scroll bar to navigate through the messages in the **Session Transcript** window.

<Note>
  The Session Transcript interface in the Conversations explorer is specifically designed for detailed analysis of past sessions, whereas the [Interaction Panel](/ai/agents/test/interaction-panel/overview) is primarily used for testing and debugging Flows.
  When you access a session transcript in the Conversations explorer, you may observe differences compared to the Interaction Panel. For example, the format and content of debug messages vary between these two tools.
</Note>

#### Debug Mode

Insights' Debug mode is not fully equal to the Debug mode in the [Interaction Panel](/ai/agents/test/interaction-panel/chat#debug-mode) documentation. Debug messages from Nodes (including Code Nodes and Debug Message Nodes) are only for immediate testing and can only be seen in the Interaction Panel.

The Debug mode offers additional context for conversation messages, including:

* Triggered Intent and its score
* Filled Slots
* User connection and disconnection events
* Voice event payloads
* xApp session details

Additionally, you can switch to a Flow related to the current session.

To use Debug mode, activate the **Debug mode** toggle at the top of the **Session Transcript** window.

#### Scenarios

You can create a [scenario](/ai/agents/test/simulator/overview) based on a session transcript. Scenarios allow you to simulate user interactions and test your Flows effectively.

<Note>
  To create a scenario, an LLM model must exist on [the LLMs page](/ai/agents/develop/gen-ai-and-llms/providers/all-providers). By default, the LLM model marked as [`default`](/ai/agents/develop/gen-ai-and-llms/other-operations#set-a-model-as-default) will be selected.
</Note>

To create a scenario from a transcript, do the following:

1. In the upper-right corner of the **Session Transcript** window, click <img alt="menu" /> to open a selection menu.
2. Select **Create Scenario**. You will be redirected to the Simulator interface, where the Scenario will be created based on the selected transcript. The persona for the scenario will be generated from scratch. Click **Create**, then [follow the instructions](/ai/agents/test/simulator/overview#2-simulate) to run the simulation.

Alternatively, you can create a scenario from a custom transcript in the [Simulator](/ai/agents/test/simulator/overview) interface.

#### Playbooks

To create a [Playbook](/ai/agents/test/playbooks) from a transcript, do the following:

1. In the upper-right corner of the **Session Transcript** window, click <img alt="menu" /> to open the selection menu.
2. Select one of the following options:
   * **Create Playbook** — generate a Playbook to test your Flows.
   * **Create Playbook with Assertions** — generate a Playbook and its Assertions, which are checks to ensure your Flows behave as expected during testing.

#### Open Message-Related Flow

To check your [Flow](/ai/agents/develop/projects-and-flows/overview) or to improve its configuration, the Conversations explorer provides the **Open Flow** context menu option.

To switch from the Conversations explorer to the corresponding Flow in Cognigy, do the following:

1. In the left-side menu of the Insights interface, select **Conversations**.
2. On the **Conversations** page, select a session from the list. The session transcript will appear in a separate window on the right side of the page.
3. In the **Session Transcript** window, right-click a specific message and select **Open Flow**. The corresponding Flow in Cognigy appears.

### Session Transcript Details

To get more session details, click <img alt="user menu" /> in the upper-left corner of the **Session Transcript** window.

The transcript details include the following sections:

* [Session Details](#session-details)
* [Contact Profile Details](#contact-profile-details)

#### Session Details

| Session Details        | Description                                                                                                                 |
| ---------------------- | --------------------------------------------------------------------------------------------------------------------------- |
| Session ID             | The unique identifier assigned to each session.                                                                             |
| First Message          | The date and time when the first message was sent.                                                                          |
| Last Message           | The date and time when the last message was sent.                                                                           |
| User Message Count     | The total number of messages sent by the user.                                                                              |
| AI Agent Message Count | The total number of messages sent by the AI Agent.                                                                          |
| Agent Message Count    | The total number of messages sent by the human agent.                                                                       |
| Flow Name              | The name of the conversation Flow in your Agent.                                                                            |
| Rating                 | The feedback received from the user in the conversation, displayed as a thumbs-up (5 to 10) or thumbs-down (0 to 4) rating. |
| Comment                | Any additional comments provided by the user along with their feedback.                                                     |

#### Contact Profile Details

| Contact Profile Details | Description                                                                                                                                                                           |
| ----------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| First Name              | User's first name.                                                                                                                                                                    |
| Last name               | User's last name.                                                                                                                                                                     |
| Email                   | User's email address.                                                                                                                                                                 |
| Gender                  | User's gender.                                                                                                                                                                        |
| Age                     | User's age.                                                                                                                                                                           |
| Birthday                | User's birthday.                                                                                                                                                                      |
| Location                | User's location.                                                                                                                                                                      |
| Accepted GDPR           | Accepted General Data Protection Regulation.                                                                                                                                          |
| Tasks Completed         | All the tasks that were completed by the user during conversations with AI Agents or human agents. The field refers to the [Tasks](/ai/agents/analyze/goals-and-tasks/tasks) feature. |

#### Evaluation Details

To view the evaluation details for a session, click <img alt="menu" /> in the upper-right corner of the **Session Transcript** window and select **Evaluation Details**.
The evaluation details include the following sections:

* [Detected Topics](/insights/conversation-analyzer/topic-discovery)
* [Basic Analysis](/insights/conversation-analyzer/basic-analysis)
* [Custom Evaluation](/insights/conversation-analyzer/custom-evaluation)

## Delete Sessions

You can delete a session and its transcript
by selecting the session from the list on the **Conversations** page
and clicking <img alt="menu" /> **> Delete Session Transcript** in the **Session Transcript Details** window.
The session will be deleted from the Transcript and Message explorers.
However, the session metadata will remain accessible through the [Cognigy OData endpoint](/ai/agents/analyze/odata), except for the data in the [Conversation](/ai/agents/analyze/odata#conversations) collection.
Also, the Contact Profile details will still be available in the [Contact Profile](/ai/agents/analyze/contact-profiles).

<Note>
  If you've migrated to the new [Insights framework](/insights/data-management/pre-aggregation), there might be a delay between deleting the session transcript and the session itself. When you click the **Delete Session Transcript** button, the transcript is deleted first, and the session is deleted from the **Conversations** page about 10 minutes later.
</Note>

## More Information

* [Insights Glossary](/insights/glossary)
* [Playbooks](/ai/agents/test/playbooks)
* [Interaction Panel](/ai/agents/test/interaction-panel/overview)


# Global Filters
Source: https://docs.cognigy.com/insights/global-filters

Use global filters in Cognigy Insights to focus reports on specific time spans, Endpoints, Snapshots, and Locales for targeted data analysis.

<a href="/release-notes/2026.13"><Badge>Updated in 2026.13</Badge></a>

*Global Filters* enable you to focus your reports and explorers on specific datasets by adjusting a set of criteria. These filters are available and effective across all reports and explorers, and they persist when you switch between them. Once you change a filter setting, the displayed graphics and charts will adapt to your selections. Reports will also reflect these changes.

<Frame>
  <img alt="Global filter bar with time span, endpoint, snapshot, and locale options" />
</Frame>

The following global filters are available:

* [Time Span](#time-span)
* [Endpoints](#endpoints)
  * [Interaction Panel and Playbook Data](#interaction-panel-and-playbook-data)
* [Snapshots](#snapshots)
* [Locales](#locales)

## Time Span

To focus on a specific time span, you can either set a preset time span or select specific dates using a date picker.

Available preset time spans include:

| Time Span     | Description                                                             |
| ------------- | ----------------------------------------------------------------------- |
| Last 7 Days   | Shows data from the past week. This time span is the default filter.    |
| Last 30 Days  | Shows data from the last 30 days.                                       |
| Last 365 Days | Shows data from the previous year. The maximum period for data storage. |
| Today         | Shows data from the current day.                                        |
| This Month    | Shows data from the current month.                                      |
| This Year     | Shows data from the current calendar year.                              |
| Custom Period | Shows data from a user-defined range of dates.                          |

If you have changed a time span, click **Apply** to confirm and activate this filter.

### Select a Custom Time Span

You can select specific hours within the chosen time span to narrow down your results to a more precise time period.
To define a custom time span, proceed as follows:

1. At the top of the Insights interface, go to the filter bar.
2. In the **Time Span** filter, select **Custom Period**.
3. Activate the **Select time** option to include hours in your selection.
4. Click **Start Date** to open a date picker calendar.
5. Select both the date and time by adjusting the clock hand to your preferred hour within the 24-hour cycle.
6. Select **End Date** to open the date picker calendar once more.
7. Similarly, select the end date and time that you desire.
8. Click **Apply** to activate the selected filter settings.

When **Select time** is turned on, the start time is included, and the end time isn't.
For example, 09:00–10:00 includes sessions starting at or after 09:00:00 and before 10:00:00. Any session starting at 10:00:00 is counted in the next time window.
This behavior only applies to the **Custom Period** setting. Date-based presets, such as **Last 7 Days** or **This Month**, are unaffected.

## Endpoints

Endpoints serve as connectors. They allow seamless interactions between Cognigy AI Agents and external systems or platforms. Examples of Endpoints include Webchat, Voice Gateway, and Microsoft Teams.
For more information, read the [Endpoints](/ai/agents/deploy/endpoints/overview) documentation.

You can select one or more Endpoints to be analyzed.

To use the **Endpoints** filter, proceed as follows:

1. At the top of the Insights interface, go to the filter bar.
2. Click **Endpoint**.
3. Use the **Type to search** field to find a specific Endpoint or select Endpoints from the list.
4. *(Optional)* Activate the **Negate** option and select Endpoints you don't want to include in your analysis. All Endpoints except the selected ones will be included in your data charts.
5. *(Optional)* To view only selected options, click <img alt="filter" /> in the **Type to search** field.
6. *(Optional)* To reset the search results, click **x** in the **Type to search** field.
7. Click **Apply** to activate the filter settings. All reports and explorers will then display the data according to your selection.

### Interaction Panel, Playbooks, and Simulations

By default, the sessions created via the Interaction Panel, Playbooks, or Simulator aren't included in the reports or explorers. To include such data, proceed as follows:

1. At the top of the Insights interface, go to the filter bar.
2. Click **Endpoint**.
3. Activate the **Interaction Panel / Playbooks / Simulations** option.
4. Click **Apply** to activate the filter settings. All reports and explorers will then display the data according to your selection.

## Snapshots

Snapshots save the current form of your AI Agent and include Flows, Lexicons, Connections, Functions, and more. The Snapshots serve as an immutable form of an AI Agent in Cognigy.
For more information, read the [Snapshots](/ai/agents/deploy/snapshots) documentation.

To use the **Snapshot** filter, proceed as follows:

1. At the top of the Insights interface, go to the filter bar. Under **Snapshots**, open the list by clicking on it.
2. Use the **Type to search** field to find a specific Snapshot or select Snapshots from the list.
3. *(Optional)* Activate the **Negate** option to select the Snapshots you don't want to include in your analysis. All Snapshots except the selected ones will be considered for analysis.
4. *(Optional)* Activate the **None** option to obtain analysis data of all available endpoint data, except for Snapshot data.
5. *(Optional)* To view only selected options, click <img alt="filter" /> in the **Type to search** field.
6. *(Optional)* To reset the search results, click **x** in the **Type to search** field.
7. Click **Apply** to activate the filter settings. All reports and explorers will then display the data according to your selection.

## Locales

The **Locales** filter allows you to filter the data to include or exclude one or more languages or locales.
For more information, read the [Localization](/ai/platform-features/translation-and-localization/localization) documentation.

To use the **Locales** filter, proceed as follows:

1. At the top of the Insights interface, go to the filter bar and click **more filters**.
2. Under **Locales**, open the list by clicking on it.
3. Use the **Type to search** field to find a specific Locale or select Locales from the list.
4. *(Optional)* Activate the **Negate** option to select the Locales you don't want to include in your analysis. All Locales except the selected ones will be considered for analysis.
5. *(Optional)* To view only selected options, click <img alt="filter" /> in the **Type to search** field.
6. *(Optional)* To reset the search results, click **x** in the **Type to search** field.
7. Click **Apply** to activate the filter settings. All reports and explorers will then display the data according to your selection.

## Goals (beta)

The **Goals** filter allows you to include or exclude one or more goals from the data.\
For more information, refer to the [Goals](/ai/agents/analyze/goals-and-tasks/goals) documentation.

To use the **Goals** filter, follow these steps:

1. In the left-side menu of the Insights interface, select where you want to filter goals: the **Goals Report** or the **Conversations Explorer**.
2. At the top of the page, go to the filter bar and click **More Filters**. Under **Goals**, click to open the list.
3. Use the **Type to search** field to find a specific goal, or select goals from the list.
4. *(Optional)* Activate the **Negate** option to exclude the selected goals from your analysis. All goals except the selected ones will be considered.
5. *(Optional)* To view only the selected options, click <img alt="filter" /> in the **Type to search** field.
6. *(Optional)* To reset the search results, click **x** in the **Type to search** field.
7. Click **Apply** to activate the filter settings. Based on your selection, the report or explorer will display the corresponding data.

## Share Insights URL with Filters

When you share a specific URL with global filters applied,
your team members will see the same data and filters you've selected.

More benefits:

* You can reload any Insights page, and it will restore the Global Filters you previously selected.
* When reloading an open transcript in the Conversations explorer, you will land on the same page with the transcript open.

## More Information

* [Cognigy Insights](/insights/overview)
* [Download Report](/insights/download-reports)
* [Interaction Panel](/ai/agents/test/interaction-panel/overview)
* [Snapshots](/ai/agents/deploy/snapshots)
* [Endpoints](/ai/agents/deploy/endpoints/overview)
* [Localization](/ai/platform-features/translation-and-localization/localization)


# Insights Glossary
Source: https://docs.cognigy.com/insights/glossary

Glossary for Cognigy Insights terms.

This glossary includes all terms related to our Insights tool listed in alphabetical order.

## A

**Annotations**

An optional setting in a Node (refer to Cognigy) called Analytics Step can be defined in a configured Flow using the Node Editor. These Annotations should provide informative references to the configured Node. For instance, when annotating a Say Node that initiates a customer conversation with a welcome, you can enter `Greetings` or `Welcome New Caller`.

## C

**Chart**

A visual tool that displays data in a graphical format, making it easier to understand trends and patterns. Insights reports use various chart types such as bars, lines, and rings to present information effectively.

**Cognigy Insights**

A powerful analytics platform for evaluating and enhancing the performance of Cognigy Agents. It provides dynamic tools and key metrics across different reports to analyze customer interactions and improve AI Agent efficiency. For more information, read the [Cognigy Insights](/insights/overview) documentation.

**Conversation**

A billable unit of interaction between a user and an AI Agent or a human agent. For more information about billed conversations, read the [Cognigy Billing](/ai/administer/billing) documentation.

## D

**Data Expiration**

Data Expiration (TTL, time-to-live) displays the time span until the expiration of the Insights data. For more information, read the [Data Expiration](/insights/data-management/data-expiration) documentation.

**Download Report**

Each widget on the Insights reports provides a **Download Report as .csv file** option in a dropdown menu. The downloaded report file contains the complete widget dataset (depending on the filter settings) for storage or further analysis of your data. For more information, read the [Download Report](/insights/download-reports) documentation.

## E

**Endpoints**

Endpoints ensure that channel-specific input and output (images, galleries, voice) are converted to the standardized Input object. This means that a Cognigy Flow only needs to be developed once, after which it can be deployed to multiple channels by configuring Endpoints. Examples of Endpoints include Webchat, Voice Gateway, Microsoft Teams, and more. In Insights, you can select one or more Endpoints for analysis. For more information, read the [Endpoints](/insights/global-filters#endpoints) documentation.

**Debug Mode**

This mode provides additional information for conversation messages, including the triggered Intent and its score, as well as the filled Slots. You can enable Debug mode by toggling the switch at the top of the Transcript panel. For more information, read the [Debug Mode](/insights/explorers/conversations#debug-mode) documentation.

**Engagement**

A report in Cognigy Insights that displays performance metrics for your agent, focusing on customer interactions in conversations. For more information, read the [Engagement Report](/insights/reports/engagement) documentation.

**Escalation**

See [Live Agent Handovers](#l).

**Explorer**

Analytical tools integrated into Cognigy Insights are designed to evaluate and display conversational data. The primary purpose of an explorer is to assist you in understanding customer reactions (messages) during a conversation with your AI Agent, ultimately improving your service.

## F

**Filter Bar**

A bar displayed at the top of each Insights report or explorer. The filter bar includes all [global filters](#g) and [local filters](#l), accessible by clicking the **more filters** button.

## G

**Global Filter**

A tool that allows users to set consistent filtering criteria, such as time span, endpoints, snapshot, locales, and message rating, across different reports and explorers. For more information, read the [Global Filter](/insights/global-filters) documentation.

## L

**Live Agent**

A report in Cognigy Insights that presents Live Agent metrics for a specific AI Agent. For more information, read the [Live Agent](/insights/reports/live-agent) documentation.

**Live Agent Handovers**

An Indicator chart that displays the number of handovers. Handovers occur when an AI Agent transfers a conversation to a human agent, and multiple handovers may occur during a session. For more information, read the [Live Agent Escalations](/insights/reports/live-agent#live-agent-handovers) documentation.

**Local Filter**

The Local Filters are specific to a report or an explorer. For more information, see the Local Filter example.

## M

**Message Rating**

A configurable Global Filter setting that allows users to refine their data analysis based on the sentiment of user feedback within a session.

## N

**NLU Performance**

A report in Cognigy Insights that provides a comprehensive view of your AI Agent's natural language understanding (NLU) capabilities. It enables you to monitor and enhance message comprehension, intent recognition, and execution efficiency. For more information, read the [NLU Performance](/insights/reports/nlu-performance) documentation.

## O

**Overview**

A report in Cognigy Insights that displays various metrics as percentages and in graphical formats such as bar charts, ring charts, tiles, or time-based line charts. The charts display data based on the Filter settings. For more information, read the [Overview](/insights/reports/overview) documentation.

## R

**Report**

A single-page canvas that serves as a visual interface to present key performance indicators, metrics, and trends through a variety of visualizations, including line charts, bar charts, ring charts, and other chart types.

## S

**Session**

A unit of interaction between a user and a Cognigy AI Agent or human agent. The session begins when a user sends their initial input to Cognigy, which can be either a chat or voice message, and ends when the user or human agent completes or terminates the interaction. The session can contain an unlimited number of user inputs.

**Snapshots**

An exported Agent that contains necessary resources. A Snapshot is an unchangeable, real-time copy of all relevant Agent resources. For more information, read the [Global Filters: Snapshots](/insights/global-filters#snapshots) documentation.

## T

**Tiles**

A UI element designed for displaying Indicator charts. These tiles are positioned at the top of a report. For more information, read the [Overview Report](/insights/reports/overview) documentation.

**Time span**

A Global Filter settings option. You can choose from preset time periods or set it individually using a date picker in a calendar. For more information, read the [Time Span](/insights/global-filters#time-span) documentation.

**Time zone**

Insights supports presenting time-based data according to the user's specified time zone in the Cognigy user profile settings. For more information, read the [Timezone](/insights/data-management/time-zone) documentation.

**Transcripts**

Logged conversations between end users and agents.

**Conversations Explorer**

An analytical tool that allows you to evaluate messages in detail to enhance your agent, customer service, and overall user experience. The Conversations explorer enables you to search for key phrases such as `booking`, `cancel`, or `message` within a list of transcripts. For more information, read the [Conversations explorer](/insights/explorers/conversations) documentation.


# Cognigy Insights
Source: https://docs.cognigy.com/insights/index



<div>
  <div>
    *Cognigy Insights* is a tool that provides businesses with actionable insights and analytics derived from conversational data collected through Cognigy.

    <p>
      Insights help businesses better understand their users, optimize chat and voice interactions, and improve the overall customer experience.
    </p>

    <p>
      Insights also enables organizations to boost AI Agent performance and deliver consistently exceptional interactions.
    </p>
  </div>

  <div>
    <video aria-label="Cognigy Insights hero animation showing horizontal and vertical bar charts, along with indicators for average success rate and average sentiment metrics.">
      <source type="video/mp4" />
    </video>

    <video aria-label="Cognigy Insights hero animation showing horizontal and vertical bar charts, along with indicators for average success rate and average sentiment metrics.">
      <source type="video/mp4" />
    </video>

    <button type="button" aria-label="Pause video">
      ⏸
    </button>
  </div>
</div>

<CardGroup>
  <Card title="About Insights" icon="info" href="/insights/overview">
    To analyze your AI Agents and take your chat and voice agents to the next level, explore the capabilities of Cognigy Insights.
  </Card>

  <Card title="View Report Metrics" icon="chart-no-axes-combined" href="/insights/reports/overview">
    Monitor key metrics by utilizing ready-made charts on reports.
  </Card>

  <Card title="Use Explorers" icon="telescope" href="/insights/explorers/conversations">
    Examine sessions separately to thoroughly investigate performance issues.
  </Card>

  <Card title="Analyze Conversations" icon="sparkles" href="/insights/conversation-analyzer/overview">
    Turn conversations into insights with LLM-powered analysis. Measure sentiment, containment, instruction following, and more using customizable criteria.
  </Card>

  <Card title="Download Reports" icon="download" href="/insights/download-reports">
    Download reports individually for each chat on the reports.
  </Card>

  <Card title="Filter and Sort" icon="list-filter-plus" href="/insights/global-filters">
    Use global filters to effectively sort and filter data both on reports and explorers.
  </Card>

  <Card title="Data Management" icon="calendar-sync" href="/insights/data-management/pre-aggregation">
    Learn how data pre-aggregation is used for reports, configure Time Zone and TTL (Time To Live) settings for your project.
  </Card>
</CardGroup>


# Cognigy Insights
Source: https://docs.cognigy.com/insights/overview

Learn about Cognigy Insights features, including reports, explorers, and key metrics to analyze AI Agent usage and performance across chat and voice channels.

<a href="/release-notes/2026.14"><Badge>Updated in 2026.14</Badge></a>

*Cognigy Insights* replaces our previous static analytics report and provides you with the tools you need to improve the customer experience your AI Agents provide.
Insights features a range of key metrics to help you understand your AI Agents' usage and performance.

<Frame>
  <img alt="Cognigy Insights main interface with reports, explorers, and key metrics" />
</Frame>

Cognigy provides two data analysis solutions catering to different preferences:

* **Cognigy OData endpoint**. Allows access to raw data through APIs and is highly customizable, making it suitable for integration with BI tools and creating custom reports. For more information about the OData endpoint, read the [Cognigy OData Analytics Endpoint](/ai/agents/analyze/odata) documentation.
* **Cognigy Insights**. Offers user-friendly prebuilt reports and tools for quick data analysis and visualization without requiring extensive technical expertise.

The comparison of the OData Endpoint and Insights is outlined in the table.

| Feature             | Cognigy OData Endpoint                                                                      | Cognigy Insights                                                                   |
| ------------------- | ------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------- |
| Purpose             | Provides access to raw data                                                                 | Analyzes and visualizes raw data                                                   |
| Data Type           | Raw data (structured format)                                                                | Processed data (visualizations, reports)                                           |
| Functionality       | Allows retrieval of data using GET requests                                                 | Offers data exploration, metric tracking, performance analysis, data export in CSV |
| User Interface      | API                                                                                         | User-friendly interface with reports                                               |
| Output              | Data in a structured format (for example, JSON, XML)                                        | Insights, trends, visualizations                                                   |
| Use Case            | Integrate data with external BI tools <br /> Build custom reports for complex data analysis | Get quick insights through pre-built reports                                       |
| Technical Expertise | API knowledge required                                                                      | Less technical knowledge required                                                  |
| Customization       | Highly customizable                                                                         | Limited customization                                                              |

## Reports

Cognigy Insights is tightly integrated with some of Cognigy's products.
Insights features a variety of reports to highlight different data types, enabling you to track your customers' behavior and the success of your AI Agents. Additionally, all collected data is accessible through OData.
Insights operates within individual Cognigy Projects. This allows you to analyze data specific to each Project and switch between Projects seamlessly. However, you cannot compare data across Projects simultaneously.

Access Cognigy Insights by opening the left-side menu in a Cognigy Project and selecting **Insights**. To switch Projects within Insights, use the dropdown menu in the upper-left corner.
When switching from Cognigy to Cognigy Insights, you will be redirected to the Insights **Report** section, starting on the Overview report.

<Frame>
  <img alt="Overview report displayed after navigating from Cognigy to Insights" />
</Frame>

Cognigy Insights features a variety of reports that visualize key metrics for your currently selected AI Agent.

Insights provides the following Reports:

* [Overview](/insights/reports/overview)
* [Engagement](/insights/reports/engagement)
* [NLU Performance](/insights/reports/nlu-performance)
* [Live Agent](/insights/reports/live-agent)
* [Goals](/insights/reports/goals)

Insights also provides the following explorer:

* [Conversations](/insights/explorers/conversations)

## Filters

Cognigy Insights offers global and local filters. The filter bar shows only a selection of global filters. To access all filters, including both global and local ones, click **more filters**. Any filter changes you make will be applied to all data visualizations throughout the platform.

**Global Filters**

Global filters apply across Cognigy Insights. Once you set them up, they persist throughout your session, even if you change from one report to another.

**Local Filters**

Local filters are specific filters available only for one or a few reports. When you switch from one board to another, the boards you were on will be temporarily disabled but not removed from your selection.

## User Menu

You can find the [User Menu](/ai/administer/user-menu/overview) in the upper-right corner of the screen.

## Main Menu

The Insights left-side menu displays all available reports and the Conversations explorer. This menu lets you easily switch between reports and the explorer to analyze your AI Agent's performance.

You can always identify the currently selected view in two ways: it is highlighted in light blue-gray in the menu and displayed in the header at the top of the screen.

The left-side menu also allows you to switch back to Cognigy by clicking **Cognigy**.

<Frame>
  <img alt="Insights left-side menu listing available reports and the Conversations explorer" />
</Frame>

<Tip>
  To learn more, watch our live presentation of Cognigy Insights:

  * [Insights product introduction (English)](https://youtu.be/hYB2iQ_dT5g)
  * [Insights product introduction (German)](https://youtu.be/kCkUrHM9dQo)
</Tip>

## More Information

* [Global Filters](/insights/global-filters)
* [Pre-Aggregation](/insights/data-management/pre-aggregation)
* [Data Expiration](/insights/data-management/data-expiration)


# Engagement
Source: https://docs.cognigy.com/insights/reports/engagement

Analyze customer interaction metrics, containment rates, session data, and handover performance with the Cognigy Insights Engagement report.

<a href="/release-notes/2026.14"><Badge>Updated in 2026.14</Badge></a>

The *Engagement Report* highlights customer interactions in conversations and visualizes the engagement performance of your AI Agent in the following charts:

* [Indicators](/insights/reports/engagement#indicators)
* [Heat maps](/insights/reports/engagement#heat-maps)
* [Bar charts](/insights/reports/engagement#bar-charts)
* [Line charts](/insights/reports/engagement#line-charts)

Charts display data based on the [global filter](https://docs.cognigy.com/insights/global-filters/) settings.

<Note>
  This report's metrics include sessions with personal data masked using the [Delete Profile](/ai/agents/develop/node-reference/analytics/delete-profile) or [Deactivate Profile](/ai/agents/develop/node-reference/analytics/deactivate-profile) Nodes.
</Note>

<Frame>
  <img alt="Engagement report showing session metrics, ratings, and handover data" />
</Frame>

## Indicators

The Indicator charts display real-time data that refreshes every 30 seconds.

### Tasks reached

This chart displays the total number of [completed tasks](/ai/agents/develop/node-reference/analytics/complete-task).

Source table:

```txt theme={null}
 | StartedAt  |  Sessions   |  Tasks                   | Tasks reached |
 | ---------- | ----------- | ------------------------ | ------------- |
 | 2022/01/01 | session-123 | "Goal A, Goal B, Goal D" |       3       |
 | 2022/01/01 | session-456 | "Goal B, Goal C, Goal E" |       3       |
 | 2022/01/03 | session-789 | "Goal B, Goal E"         |       2       |
 | 2022/01/03 | session-abc | "Goal B"                 |       1       |
 | 2022/01/05 | session-def | "Goal A, Goal B"         |       2       |
```

Result: `11`

### Avg. Session length

Displays how long the average session lasted.

The session length is defined as the time difference between the timestamp of the first user input and the last user input.
The Average session length is the mean of all session lengths within the specified time span in your Insights filter.

Source table:

```txt theme={null}
 | StartedAt           |  EndedAt            | Sessions    |          
 | ------------------- | ------------------- | ----------- |
 | 2022/01/01 12:15 00 | 2022/01/01 12:18 00 | session-123 |
 | 2022/01/01 14:00 00 | 2022/01/01 15:00 00 | session-456 |
 | 2022/01/03 09:05 30 | 2022/01/03 09:15 30 | session-789 |
 | 2022/01/03 11:30 00 | 2022/01/03 11:44 00 | session-abc |
 | 2022/01/05 06:10 00 | 2022/01/05 06:20 00 | session-def |
```

Calculation:

1. Length for a particular session `Max(timestamp) - Min(timestamp)`.
2. `Length of all sessions / total number of sessions`.

Example:

1. Total Length = `3 minutes + 1 hour + 10 minutes + 14 minutes + 10 minutes = 1 hour 37 minutes`
2. Average Session Length = `1 hour 37 minutes / 5 = 19 minutes and 24 seconds`

Result: `00:19:24`

### Positive Ratings

Displays the total number of Positive Ratings the end user has given.

Source table:

```txt theme={null}
| Sessions    | Positive Rating | Negative Rating | Rated Sessions |
| ----------- | --------------- | --------------- | -------------- |
| session-123 |        1        |        0        |       1        |
| session-456 |        0        |        1        |       1        |
| session-789 |        0        |        0        |       0        |
| session-abc |        1        |        0        |       1        |
| session-def |        1        |        0        |       1        |
| ---------------------------------------------------------------- |
| TOTAL       |        3        |        1        |       4        |
```

Result: `3`

### Containment Rate

Displays containment rate. The containment rate is calculated by the ratio of sessions without any handover to all sessions.

Source table:

```txt theme={null}
| Session ID  | Handovers | Handover Sessions | Sessions |
| ----------- | ----------| ----------------- | -------- |
| session-123 |     3     |        1          |    1     |
| session-456 |     0     |        0          |    1     |
| session-789 |     0     |        0          |    1     |
| session-abc |     2     |        1          |    1     |
| session-def |     1     |        1          |    1     |
| ------------------------------------------------------ |
| TOTAL       |     6     |        3          |    5     |
```

Formula:

1. Containment Rate = `1 - (Handover included Sessions)/(Total Sessions)`.

2. Percentage `Containment Rate * 100%`.

Result: `40%`

## Heat maps

### Handovers by time of the day

Displays the number of times a human agent is contacted during the day. Multiple handovers can occur within a single session.

Source table:

```txt theme={null}
| DateTime        | Handover   |
| --------------- | ---------- |
| 01/8/2023 16:00 |	   0       |
| 01/8/2023 17:00 |	   0       |
| 01/9/2023 00:00 |    0       |
| 1/11/2023 15:00 |    0       |
| 1/15/2023 18:00 |    0       |
| 1/16/2023 21:00 |    0       |
| 1/17/2023 11:00 |    0       |
| 1/18/2023 06:00 |    1       |
| 1/18/2023 21:00 |    1       |
| 1/19/2023 00:00 |    2       |
| 1/19/2023 10:00 |    1       |
| 1/23/2023 13:00 |    0       |
| 1/24/2023 23:00 |    0       |
```

Result:

<Frame>
  <img alt="Heat map showing handover frequency by time of day and day of week" />
</Frame>

## Bar charts

### Containment Rate

<a href="/release-notes/4.48"><Badge>Updated in 4.48</Badge></a>

Indicates the percentage of sessions without any handover during the timeline.

Source table:

```txt theme={null}
| Date      | Containment Rate | Total Sessions | Handover included Sessions |
| --------- | ---------------  | -------------- | -------------------------- |
| 3/1/2023  | 	87.5           |       40       | 	       35                |
| 3/2/2023  | 	82.60869565    |       46       | 	       38                |
| 3/3/2023  | 	96.77419355    |       62       | 	       60                |
| 3/6/2023  | 	92.5           |       40       | 	       37                |
| 3/7/2023  | 	100            |       58       | 	       58                |
| 3/8/2023  | 	100            |       92       | 	       92                |
| 3/10/2023 | 	100            |       10       | 	       10                |
```

Formula:

1. Containment Rate = `1 - (Handover included Sessions)/(Total Sessions)`.

2. Percentage `Containment Rate * 100%`.

Example for the 3/1/2023: `(35/40)* 100% = 87.5%`

Result:

<Frame>
  <img alt="Bar chart showing the containment rate percentage over time" />
</Frame>

## Line charts

### Handovers in Time

Displays the number of times a human agent was contacted within the filtered time span. Multiple handovers can occur within a single session.

Source table:

```txt theme={null}
| DateTime   | Handovers per day |
|------------|-------------------|
| 4/11/2023  |         3         |           
| 4/12/2023  |         0         |            
| 4/13/2023	 |        24         |           
| 4/14/2023	 |         8         |             
| 4/17/2023	 |        17         |        
| 4/18/2023	 |        12         |       
| 4/20/2023	 |         5         |         
| 4/21/2023	 |         5         |        
```

Result:

<Frame>
  <img alt="Line chart showing the number of handovers to human agents over time" />
</Frame>

### Positive Ratings over Time

Displays the total number of Positive Ratings given as feedback by the end user.

Source table:

```txt theme={null}
| Date (weeks) | Positive | Negative | Number of Ratings |
| ------------ | -------- | -------- | ----------------- |	
| 2023-W01	   |     1	  |    0     |         1         |
| 2023-W04	   |     1	  |    0     |         1         |
| 2023-W05	   |     2	  |    0     |         2         |
```

Result:

<Frame>
  <img alt="Line chart showing positive user ratings over time" />
</Frame>

### Unique Contacts

Displays the number of unique contacts (`userid`) in relation to the selected timespan. If the user creates more than one session in a single day, it will be counted as one unique contact for that day.

Source table:

```txt theme={null}
| Date    	| Contacts |
| --------- | -------- |
| 1/16/2023	|    1     |
| 1/17/2023	|    8     |
| 1/18/2023	|    15    |
| 1/19/2023	|    6     |
| 1/20/2023	|    18    |
```

Result:

<Frame>
  <img alt="Line chart showing the number of unique contacts per day" />
</Frame>

## Download Report

To download a report as a CSV file for the chart,
click <img alt="menu" /> in the upper-right corner of the chart.
To create a report file, refer to the [Download Report](https://docs.cognigy.com/insights/download-reports/) article.

## More Information

* [Overview](/insights/reports/overview)
* [Goals Report](/insights/reports/goals)
* [NLU Performance Report](/insights/reports/nlu-performance)
* [Live Agent Report](/insights/reports/live-agent)


# Goals
Source: https://docs.cognigy.com/insights/reports/goals

Track AI Agent goal completion, time savings, cost savings, and revenue impact with the Insights Goals report. Requires the new Insights framework.

<a href="/release-notes/2026.14"><Badge>Updated in 2026.14</Badge></a>

<Note>
  The Goals Report is in beta and is available if you have migrated to the new [Insights framework](/insights/data-management/pre-aggregation).
  Before migration, you can also explore the Goals Report in the Cognigy trial environment.
  We encourage you to try this feature out and provide us with feedback.
</Note>

The *Goals Report* highlights goal tracking, completion status, time, revenue, and cost metrics, and the business impact of AI Agents in the following charts:

* [Indicators](/insights/reports/engagement#indicators)
* [Bar charts](/insights/reports/engagement#bar-charts)

Charts display data based on the [global filter](https://docs.cognigy.com/insights/global-filters/) settings.

<Note>
  This report's metrics include sessions with personal data masked using the [Delete Profile](/ai/agents/develop/node-reference/analytics/delete-profile) or [Deactivate Profile](/ai/agents/develop/node-reference/analytics/deactivate-profile) Nodes.
</Note>

<Frame>
  <img alt="Goals report showing completion status, time saved, cost savings, and revenue metrics" />
</Frame>

## Indicators

### Completed

Displays the total number of [completed goals](/ai/agents/develop/node-reference/analytics/track-goal).

### Partially Completed

Displays the total number of partially [completed goals](/ai/agents/develop/node-reference/analytics/track-goal).
This chart shows the goals for which some steps were not completed.

### Time Saved

Displays the total time saved for completed and partially completed goals.

### Cost Savings

Displays the total cost savings for completed and partially completed goals.

## Bar charts

### Goals Performance (in numbers)

Displays the number of goals achieved within the selected time span.
The goals have three distinct categories, each has a separate bar:

* **Incomplete** — goals that have been started, with only the initial step triggered.
* **Partial Completion** — goals that have been started and progressed through some steps, but the completion step was not achieved.
* **Total Completion** — goals that have been started, progressed through some steps, and at least one completion step was achieved.

Source table:

```txt theme={null}
| Dimension | Measure                   | Value |
|-----------|---------------------------|-------|
| Goal_01   | completed_count           | 13667 |
| Goal_04   | completed_count           | 3601  |
| Goal_05   | completed_count           | 3425  |
| Goal_03   | completed_count           | 1396  |
| Goal_02   | completed_count           | 1358  |
| Goal_08   | completed_count           | 99    |
| Goal_01   | incompleted_count         | 6640  |
| Goal_04   | incompleted_count         | 3691  |
| Goal_05   | incompleted_count         | 3573  |
| Goal_03   | incompleted_count         | 3124  |
| Goal_02   | incompleted_count         | 11687 |
| Goal_08   | incompleted_count         | 0     |
| Goal_01   | partially_completed_count | 3229  |
| Goal_04   | partially_completed_count | 3494  |
| Goal_05   | partially_completed_count | 3632  |
| Goal_03   | partially_completed_count | 0     |
| Goal_02   | partially_completed_count | 0     |
| Goal_08   | partially_completed_count | 0     |
```

Result:

<Frame>
  <img alt="Bar chart showing goal completion counts broken down by incomplete, partial, and total" />
</Frame>

### Time Saved (in hours)

Displays time saved in hours for each goal that was completed within the filtered time span.
You can track efficiency gains and compare time saved across different goals.

Source table:

```txt theme={null}
| Date       | Goal_1   | Goal_3 | Goal_4   | Goal_5   |
|------------|----------|--------|----------|----------|
| 27.10.2024 | 37:27:00 | -      | 22:27:00 | 28:20:00 |
| 28.10.2024 | 216 h    | 25:30  | 37:06:00 | 48:50:00 |
| 29.10.2024 | 165 h    | 23:30  | 22:42:00 | 27:50:00 |
| 30.10.2024 | 108 h    | 8:30   | 27:48:00 | 36:20:00 |
| 31.10.2024 | 82:53:00 | 5:30   | 38:03:00 | 49:30:00 |
| 01.11.2024 | 144 h    | 18:30  | 16:12:00 | 25:10:00 |
| 02.11.2024 | 31:22.00 | -      | 18:30:00 | 21:10:00 |
```

Result:

<Frame>
  <img alt="Bar chart displaying hours saved per goal over the selected time span" />
</Frame>

### Cost Savings

Displays the amount of the money saved for each goal completed within the selected time span.
You can analyze cost savings across different goals, identify the most cost-effective processes, and optimize resource allocation.

Source table:

```txt theme={null}
| Date       | Goal_1 | Goal_3 | Goal_4 | Goal_5 | Goal_6 |
|------------|--------|--------|--------|--------|--------|
| 27.10.2024 | €594   | -      | €1167  | €253   | -      |
| 28.10.2024 | €1539  | €1020  | €1929  | €440   | €43    |
| 29.10.2024 | €930   | €900   | €1182  | €254   | €47    |
| 30.10.2024 | €942   | €340   | €1448  | €319   | €22    |
| 31.10.2024 | €1078  | €220   | €1981  | €440   | €5     |
| 01.11.2024 | €888   | €740   | €842   | €219   | €44    |
| 02.11.2024 | €476   | -      | €962   | €189   | -      |
```

Result:

<Frame>
  <img alt="Bar chart displaying cost savings per goal over the selected time span" />
</Frame>

### Revenue Earned

Displays the total revenue earned for each goal completed within the selected time period.
You can analyze earnings across different goals to identify the most efficient processes.

Source table:

```txt theme={null}
| Date       | Goal_01 | Goal_04 | Goal_05 | Goal_08 |
|------------|---------|---------|---------|---------|
| 11/11/2024 | €0      | €0      | €0      | €0      |
| 11/12/2024 | €3984   | €861    | €546    | €96     |
| 11/13/2024 | €0      | €1332   | €783    | €0      |
| 11/14/2024 | €0      | €1533   | €842    | €0      |
| 11/15/2024 | €0      | €297    | €170    | €0      |
| 11/16/2024 | €0      | €0      | €0      | €0      |
```

Result:

<Frame>
  <img alt="Bar chart displaying revenue earned per goal over the selected time period" />
</Frame>

## More Information

* [Goals](/ai/agents/analyze/goals-and-tasks/goals)
* [Overview Report](/insights/reports/overview)
* [NLU Performance Report](/insights/reports/nlu-performance)
* [Live Agent Report](/insights/reports/live-agent)


# Live Agent
Source: https://docs.cognigy.com/insights/reports/live-agent

Monitor human agent handovers, response times, resolution rates, and conversation statuses with the Live Agent report in Cognigy Insights.

<a href="/release-notes/2026.14"><Badge>Updated in 2026.14</Badge></a>

The *Live Agent Report* displays Live Agent metrics for a specific AI Agent.

Using this report, you can:

* Keep track of how quickly human agents respond and resolve customer inquiries.
* Monitor the total number of handovers for the AI Agent.
* Summarize conversation statuses.
* Evaluate conversation types based on the assigned label.

These metrics are visualized in the following charts:

* [Indicators](#indicators)
* [Bar charts](#bar-charts)
* [Line charts](#line-charts)

Charts display data based on the [global filter](https://docs.cognigy.com/insights/global-filters/) settings.

<Note>
  This report's metrics include sessions with personal data masked using the [Delete Profile](/ai/agents/develop/node-reference/analytics/delete-profile) or [Deactivate Profile](/ai/agents/develop/node-reference/analytics/deactivate-profile) Nodes.
</Note>

<Frame>
  <img alt="Live Agent report showing handover metrics, response times, and resolution data" />
</Frame>

## Indicators

The Indicator charts present real-time data that refreshes every 30 seconds.

### Live Agent Handovers

This overview shows the number of handovers. A handover occurs when an AI Agent transfers the conversation to a human agent. Multiple handovers can occur within a single session.

Source table:

```txt theme={null}
| Sessions    | DateTime  | Handovers per session  | Escalation |
| ----------- | --------- | ---------------------- | ---------- |
| session-123 | 1/08/2023 |          1             |     1      |
| session-456 | 1/09/2023 |          0             |     0      |
| session-789 | 1/11/2023 |          0             |     0      |
| session-abc | 1/15/2023 |          2             |     1      |
| session-def | 1/16/2023 |          4             |     1      |
```

Result: `7`

### Handover Rate

Displays the percentage of sessions needing human assistance. The handover rate is calculated as the ratio of Live Agent Escalations divided by total sessions. A lower value indicates fewer sessions needing human assistance.
If multiple handovers occur during a session, Cognigy counts them as one escalation.

Source table:

```txt theme={null}
| Sessions    | DateTime  | Handovers per session  |  Escalation   |
| ----------- | --------- | ---------------------- | ------------- |
| session-123 | 1/08/2023 |          1             |       1       |
| session-456 | 1/09/2023 |          0             |       0       |
| session-789 | 1/11/2023 |          0             |       0       | 
| session-abc | 1/15/2023 |          2             |       1       | 
| session-def | 1/16/2023 |          4             |       1       | 
```

Calculation: `(Number of Live Agent Escalations) / (Total Sessions) * 100%`

Result: `60%`

### Avg. First Response Time

Displays the average time it takes for a human agent to send the first response to an end user.

Source table:

```txt theme={null}
 | Request time        | First response time | Sessions    |          
 | ------------------- | ------------------- | ----------- |
 | 2022/01/01 12:15 00 | 2022/01/01 12:16 00 | session-123 |
 | 2022/01/01 14:00 00 | 2022/01/01 14:02 00 | session-456 |
 | 2022/01/03 09:05 30 | 2022/01/03 09:05 30 | session-789 |
 | 2022/01/03 11:30 00 | 2022/01/03 11:31 00 | session-abc |
 | 2022/01/05 06:10 00 | 2022/01/05 06:12 30 | session-def |
```

Calculation:

1. For each session, `First response time - Request time`.
2. Average first response time = `Total time to send a first response / Number of sessions`

Result: `1 min 10 sec`

### Avg. Resolution Time

Displays the average time it takes a human agent to resolve a conversation after escalation.
Source table:

```txt theme={null}
 | Request time        | Resolving time      | Sessions    |          
 | ------------------- | ------------------- | ----------- |
 | 2022/01/01 12:15 00 | 2022/01/01 12:30 00 | session-123 |
 | 2022/01/01 14:00 00 | 2022/01/01 14:15 00 | session-456 |
 | 2022/01/03 09:05 00 | 2022/01/03 09:14 30 | session-789 |
 | 2022/01/03 11:30 00 | 2022/01/03 11:42 00 | session-abc |
 | 2022/01/05 06:10 00 | 2022/01/05 06:12 30 | session-def |
```

Calculation:

1. For each session, `Resolving time - Request time`.
2. Average resolution time = `Total Resolution time / Total number of sessions`

Result: `9 min`

### Opened

Indicates the number of conversations that were opened by a user or reopened by a human agent.

### Assigned

Indicates the number of conversations that were assigned to a human agent.

### Pending

Indicates the number of conversations that were paused until a customer or human agent responded.

### Snoozed

Indicates the number of conversations that were temporarily suspended until a time trigger or customer response trigger was activated.

## Bar charts

### Handover End Reason

Displays the following conversation statuses on a timescale:

* Abandoned — a conversation is closed because the user left the conversation.
* Resolved — a conversation is closed because the user's problem was resolved.

The chart indicates how the Live Agents' performance varied over time.

Source table:

```txt theme={null}
| Date      | Status    |  Conversations  |
| --------- | --------- | --------------- |
| 3/20/2023 | abandoned |        3        |
| 3/21/2023 | abandoned |        2        |
| 3/21/2023 | resolved  |        1        |
| 3/29/2023 | abandoned |        2        |
| 3/29/2023 | resolved  |       24        |
| 3/30/2023 | resolved  |        5        |
| 4/4/2023  | resolved  |        1        |
| 4/13/2023 | resolved  |       15        |
| 4/13/2023 | abandoned |       13        |
| 4/13/2023 | pending   |        1        |
| 4/13/2023 | snoozed   |        1        |
| 4/14/2023 | assigned  |       22        |
| 4/14/2023 | abandoned |        3        |
| 4/14/2023 | resolved  |        5        |
| 4/17/2023 | assigned  |        7        |
| 4/17/2023 | resolved  |       15        |
| 4/18/2023 | assigned  |        6        |
| 4/18/2023 | abandoned |        6        |
```

Result:

<Frame>
  <img alt="Bar chart showing handover end reasons such as abandoned and resolved over time" />
</Frame>

### Label Summary

Indicates Live Agents labels for categorizing and prioritizing conversations.

```txt theme={null}
| Label name | Number of labels | 
| ---------  | ---------------- | 
|   QA       |       4          |
|   test 	 |       1          |
```

Result:

<Frame>
  <img alt="Bar chart summarizing conversation labels assigned by human agents" />
</Frame>

## Line charts

### Live Agent Performance Times

Displays the average time to first response and resolution time for human agents.
This metric is calculated based on the number of sessions and helps evaluate human agent effectiveness by comparing response and resolution times.

The First Response Time metric refers to the average time a human agent takes to send a first response to a customer's request.
The Resolution Time metric refers to the average time of a customer's interaction with a human agent, from the start of the handover sequence until the end of the session.

Source tables:

* [Avg. First Response Time](#avg-first-response-time)
* [Avg. Resolution Time](#avg-resolution-time)

Result:

<Frame>
  <img alt="Line chart comparing average first response time and resolution time for human agents" />
</Frame>

## Download Report

To download a report as a CSV file for the chart,
click <img alt="menu" /> in the upper-right corner of the chart.
To create a report file, refer to the [Download Report](https://docs.cognigy.com/insights/download-reports/) article.

## More Information

* [Overview Report](/insights/reports/overview)
* [Engagement Report](/insights/reports/engagement)
* [NLU Performance Report](/insights/reports/nlu-performance)


# NLU Performance
Source: https://docs.cognigy.com/insights/reports/nlu-performance

Analyze NLU Intent recognition accuracy, execution times, and slot performance for your AI Agent with the Cognigy Insights NLU Performance report.

<a href="/release-notes/2026.14"><Badge>Updated in 2026.14</Badge></a>

The *NLU Performance Report* provides valuable insights into the performance of your AI Agent's natural language understanding (NLU) capabilities. It allows you to monitor and enhance message comprehension, Intent recognition, and execution efficiency. Moreover, it offers easy access to resources like the Intent Trainer for further performance improvement.

You can find more information about our NLU and Intents in the [NLU Overview](/ai/platform-features/nlu/overview) article.

The NLU Performance data of your AI Agent is visualized in the following charts:

* [Indicators](/insights/reports/engagement#indicators)
* [Bar charts](/insights/reports/engagement#bar-charts)
* [Line charts](/insights/reports/engagement#line-charts)

These charts display all data based on the [global filter](/insights/global-filters) and local filter settings.

<Note>
  This report's metrics include sessions with personal data masked using the [Delete Profile](/ai/agents/develop/node-reference/analytics/delete-profile) or [Deactivate Profile](/ai/agents/develop/node-reference/analytics/deactivate-profile) Nodes.
</Note>

<Frame>
  <img alt="NLU Performance report showing intent scores, execution times, and slot metrics" />
</Frame>

## Indicators

The Indicator charts present real-time data that refreshes every 30 seconds.

### Average Execution Time

Displays the average execution time across a specific timespan.

Formula:

Average Execution Time = `Sum of Execution Times / Total number of inputs`

### Avg Intent Score

Displays the average Intent score for the selected time span. Higher Intent scores indicate better AI Agent performance in matching user queries.
For more information, see [Machine Learning Intents](/ai/platform-features/nlu/intents/overview).

To find the average intent score, Insights sums up all the intent scores and divides the sum by the total number of records.

Source table:

```txt theme={null}
| Analytics Record | Intent Score |
| ---------------- | ------------ |
| Record-1         | 0.30         |
| Record-2         | 0.90         |
| Record-3         | 0.10         |
| Record-4         | 1            |
| Record-5         | 0.20         |
| Record-6         | 0.02         |
| Record-7         | 0.50         |
```

Calculation:

1. Sum of Intent Scores = `0.30 + 0.90 + 0.10 + 1 + 0.20 + 0.02 + 0.50 = 3.02`
2. Number of Records = `7`
3. Average Intent Score = `(Sum of Intent Scores) / (Number of Records)`
4. Average Intent Score = `3.02 / 7`

Result: `0.43`

### Max Intent Score

Indicates the number of Intents that were matched with 100% efficacy.

Source table:

```txt theme={null}
| Analytics Record | Intent Score |
| ---------------- | ------------ |
| Record-1         | 0.30         |
| Record-2         | 0.90         |
| Record-3         | 0.10         |
| Record-4         | 1            |
| Record-5         | 0.20         |
| Record-6         | 0.02         |
| Record-7         | 0.50         |
```

Result: In the provided source table, there is one record with an intent score of `1`, so the max intent score is `1`.

## Bar charts

### Top Intents

Displays the top-rated number of Intents. The summary view shows a maximum of 6 Intents, while the detailed view displays up to 25. Matching an Intent triggers its recording in the analytics data. The data is then grouped by the matched Intent and sorted in descending order.
Source table:

```txt theme={null}
|  Intent           | Total number  |
| ----------------- | ------------- |
| Flight discounts  |      12       |
| Book a flight     |      5        |
| Change a flight   |      5        |
| Additional fee    |      4        |
| Reject a flight   |      3        |
| Transfer flights  |      3        |
```

Result:

<Frame>
  <img alt="Horizontal bar chart ranking the most frequently matched intents" />
</Frame>

### Top Slots

Similar to **Top Intents**, this chart displays the top-rated number of Slots in a horizontal bar chart.

Source table:

```txt theme={null}
| Slot        | Number      |
| ----------- | ----------- |
| DATE        | 2           |
| NUMBER      | 1           |
```

Result:

<Frame>
  <img alt="Horizontal bar chart showing the most frequently filled slots" />
</Frame>

### Intents by Score Range

Displays the scoring range of matched Intents. Each score range is displayed as a bar with minimum, maximum, and average values.
Clicking on the three-dots menu item,
you can select the **Go to Intent Trainer** that leads you to the application in Cognigy to improve your Flow.
For more information, see [Intent Trainer](/ai/platform-features/nlu/intents/intent-trainer).

Source table:

```txt theme={null}
| Date               | intentScoreAvg     | intentScoreMax     | intentScoreMin     |
| ------------------ | --------------     | --------------     | ------------------ |
| cakes              | 0.6517875252045072 | 0.6517875252045072 | 0.6517875252045072 |
| pizza              | 0.6863160928710601 | 0.7237952611766103 | 0.6491475952551133 |
| drinks             | 0.698537294767418  | 0.698537294767418  | 0.698537294767418  |
| ice cream          | 0.7862064089630107 | 0.7862064089630107 | 0.7862064089630107 |
| transfer with refer| 1                  | 1                  | 1                  |
| transfer with dial | 1                  | 1                  | 1                  |
| play               | 1                  | 1                  | 1                  |
| DTMF               | 1                  | 1                  | 1                  |

```

Result:

<Frame>
  <img alt="Bar chart displaying intent score ranges with minimum, maximum, and average values" />
</Frame>

### Top Intent Scores

Indicates the top-rated numbers of Intents displayed in a bar chart.
For more information, see [Intent Analyzer](/ai/platform-features/nlu/intents/intent-analyzer).

Source table:

```txt theme={null}
|  Intents             | Intent Score       |
| -------------------- | ------------------ |
| play                 | 1                  |
| transfer with dial   | 1                  |
| transfer with refer  | 1                  |
| DTMF                 | 1                  |
| ice cream            | 0.7862064089630107 |
| drinks               | 0.698537294767418  |
| pizza                | 0.6863160928710601 |
| cakes                | 0.6517875252045072 |
```

Result:

1. Intent Score Max: `1`
2. Intent Score Min: `0.65`

<Frame>
  <img alt="Bar chart ranking intents by their confidence scores" />
</Frame>

## Line charts

### Average Execution Time

Displays the average time in milliseconds it takes to execute a Flow process over a selected time span.

Formula:

Average Execution Time = `Sum of Execution Times / Total number of inputs`

Source table:

```txt theme={null}
|  Date        | Execution Time  |
| ------------ | --------------- |
| 2023-09-13   |      130        |
| 2023-09-14   |      0          |
| 2023-09-15   |      123        |
| 2023-09-16   |      0          |
| 2023-09-17   |      0          |
| 2023-09-18   |      0          |
| 2023-09-19   |      0          |
```

Result: `(130ms + 123ms) / 2 = 127.5ms`

<Frame>
  <img alt="Line chart showing average Flow execution time in milliseconds over a time span" />
</Frame>

## Download Report

To download a report as a CSV file for the chart,
click <img alt="menu" /> in the upper-right corner of the chart.
To create a report file, refer to the [Download Report](https://docs.cognigy.com/insights/download-reports/) article.

## More Information

* [Engagement Report](/insights/reports/engagement)
* [Overview Report](/insights/reports/overview)
* [Live Agent Report](/insights/reports/live-agent)


# Overview
Source: https://docs.cognigy.com/insights/reports/overview

Get a quick summary of key AI Agent performance metrics, including active sessions, top Intents, and user interactions in Cognigy Insights.

<a href="/release-notes/2026.14"><Badge>Updated in 2026.14</Badge></a>

The *Overview Report* gives you a quick summary of how customers are interacting with your AI Agent.

The summary of key metrics is represented in the following charts:

* [Indicators](/insights/reports/engagement#indicators)
* [Bar charts](/insights/reports/engagement#bar-charts)
* [Line charts](/insights/reports/engagement#line-charts)

All charts display data based on the global filter settings you've selected.

<Note>
  This report's metrics include sessions with personal data masked using the [Delete Profile](/ai/agents/develop/node-reference/analytics/delete-profile) or [Deactivate Profile](/ai/agents/develop/node-reference/analytics/deactivate-profile) Nodes.
</Note>

<Frame>
  <img alt="Overview report displaying key metrics including indicators, bar charts, and line charts" />
</Frame>

## Indicators

### Active Sessions

Indicates the total number of currently active user sessions in the last 15 minutes. This widget updates in a regular interval of 30 seconds.

### Total Sessions

Indicates the total number of sessions.

### Positive Ratings

Indicates the total number of Positive Ratings given as feedback by the end user.

Source table:

```txt theme={null}
| Sessions    | Positive Rating | Negative Rating | Rated Sessions |
| ----------- | --------------- | --------------- | -------------- |
| session-123 |        1        |        0        |       1        |
| session-456 |        0        |        1        |       1        |
| session-789 |        0        |        0        |       0        |
| session-abc |        1        |        0        |       1        |
| session-def |        1        |        0        |       1        |
| ---------------------------------------------------------------- |
| TOTAL       |        3        |        1        |       4        |
```

Result: `3`

## Ring charts

### Sessions by Channel

Indicates the total number of conversations displayed in a pie chart based on the channels used.

Source table:

```txt theme={null}
| Sessions    | Channel          | Conversations   | Percentage     |
| ----------- | ---------------  | --------------- | -------------- |
| session-123 | Demo Webchat     |        14       |       46.67    |
| session-456 | Voice Gateway    |        12       |       40       |
| session-789 | Interaction Panel|        3        |       10       |
| session-abc | Webchat          |        1        |       3.33     |
| ----------------------------------------------------------------- |
| TOTAL       |                  |        30       |       100      |
```

Result:

<Frame>
  <img alt="Ring chart showing sessions distributed by channel type" />
</Frame>

## Bar charts

### Top Intents

Indicates the top-rated number of used Intents, displayed in a horizontal bar chart. The chart showcases the top-rated number of Intents used, with a maximum of 6 Intents shown in the summary view and 25 in the detailed view. Whenever an Intent is matched, it is recorded in the analytics data. The data is then grouped by the matched Intent and sorted in descending order.

Source table:

```txt theme={null}
|  Intent           | Total number  |
| ----------------- | ------------- |
| Flight discounts  |      12       |
| Book a flight     |      5        |
| Change a flight   |      5        |
| Additional fee    |      4        |
| Reject a flight   |      3        |
| Transfer flights  |      3        |
```

Result:

<Frame>
  <img alt="Horizontal bar chart ranking the most frequently matched Intents" />
</Frame>

### Users by Locale

Indicates time-related recognized conversations depending on different locales given in percentages.

Source table:

```txt theme={null}
| Date       | Conversations   | Locales         | Percentage      |
| -----------| --------------- | --------------- | --------------- |
| 2023-08-27 | 0               |     -           |  -              |
| 2023-08-28 | 1               | "en-US"         | 100             |
| 2023-08-29 | 4               | "en-US"         | 100             |
| 2023-08-30 | 27              | "en-US"         | 100             |
| 2023-08-31 | 1               | "en-US"         | 100             |
| 2023-09-01 | 1               | "en-US"         | 100             |
| 2023-09-02 | 0               |     -           |  -              |
```

Result:

<Frame>
  <img alt="Bar chart showing conversation distribution by user locale over time" />
</Frame>

### Top Tasks

Indicates the number of sessions in which each task was completed. The task completed in the most sessions is ranked first. A task completed several times within the same session is counted once.

Source table:

```txt theme={null}
| Sessions    | Tasks                  |
| ----------- | ---------------------- |
| session-123 | Task A, Task B, Task D |
| session-456 | Task B, Task C, Task E |
| session-789 | Task A, Task B, Task E |
| session-abc | Task A                 |
| session-def | Task, Task A           |
| session-ghi | Task, Task A           |
| session-jkl | Task A, Task B, Task A |
```

In `session-jkl`, Task A appears twice but is counted only once.

Result:

<Frame>
  <img alt="Bar chart ranking tasks by the number of sessions in which they were completed" />
</Frame>

### Top Slots

Similar to Top Intents, this chart displays the top-rated number of Slots in a bar chart.

Source table:

```txt theme={null}
| Slot        | Number      |
| ----------- | ----------- |
| DATE        | 2           |
| NUMBER      | 1           |
```

Result:

<Frame>
  <img alt="Bar chart showing the most frequently filled Slots" />
</Frame>

### Number of Inputs

Displays the number of inputs displayed over a period of time.

Source table:

```txt theme={null}
| Date       | Inputs       | 
| -----------| ------------ |
| 2023-08-27 | 0            | 
| 2023-08-28 | 1            |
| 2023-08-29 | 23           | 
| 2023-08-30 | 72           | 
| 2023-08-31 | 8            | 
| 2023-09-01 | 3            | 
| 2023-09-02 | 0            | 
```

Result:

<Frame>
  <img alt="Bar chart displaying the number of user inputs over a period of time" />
</Frame>

## Line charts

### Sessions

Displays the number of sessions on a timescale.

Source table:

```txt theme={null}
| Date       | Sessions     | 
| -----------| ------------ |
| 2023-08-27 | 0            | 
| 2023-08-28 | 1            |
| 2023-08-29 | 4            | 
| 2023-08-30 | 30           | 
| 2023-08-31 | 6            | 
| 2023-09-01 | 1            | 
| 2023-09-02 | 0            | 
```

Result:

<Frame>
  <img alt="Line chart showing the number of sessions over a timescale" />
</Frame>

### Unique Contacts

Displays the number of unique users (`userid`) over a selected time span. If a user has multiple sessions in a single day, they are counted as one unique contact for that day.

Source table:

```txt theme={null}
| Date    	| Contacts |
| --------- | -------- |
| 1/16/2023	|    1     |
| 1/17/2023	|    8     |
| 1/18/2023	|    15    |
| 1/19/2023	|    6     |
| 1/20/2023	|    18    |
```

Result:

<Frame>
  <img alt="Line chart showing the number of unique contacts per day" />
</Frame>

## Download Report

To download a report as a CSV file for the chart,
click <img alt="menu" /> in the upper-right corner of the chart.
To create a report file, refer to the [Download Report](https://docs.cognigy.com/insights/download-reports/) article.

## More Information

* [Engagement Report](/insights/reports/engagement)
* [NLU Performance Report](/insights/reports/nlu-performance)
* [Live Agent Report](/insights/reports/live-agent)


# Questions and Answers
Source: https://docs.cognigy.com/live-agent/FAQs

Answers to common Live Agent questions covering user removal, roles, inbox management, OData reporting, and troubleshooting integration issues.

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

## Questions and Answers about Live Agent

### How to remove a user?

Removing a user (human agent) from Live Agent can be done from the Cognigy [Members page](/ai/administer/access/members). The `liveAgentAdmin`/`liveAgentAgent` roles need to be removed, and this will prevent the human agent from logging into Live Agent. This can also be achieved by deleting the user from Cognigy.

Head over to [Cognigy Access Control](/ai/administer/access/admin-center/access-control) for more information.

### How to create an Agent with multiple roles?

To create an Agent with multiple roles, read the [Roles](/live-agent/roles#assign-roles) article.

### What happens if an Inbox is deleted?

Inboxes created manually can be safely deleted, except those with the `-default` suffix. Deleting Inboxes with `-default` breaks the integration between Live Agent and Cognigy. If this Inbox is accidentally deleted, see the solution in the question below.

### What should I do if the Inbox is deleted?

You need to create a new Inbox. To do that, use the [Adding another Inbox for your Project](/live-agent/getting-started/live-agent-setup/live-agent-setup-additional-inbox) article.

### Where can I check what the default Inbox is? Does it have a special account id?

The default Inbox has a pattern name `XXXX-default`, for example, `7119-default`. You can rename it in [Inbox Settings](/live-agent/settings/inboxes).

### How to fix custom SSL certificate verification error?

That may happen because Cognigy is running under a custom SSL certificate (Error cannot validate certificate). And Live Agent is failing to perform the requests to it as SSL requests cannot be verified.

Head over to [Rest Client SSL Helm values](/live-agent/installation/helm-values/helm-values#rest-client-ssl) for configuring it.

### How to fix SMTP for no authentication provider?

For on-premises installations, the SMTP configuration needs to be modified.

Head over to [Helm SMTP values](/live-agent/installation/helm-values/smtp) for more information.

### How to create labels and assign them to conversations?

Once a label has been created, navigate to a conversation. There is a section called ***Conversation Labels*** on the right side, where labels can be added.

Head over to [Labels](/live-agent/settings/labels) for more information.

### How to request a file from a conversation?

Head over to [Attachments](/live-agent/settings/attachments#request-attachments-with-the-plugin) for more information.

### How to add Agents to an Inbox?

Only users with `Administrator` roles can perform this action.

To add an Agent to an Inbox, follow these steps:

1. In the left-side menu of the Live Agent interface, navigate to **Settings**, then select **Inboxes**.
2. Locate the Inbox that you want to edit and click the **Settings** icon.
3. On the **Collaborators** tab, select the Agent name from the **Agents** list.
4. Apply changes by clicking **Update**.
   Head over to [Inboxes](/live-agent/settings/inboxes#add-agents-to-inboxes) for more information.

### How to create a team and add members to it?

Navigate to **Settings > Teams** and click the **Create new team** button. In the second step, add the desired Agents. The added Agents can be updated later.

Head over to [Teams](/live-agent/settings/teams) for more information.

### How to get the user access token (API Key)?

The user access token is displayed at the bottom of the [Profile Settings](/live-agent/profile-settings).

### How to view the account ID and the inbox ID?

This information can be obtained from the Live Agent URL.

A quick way is to select an Inbox on the dashboard screen. The URL will look like: `https://<live-agent-domain>/app/accounts/1/inbox/4`.

The URL follows the pattern `accounts/{Account ID}/inbox/{Inbox ID}`, where the account ID is `1` and inbox ID is `4`.

### Can a TTL be set for attachments contained in messages?

Yes, a TTL for attachments contained in messages per account can be defined under the [Account Settings](/live-agent/settings/account-settings#ttl-for-attachments).

### Is antivirus file upload scanning enabled for App and Trial environments?

The antivirus file upload scanning needs to be configured at the infrastructure level, and it is not enabled for the Trial and App instances at the moment.
Head over to [File Upload Antivirus Scan Helm Values](/live-agent/installation/helm-values/helm-values#file-upload-antivirus-scan) for learning how to configure it.

### How to navigate to Cognigy from Live Agent?

Under the profile menu, there is a link called Cognigy that heads over to the Cognigy. If it is not present, head over to [URLs Helm values](/live-agent/installation/helm-values/helm-values#urls) for configuring it.

<Frame>
  <img alt="Link to Cognigy" />
</Frame>

### Get a Google Maps API key

To see a location sent from WhatsApp with Google Maps, you need to get a [Google Maps API key](https://developers.google.com/maps/documentation/javascript/get-api-key). Head over to the [App](/live-agent/installation/helm-values/helm-values#app) to configure it.

## Live Agent OData service

### Why is the OData service not returning all the information from the database?

The OData service returns data belonging to your account.

Head over to [OData Endpoint](/live-agent/tools/odata-endpoint) for more information.

### Why is the OData endpoint throwing an unauthorized error while querying?

If you encounter an unauthorized error, ensure that the API access token belongs to a user with the `Administrator` role for the account being used in the query.

### How to retrieve all content from a particular collection?

You can retrieve all content from a particular collection, for example, Conversation, Message.

The following filtering query can be performed:

For example, `http://{{host}}:{{port}}/v1.0/odata/Conversation?$filter=created_at le '2126-11-23T00:00:00'`.

It will return all the fields for the Conversation model created before 23/11/2126.

Head over to [OData Endpoint](/live-agent/tools/odata-endpoint) for more information.

### How to filter OData with specific fields?

To filter OData results by specific fields, you can use the `$select` and `$filter` query parameters.

For example, `http://{{host}}:{{port}}/v1.0/odata/Conversation?$select=id,account_id&$filter=created_at le '2026-11-23T00:00:00'`.

In this example, the request is retrieving the `id` and `account_id` fields from the Conversation model, filtered by the `created_at` field being less than or equal to `2026-11-23T00:00:00`.

For more detailed information, refer to the [OData Endpoint](/live-agent/tools/odata-endpoint) article.


# Cognigy Live Agent API
Source: https://docs.cognigy.com/live-agent/api

The Cognigy Live Agent API provides access to conversations, teams, and agents. It enables you to integrate, automate, and analyze workflows beyond the Live Agent GUI.

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

The [Cognigy Live Agent API](https://liveagent-trial.cognigy.ai/openapi) provides access to conversations, teams, and agents. It enables you to integrate, automate, and analyze workflows beyond the Live Agent GUI.

## Access the API

The API Base URL and an access token are required to access the API.

### Choose the API Base URL

The API Base URL is the endpoint for accessing the Cognigy Live Agent OpenAPI.
The table outlines the base URLs for different environments.

| Environment | API Base URL                                                                                                                                                                                                                                                                                                                                                                                                                            |
| ----------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Trial       | `https://liveagent-trial.cognigy.ai`                                                                                                                                                                                                                                                                                                                                                                                                    |
| SaaS        | Choose the environment you use:<ul><li>**App (app.cognigy.ai)** - `https://liveagent-app.cognigy.ai`</li><li>**App-US** - `https://liveagent-app-us.cognigy.ai`</li> <li>**Other SaaS** - use the `https://api-<company-name>.liveagent.cloud/openapi` pattern where `company-name` is your company name. For example, if your company name is `abc123`, the URL would be:<br /> `https://api-abc123.liveagent.cloud/openapi`</li></ul> |
| On-premises | In your Live Agent `values.yaml` file, replace live-agent.example.com in `host: "live-agent.example.com"` with the API URL from your installation.                                                                                                                                                                                                                                                                                      |

### Get an Access Token

To use the Live Agent API, you need to get an access token. The token must be passed for each method in the HTTP-Authorization header.

1. In the bottom-left corner of the Live Agent interface, click your profile icon and select **Profile Settings**.
2. In the **Access Token** section, copy the token and save it for later use.

## Make your First API Request

Create a request to get conversation metadata. Select one of the options to make the request:

<Tabs>
  <Tab title="cURL">
    1. Open your terminal to create a `GET` request:
       ```bash theme={null}
       curl -X GET "https://{api_base_url}/api/v1/accounts/{account_id}/conversations/meta" \
       -H "api_access_token: your_access_token"
       ```
    2. Replace the following placeholders in request:
       * Replace `{api_base_url}` with the base URL of your environment. For example, for the trial environment, use `liveagent-trial.cognigy.ai`.
       * Replace `{account_id}` with your Live Agent account ID. Find the ID in the Live Agent URL. For example, in `https://liveagent-trial.cognigy.ai/accounts/12345/conversations`, the account ID is `12345`.
       * Replace `your_access_token` with the access token you saved.
  </Tab>

  <Tab title="Postman">
    1. Open Postman and create a `GET` request with the following URL: `https://{api_base_url}/api/v1/accounts/{account_id}/conversations/meta`.
    2. Replace the following placeholders in request:
       * Replace `{api_base_url}` with the base URL of your environment. For example, for the trial environment, use `liveagent-trial.cognigy.ai`.
       * Replace `{account_id}` with your Live Agent account ID. Find the ID in the Live Agent URL. For example, in `https://liveagent-trial.cognigy.ai/accounts/12345/conversations`, the account ID is `12345`.
    3. Add the authorization header:
       1. Go to the **Headers** tab.
       2. Add the `api_access_token` key and set the value to `your_access_token`.
    4. Click **Send** to execute the request.
       You should receive a `200 OK` response containing the following JSON:

    ```json theme={null}
    {
        "meta": {
            "mine_count": 0,
            "assigned_count": 20,
            "unassigned_count": 512,
            "all_count": 532
        }
    }
    ```
  </Tab>
</Tabs>

## Troubleshooting

<Accordion title="Error 401 (unauthorized)">
  If you receive a `401 Unauthorized` error, ensure that you have included the correct access token in the request header.
</Accordion>

## More Information

* [Account Settings](/live-agent/settings/account-settings)


# Agent Copilot in Live Agent
Source: https://docs.cognigy.com/live-agent/assistants/agent-copilot

Use Agent Copilot in Live Agent to provide human agents a dedicated workspace with AI tools, knowledge bases, and customer context during handovers.

<a href="/release-notes/4.51"><Badge>Updated in 4.51</Badge></a>

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

## Before you Begin

Check the Agent Copilot availability for your installation:

* For trial and SaaS, Agent Copilot is automatically activated.
* For on-premises, install Agent Copilot with [Cognigy Helm Chart](https://github.com/Cognigy/cognigy-ai-helm-chart).

## Set up Flows

To provide additional real-time information to your agents, this assistant uses a Flow.

To set up an assistant, you need to have two separate flows:

* The Main Flow where the primary logic is defined.
* A Flow where the logic for your assistant will be implemented.

During an active Live Agent handover, this Flow will process all customer inputs and display the outputs as part of the conversation for the human agent.

Assistant's messages are only shown based on customer inputs received during an active handover after the human agent has initially responded.

<Warning>
  You don't need to include the Handover Node in the assistant's Flow, as using the Handover to Human Agent Node in the main Live Agent Flow is enough for the configuration process.

  To provide helpful information to your agents, create Intents and use Lookup Nodes with Say Nodes based on customer inputs.

  <Frame>
    <img alt="Agent assist Flow with Lookup Nodes and Say Nodes for customer inputs" />
  </Frame>
</Warning>

## Set up Endpoints

To configure your Endpoint:

1. Go to **Deploy > Endpoints**.
2. Go to your **Webchat** Endpoint.
3. Scroll down to **Handover Settings** and select **Cognigy Live Agent** from the list.
4. Go to the **Copilot** section.
5. In the **Copilot Config** field, select a config you recently created via [API](https://api-trial.cognigy.ai/openapi#post-/v2.0/agentassistconfigs).
6. Click **Save**.
7. In Live Agent, go to **Profile settings > Agent Copilot** and select **Enable Agent Copilot Workspace**. Alternatively, you can activate the Workspace for all human agents in your project. To do this, go to **Settings > Account Settings > Agent Copilot Workspace** and select **Enable Agent Copilot Workspace**, then click **Update Settings**.

The Endpoint setup is completed.

## Test Agent Copilot Workspace

To test Agent Copilot, follow these steps:

1. Start a Webchat session.
2. Complete the handover process to Live Agent.
3. Once the handover is complete, send a message as a human agent.
4. After that, send a message as the user in Webchat.

<Frame>
  <img alt="Webchat session during a handover to a Live Agent human agent" />
</Frame>

Once the human agent receives this message, the testing behavior will vary depending on how you configured the Endpoint beforehand.

This Workspace is accessible from the right side of the conversation chat.

<Frame>
  <img alt="Agent Copilot workspace panel on the right side of the conversation chat" />
</Frame>

If the testing is successful, the Agent Copilot setup is complete.

## Troubleshooting

### Transcript Tile Not Sending Messages

To send messages from the Transcript Tile chat to the end user's conversation chat, assign a human agent.
If the conversation remains unassigned, the Transcript Tile will not operate properly, as there will be no human agent to associate with the message when it is created.

## More Information

* [Agent Copilot Overview](/agent-copilot/overview)
* [Getting started with Agent Copilot](/agent-copilot/getting-started/chat)


# Agent Copilot Bot
Source: https://docs.cognigy.com/live-agent/assistants/agent-copilot-bot

Set up Agent Copilot Bots as personal virtual assistants for human agents in Live Agent to automate workflows and assist with complex customer inquiries.

<a href="/release-notes/4.67"><Badge>Updated in 4.67</Badge></a>

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

*Agent Copilot Bots* are personal virtual assistants for human agents. These assistants automate service processes, give human agents access to relevant resources and assist in answering common customer questions.

Human agents can interact with Agent Copilot Bots in every customer conversation to solve complex customer issues more efficiently and faster to improve customer experience.

Only users with the `Administrator` role can create, edit, or delete Agent Copilot Bots in Live Agent.

Differing from other Agent Copilot assistants, a human agent can use multiple Agent Copilot Bots within a single conversation.
While the Agent Copilot assistants listen to the conversation and attempt to follow it,
an Agent Copilot Bot acts as an independent secondary bot with built-in logic to assist the human agent.
For instance, if the human agent requires additional expertise during the conversation,
they can activate a specialized Agent Copilot Bot, such as one for technical support or order processing.

<Frame>
  <img alt="Live Agent Agent Copilot Bot Overview" />
</Frame>

## Before you Begin

To use Agent Copilot Bots in a Live Agent conversation, you need to create two Flows and two Endpoints in Cognigy that comply with the following conditions:

* The **Main** Flow for interaction between an AI Agent and an end user. Within this Flow, add a [Handover to Human Agent](/ai/agents/develop/node-reference/service/handover-to-agent) Node to transition the conversation from the AI Agent to a human agent. The Inbox ID must be included in the Handover to Human Agent Node.
* The **Endpoint** for the **Main** Flow.
* The **Agent Copilot Bot** Flow for interaction between an Agent Copilot Bot and a human agent.
* The **Endpoint** for the **Agent Copilot Bot** Flow.

## Create Agent Copilot Bots

You can create multiple Agent Copilot Bots within an Inbox.

To create an Agent Copilot Bot, do the following:

1. In the Live Agent interface, go to **Settings > Inboxes**.
2. To edit the previously created Inbox settings, click **Settings** next to the Inbox.
3. On the **Settings** tab, click **Add Bots**.
4. In the **Create Agent Copilot Bot** window, fill in the following fields:
   * **Name** — enter a unique name for your bot. This name is helpful if there are multiple bots within the same Inbox, allowing you to distinguish them easily.
   * **Endpoint Config URL** — enter the URL copied from the Endpoint settings in Cognigy. You can find and copy the **Config URL** in the Webchat Endpoint settings for the previously created Agent Copilot Bot Flow in Cognigy.
5. Click **Add**.

The created Agent Copilot Bot will be available for all assigned human agents within the Inbox.

## Use Agent Copilot Bots in Conversations

Note that you can use multiple Agent Copilot Bots within a conversation. Each bot creates a separate session that needs to be [manually closed](#complete-agent-copilot-bot-sessions) after ending the conversation.

To use Agent Copilot Bots in a conversation, do the following:

1. Go to your conversation.
2. At the top of the conversation window, click **More Details**.
3. On the right side of the **Conversation Details** window, select the **Copilot** tab.
4. Open the dropdown list and select one of the available **Agent Copilot Bots**.
5. In the **Edit Message** field at the bottom of the window, write your request to the Agent Copilot Bot. The Agent Copilot Bot performs research as configured in the Agent Copilot Bot Flow and sends back the result in a separate Live Agent chat, such as a product video, image, text, and more.
6. To forward the provided content to the Live Agent conversation, click directly on the content item. It will be sent to the **Reply to User** field in the main conversation chat.
7. Click **Send** to provide the content item to the customer in the main conversation chat.

This way, the Agent Copilot Bot becomes a dependable assistant for human agents, enabling them to delegate tasks to the bot and then share the outcomes with the customer.

## Complete Agent Copilot Bot Sessions

To complete a session with Agent Copilot Bot, do the following:

1. Go to your conversation.
2. At the top of the conversation window, click **More Details**.
3. On the right side of the conversation window, select the **Copilot** tab.
4. Click <img alt="menu" />.
5. In the dropdown list, select **Reset Session**.

The session will stop, the interaction with the bot will finish, and the Agent Copilot Bot communication history will be cleared.

## Send Transcripts to Assigned Agents

In case you want to send a conversation transcript to a specific email address or to the assigned human agents, do the following:

1. Go to your conversation.
2. In the upper-right corner of the conversation window, click **Send Transcript**.
3. In the **Send conversation transcript** window, select one of the following options:
   * **Send transcript to the assigned agent** — the conversation transcript will be sent to a human agent currently assigned to handle the interaction. This is useful when you want the assigned human agent to have a record of the conversation for reference or follow-up.
   * **Send transcript to another email address** — the conversation transcript will be sent to a different email address. This is helpful when you need to share the conversation details with someone other than the assigned human agent, such as a supervisor, team member, or any other relevant party.
4. Click **Submit**.

A transcript will be sent to the selected receiver.

<Note>
  The email transcript includes only the conversation history between a human agent and their customer. If the human agent calls an Agent Copilot Bot for help, that interaction takes place on a separate channel and in a different session, and it will not be included in the transcript.
</Note>

## More Information

* [Assistants](/live-agent/assistants/overview)
* [Endpoints](/ai/agents/deploy/endpoints/overview)
* [Inboxes](/live-agent/settings/inboxes)


# Agent Copilot Whisper
Source: https://docs.cognigy.com/live-agent/assistants/agent-copilot-whisper

Configure Agent Copilot Whisper in Live Agent to provide human agents with AI-powered reply suggestions and next-action hints directly in the chat window.

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

## Set up Flows

To provide additional real-time information to your agents, this assistant uses a Flow.

To set up an assistant, you need to have two separate flows:

* The Main Flow where the primary logic is defined.
* A Flow where the logic for your assistant will be implemented.

During an active Live Agent handover, this Flow will process all customer inputs and display the outputs as part of the conversation for the human agent.

Assistant's messages are only shown based on customer inputs received during an active handover after the human agent has initially responded.

<Warning>
  You don't need to include the Handover Node in the assistant's Flow, as using the Handover to Human Agent Node in the main Live Agent Flow is enough for the configuration process.

  To provide helpful information to your agents, create Intents and use Lookup Nodes with Say Nodes based on customer inputs.

  <Frame>
    <img alt="Agent assist Flow with Lookup Nodes and Say Nodes for customer inputs" />
  </Frame>
</Warning>

### Supported Output Types

Agent Copilot Whisper supports all output types that can be selected in the [Say](/ai/agents/develop/node-reference/overview), [Question](/ai/agents/develop/node-reference/overview), or [Optional Question](/ai/agents/develop/node-reference/overview) Nodes.

## Set up an Endpoint

To configure your Endpoint:

1. Go to **Deploy > Endpoints**.
2. Go to your **Webchat** Endpoint.
3. Scroll down to **Handover Settings** and select **Cognigy Live Agent** from the list.
4. Go to the **Copilot** section.
5. In the **Copilot Config** field, select **None**.
6. In the **Copilot Flow** field, select the name of the Agent Copilot Flow you've created in the previous step.
7. Click **Save**.

The Endpoint setup is completed.

## Test Agent Copilot Whisper

To test Agent Copilot Whisper, follow these steps:

1. Start a Webchat session.
2. Complete the handover process to Live Agent.
3. Once the handover is complete, send a message as a human agent.
4. After that, send a message as the user in Webchat.

<Frame>
  <img alt="Webchat session during a handover for testing Agent Copilot Whisper" />
</Frame>

Once the human agent receives this message, the testing behavior will vary depending on how you configured the Endpoint beforehand:

1. The human agent will receive a dropdown menu **Agent Copilot**.
2. In this menu, the human agent will find the results from the Flow sent via Agent Copilot Whisper.
   <Frame>
     <img alt="Agent Copilot dropdown menu showing whisper results from the Flow" />
   </Frame>
3. The human agent can click the blue arrow to forward content to the user.

   <Frame>
     <img alt="Blue arrow button to forward Agent Copilot content to the user" />
   </Frame>

   <Frame>
     <img alt="Webchat showing the forwarded Agent Copilot Whisper response to the end user" />
   </Frame>

If the testing is successful, the Agent Copilot Whisper setup is complete.

## More Information

* [Agent Copilot in Live Agent](/live-agent/assistants/agent-copilot)
* [Agent Copilot Overview](/agent-copilot/overview)


# Live Agent Assistants
Source: https://docs.cognigy.com/live-agent/assistants/overview

Overview of Live Agent assistant types, including Agent Copilot Bot, Whisper, and Copilot workspace, to support human agents during customer interactions.

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

Assistants in Cognigy are tools designed to support human agents in their interactions with customers. They often involve AI-powered features, helping agents provide more effective assistance and enhance the overall customer experience.

With assistants, among other things, the following can be achieved:

* Provide contextual information for the agent (CRM records, ERP data, small talk clues, and more).
* Enable quick access to resources such as knowledge base articles.
* Suggest replies (including media) that can be forwarded to the customer.

To set up your Project for assistants, choose a virtual assistant that best suits your requirements:

* [Agent Copilot Bot](/live-agent/assistants/agent-copilot-bot). Refers to the personal tool for human agents. It offers a unique Flow for agents to access additional information as needed.
* [Agent Copilot Whisper](/live-agent/assistants/agent-copilot-whisper). Refers to the functionality that provides assistance to live agents during their interactions with customers. It typically includes real-time guidance, suggestions, and recommendations to help agents effectively handle customer inquiries or issues. Agent Copilot Whisper focuses on providing support directly from the conversation window.
* [Agent Copilot](/live-agent/assistants/agent-copilot). Refers to the dedicated workspace or environment where human agents access various tools, resources, and features to assist them in their customer interactions. It encompasses the entire workspace setup, which may include customer information, knowledge bases, collaboration tools, and other relevant components. Agent Copilot aims to provide a comprehensive and efficient working environment for human agents to manage their interactions with customers.

## More Information

* [Agent Copilot Overview](/agent-copilot/overview)


# Conversation Assignment
Source: https://docs.cognigy.com/live-agent/conversation/assign-conversations

Learn how to assign agents, teams, priority, and labels to conversations in Live Agent. Transfer and manage conversation routing between human agents.

<a href="/release-notes/2026.1"><Badge>Updated in 2026.1</Badge></a>

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

Within Live Agent, you can assign agents, teams, priority, labels, and change inboxes.

<Frame>
  <img alt="Live Agent Conversation Actions" />
</Frame>

## Assign Agents

You can transfer a conversation to another human agent by selecting a different agent from the **Assigned Agent** list.
The entire conversation will be visible to the newly selected agent.

To assign an agent to a conversation within Live Agent, do the following:

1. In the left-side menu of Live Agent, click **Conversations**.
2. From the **Inboxes** list, select an Inbox.
3. From the **Conversations** list, select your conversation where you want to assign a priority.
4. On the right side of the conversation chat, go to **Details > Conversation Actions**.
5. From the **Assigned Agent** list, select a different agent by scrolling or typing to search for an agent, then click the intended agent.

The agent will be assigned to the conversation.

### Assign Supervisors

You can select an agent with any role in the **Assigned Agent** list. However, if you want to specifically assign a supervisor, you can do so by escalating the conversation.

To escalate a conversation to a human agent with the `Supervisor` role, click the **↑ Escalate to Supervisor** button. The button appears in the **Conversation Actions** section if the Inbox settings [allow agents to escalate conversations to supervisors](/live-agent/settings/inboxes#escalate-conversations-to-supervisors).
The supervisor will be assigned to the conversation in the Inbox.

Additionally, supervisors can track the number of escalated conversations in the [Supervisor Overview Report](/live-agent/reports/supervisor-overview).

## Assign Teams

To assign a team to a conversation within Live Agent, do the following:

1. In the left-side menu of Live Agent, click **Conversations**.
2. From the **Inboxes** list, select an Inbox.
3. From the **Conversations** list, select your conversation where you want to assign a priority.
4. On the right side of the conversation chat, go to **Details > Conversation Actions**.
5. From the **Assigned Team** list, select a desired team by scrolling or typing to search for a team, then click the intended team.

The team will be assigned to the conversation.

## Assign Priority

To assign a priority to a conversation within Live Agent, do the following:

1. In the left-side menu of Live Agent, click **Conversations**.
2. From the **Inboxes** list, select an Inbox.
3. From the **Conversations** list, select your conversation where you want to assign a priority.
4. On the right side of the conversation chat, go to **Details > Conversation Actions**.
5. From the **Priority** list, select a desired priority by scrolling or typing to search for a priority level, then click the intended level.

The priority will be assigned to the conversation.
The icon representing the matching priority level will now be visible next to the conversation in the Conversations list.

## Assign Labels

To assign a label to a conversation within Live Agent, do the following:

1. In the left-side menu of Live Agent, click **Conversations**.
2. From the **Inboxes** list, select an Inbox.
3. From the **Conversations** list, select your conversation where you want to assign a priority.
4. On the right side of the conversation chat, go to **Details > Conversation Actions**.
5. Under the **Conversation Labels** section, click **Add Labels**.
6. Scroll or type to search for a label, then click a label. You can add multiple labels at once.

The label will be added to the conversation.

## Change Inboxes

When a conversation is initiated, it is automatically assigned to the inbox defined in the Cognigy Flow. However, you can change the inbox for an existing conversation if needed.

Agents can only assign conversations to the inboxes they have access to. Supervisors and Administrators, who have access to all inboxes, can assign conversations to any inbox.

To change an inbox for a specific conversation within Live Agent, follow these steps:

1. In the left-side menu of Live Agent, click **Conversations**.
2. From the **Inboxes** list, select an Inbox. All available inboxes are visible depending on the user's role.
3. From the **Conversations** list, select the conversation where you want to change the inbox.
4. On the right side of the conversation chat, go to **Details > Conversation Actions**.
5. From the **Change Inboxes** list, select the desired inbox.

After selecting a new inbox, the page will refresh to ensure all agents can see the conversation in the newly assigned inbox.

If you want to change inboxes for multiple conversations, follow these steps:

1. In the left-side menu of Live Agent, click **Conversations**.
2. From the **Conversations** list, select the conversations where you want to change the inbox and click <img alt="change inbox" />.
3. From the **Inboxes** list, select an Inbox. All available inboxes are visible depending on the user's role.
4. Confirm the action by clicking **Assign**.

After selecting a new inbox, the page will refresh to ensure all agents can see the conversations in the newly assigned inbox.

## More Information

* [Manual Assignment](/live-agent/conversation/conversation-routing/manual-mode)
* [Automatic Assignment](/live-agent/conversation/conversation-routing/automatic-mode)


# Conversation Queue
Source: https://docs.cognigy.com/live-agent/conversation/conversation-queue/overview

Understand how the Live Agent conversation queue works, including queue priorities, escalated conversations, and human agent capacity limits.

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

*Conversation Queue* is a list of conversations awaiting agent attention. In Live Agent, a queue works like a waiting area, where conversations are held until an agent sends the first reply.

If a conversation limit is set, assigned conversations appear on the **Mine** tab, while unassigned conversations waiting for an available agent appear on the **Unassigned** tab.

In the conversation queue, conversations can have different priorities:

* **Standard Conversations** – new conversations are added to the agent's queue. An agent with the `Agent` role takes ownership of the conversation when they send the first reply.
* **Escalated Conversations** – when a [conversation is escalated](/live-agent/conversation/conversation-routing/manual-mode), its priority changes to `Escalation`. Escalated conversations move to the top of the same queue but remain subject to conversation limits. An agent with the `Supervisor` role automatically takes ownership of the conversation after escalation.

## Examples

The provided examples show how the conversation queues work:

The [conversation limits](/live-agent/settings/account-settings#conversation-limit-per-agent) and specific queue behavior may vary depending on the Live Agent configuration and settings in use.

**Example 1**

Conditions:

* Conversations are assigned to an agent, but they still need to send the first reply.
* The Conversation limit is set to 2 conversations per agent.
* There is one agent available.
* 3 conversations are created.

Result: The first two conversations are assigned to the agent and are moved to the queue, awaiting the agent's reply. One conversation remains unassigned.

**Example 2**

Conditions:

* Conversations are assigned to an agent, but they still need to send the first reply.
* The Conversation limit is turned off, meaning there is no limit on the number of conversations per agent.
* There is 1 agent available.
* 3 conversations are created.

Result: All three conversations are assigned to the agent and are moved to the queue, waiting for the agent to send the first reply. None of the conversations has been assigned yet.

**Example 3**

Conditions:

* Both settings in the Inbox are activated: Automatic Conversation Assignment and Consider Conversation Priority.
* Conversations are assigned to an agent, but they still need to send the first reply.
* The Conversation limit is set to 2 conversations per agent.
* There is one agent available.
* 3 conversations with different [priorities](/live-agent/conversation/conversation-routing/automatic-mode#additional-automatic-assignment-parameters) are created:
* Conversation 1 is Medium.
* Conversation 2 is Low.
* Conversation 3 is High.

Result: The agent is assigned Conversation 3 (High), followed by Conversation 1 (Medium), both in the queue. Conversation 2 (Low) remains unassigned.

**Example 4**

Escalation behavior may vary depending on Live Agent configuration and settings in use.

Conditions:

* Conversations are assigned to an agent, but the agent hasn't sent the first reply.
* The conversation limit is set to 2 conversations per agent.
* One agent with the **Agent** role and one agent with the **Supervisor** role are available. The supervisor isn't included in the list of agents.
* 2 conversations are created and assigned to the agent with the **Agent** role.
* 1 conversation remains unassigned.
* The agent escalates **Conversation 2** to the Supervisor.

Result: Conversation 1 and Conversation 3 are assigned to the agent, while Conversation 2 moves to the top of the same queue with the `Escalation` priority and becomes visible to the supervisor.

## More Information

* [Real-Time Queue Notifications](/ai/escalate/handover-features/real-time-queue-notifications)
* [Conversation Routing: Automatic mode](/live-agent/conversation/conversation-routing/automatic-mode)
* [Conversation Routing Overview](/live-agent/conversation/conversation-routing/overview)


# Automatic Mode
Source: https://docs.cognigy.com/live-agent/conversation/conversation-routing/automatic-mode

Set up automatic conversation assignment and reassignment in Live Agent to evenly distribute incoming chats among human agents.

<a href="/release-notes/2025.16"><Badge>Updated in 2025.16</Badge></a>

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

*Automatic mode* includes two key features: Auto Assignment and Allow conversations to be reassigned, that allocate conversations to agents. Automatic assignment distributes conversations among online agents, while automatic reassignment reassigns conversations in real time based on agent availability changes. These features optimize conversation handling and improve customer support efficiency.

In automatic mode, conversations can be automatically assigned or reassigned to users with specific Live Agent roles within a specific inbox: `Agent`, `Supervisor`, or `Administrator`.

There are different scenarios for distributing conversations to human agents based on their online status and workload:

* **Online Agents**. When a conversation is created in the Inbox, and there are agents online with access to it, the system distributes the conversations evenly among them.
  Unassigned conversations are automatically assigned to agents when the [limit threshold](/live-agent/settings/account-settings#conversation-limit-per-agent) is lower than the maximum value, and they take precedence over newly incoming conversations.
* **Offline Agents**. When a conversation is created in the Inbox, and there are no agents online with access to it, the system will leave all conversations unassigned.

## Restrictions

* You can't trigger automatic assignment and automatic reassignment manually. For example, if Agent 1 manually removes themselves from a conversation and Agent 2 has available capacity, the system won't reassign the conversation to Agent 2. Instead, reassignment will only occur when a triggering event, such as a logout/login action, takes place.

## Automatic Assignment

Live Agent provides an auto-assignment system for conversations based on *round-robin* logic by default. Using this logic, the system automatically distributes incoming conversations in a sequential manner among your human agents.

When set up, the action lists users and assigns conversations one by one. For example, if your Customer Support team consists of three users, the first conversation goes to Alex, the second to David, and the third to Sara. When the fourth conversation arrives, the cycle starts again, and it will be assigned to Alex.

<Frame>
  <img alt="Round-robin assignment cycling conversations among three agents" />
</Frame>

You can configure the auto-assignment system by using the **Auto Assignment** setting. This setting is activated by default for all Inboxes. You can change this behavior for a specific Inbox.

To configure auto assignment in your Inbox, follow these steps:

1. Go to **Settings > Inboxes**.
2. Select an Inbox from the list and click <img alt="Inbox settings" />.
3. On the **Collaborators** tab, go to the **Conversation Assignment** section.
4. From the **Allow auto assignment** list, select one of the following options:
   * Enable — allows the automatic assignment of agents to conversations. The setting is activated by default.
   * Disable — allows the manual assignment of agents to conversations.

The changes will be applied.

### Additional Automatic Assignment Parameters

By default, [automatic assignment](#automatic-assignment) doesn't take into account additional parameters,
such as the number of conversations being handled by each agent, agent skills, or conversation priority.
In this case,
new conversations may be assigned to any available agent without considering their current workload, skills, or language.

You can control this behavior by using the following settings:

| Setting                         | Description                                                                                                                                                                                                                                                                                                                                                                                                  | Example                                                                                                                                                                                                                                                                                                                               |
| ------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Consider conversation count     | When this setting is enabled, Live Agent takes into account the number of conversations that each agent is already handling before assigning new conversations to them. Using this setting, the system automatically distributes incoming conversations in a balanced manner among your human agents. It prevents your agents from becoming overwhelmed and ensures your customers receive prompt responses. | If Alex is currently handling three conversations, and Sara is handling one conversation, the system will assign new incoming conversations to Sara, as she has a lower workload. More examples of how this setting works can be found in [Scenario 6](#common-scenarios).                                                            |
| Consider conversation skills    | When this setting is enabled, Live Agent considers an agent's proficiency in specific skills when assigning new conversations to ensure a more balanced workload distribution among agents with the relevant expertise                                                                                                                                                                                       | If a conversation requires expertise in the Billing skill, and if Alex possesses expertise in Technical Support while Sara specializes in Billing, the system will assign new conversations requiring the Billing skill to Sara. More examples of how this setting works can be found in [Scenario 6](#common-scenarios).             |
| Consider conversation languages | When this setting is enabled, Live Agent takes into account the languages an agent is proficient in when assigning new conversations. This ensures that customer inquiries in different languages are directed to agents who can effectively communicate in those languages.                                                                                                                                 | If a customer inquiry is in Spanish, and Alex is not fluent in Spanish while Sara is, the system will assign that conversation to Sara. More examples of how this setting works can be found in [Scenario 7](#common-scenarios).                                                                                                      |
| Consider conversation priority  | When this setting is enabled, Live Agent prioritizes conversations based on their urgency. The priority option is primarily applicable to reassigned or queued conversations, not to new incoming ones.                                                                                                                                                                                                      | If there are both High priority and Urgent priority conversations in the Conversation Queue, the system will assign the Urgent conversation first to Alex. If Sara is online, the second High priority conversation will be assigned to her. More examples of how this setting works can be found in [Scenario 8](#common-scenarios). |

When agents have an equal number of active conversations, they'll be assigned the next conversation in order using standard round-robin logic. If online agents lack the necessary skill, the conversation will be directed to the nearest available agent through the standard round-robin logic.

To consider additional automatic assignment parameters within your Inbox, follow these steps:

1. Go to **Settings > Inboxes**.
2. Select an Inbox from the list and click <img alt="Inbox settings" />.
3. On the **Collaborators** tab, go to the **Conversation Assignment** section.
4. Activate the following options:
   * Consider conversation count — consider agent workload.
   * Consider conversation skills — consider agent skills.
   * Consider conversation languages — consider agent languages.
   * Consider conversation priority — consider conversation urgency.

The changes will be applied.

## Automatic Reassignment

Live Agent provides an auto-reassignment system for conversations. This system follows a round-robin approach, similar to [automatic assignment](#automatic-assignment).

The auto-reassignment system operates in real-time, triggered by specific agent-related events. These events include:

* When a human agent changes their availability to `Offline` or `Away`.
* When a human agent logs out.

When one of these events is triggered, the conversations assigned to the human agent are automatically reassigned to other available human agents. If no human agents are currently online, the conversations are marked as unassigned.

You can configure the auto-reassignment system by using the **Allow conversations to be reassigned** setting. This setting is activated by default for all Inboxes. You can change this behavior for a specific Inbox.

To configure auto reassignment in your Inbox, follow these steps:

1. Go to **Settings > Inboxes**.
2. Select an Inbox from the list and click <img alt="Inbox settings" />.
3. On the **Collaborators** tab, go to the **Conversation Assignment** section.
4. From the **Allow conversations to be reassigned** list, select one of the following options:
   * Enable — allows the automatic reassignment of agents to conversations. The setting is activated by default.
   * Disable — allows the manual reassignment of agents to conversations.

The changes will be applied.

## Automatic Assignment and Reassignment for Busy Agents

By default, human agents with `Busy` status can't be assigned or reassigned automatically.

If an agent is already assigned conversations before they become busy,
those assigned conversations will remain in the [Conversation Queue](/live-agent/conversation/conversation-queue/overview).
However, new conversations will not be assigned to that agent until they're no longer in the `Busy` status.

You can modify this behavior by enabling the [Auto Assign Conversations to a Busy Agent](/live-agent/settings/account-settings#auto-assign-conversations-to-busy-agents) option in **Account Settings**. Note that this setting is activated at the Live Agent installation level, and you can't modify the behavior for a specific Inbox.

When the setting is activated, busy agents will be treated as having an `Online` status and following the same assignment and reassignment rules as other available agents.

## Automatic Assignment and Reassignment of Snoozed Conversations

### Conversation is Snoozed

<AccordionGroup>
  <Accordion title="Behavior for Until Next Reply">
    When a conversation is snoozed `until next reply`, the system evaluates the following conditions:

    ```text theme={null}
     Is the agent at their conversation limit?
     ├─ Yes → Conversation is unassigned to free up the agent.
     └─ No
       ├─ Agent is online or busy → Keep the conversation assigned
       └─ Agent is away or offline → Unassign
    ```

    The agent limitation check has the highest priority.
  </Accordion>

  <Accordion title="Behavior for Other Snooze Options">
    Snoozed conversations set to `until next day` and `until next week` automatically trigger reassignment. However, if the human agent who snoozed the conversation goes `offline` or becomes `away`, the system may unassign the conversation based on availability logic.

    **Example:**

    Agent Sara snoozes a conversation until the next day, expecting to pick it up during her next shift. However, she logs off shortly afterward. When she returns, the conversation is no longer assigned to her because the system unassigned it when she went offline.
    This behavior follows standard availability logic and applies even to snoozed conversations.
  </Accordion>
</AccordionGroup>

### Reopen Snoozed Conversations

<AccordionGroup>
  <Accordion title="Behavior for `Until Next Reply`">
    ```text theme={null}
    ├─ Still assigned? → Status is updated to Open.
    └─ Unassigned? → System searches for an available agent and reassigns.
    ```
  </Accordion>

  <Accordion title="Behavior for Other `Snooze` Options">
    * When reopened:
      * Status is updated to `Open`.
      * Reassignment will be attempted. If no available human agent is found, the conversation remains unassigned.
  </Accordion>
</AccordionGroup>

## Common Scenarios

The routing process depends on incoming data and the combination of settings you have configured at the Account Settings level and Inbox level.

There are various common scenarios or situations relevant to the automatic mode.

<AccordionGroup>
  <Accordion title="Default Scenario">
    This is the default scenario where all conversations are assigned to the first available agent.

    Source table:

    ```txt theme={null}
    | Agent name | Status  |
    |------------|---------|
    | Agent 1    | Online  |
    | Agent 2    | Online  |
    | Agent 3    | Online  |
    | Agent 4    | Offline |
    ```

    Conditions:

    * No [conversation limit](/live-agent/settings/account-settings#conversation-limit-per-agent) per agent.
    * 4 conversations are created.
    * Activated:
      * Automatic Conversation Assignment.
      * Automatic Conversation Reassignment.
    * Deactivated:
      * Auto Assign Conversations to a Busy Agent.
      * Consider Conversation Count.

    Result: Agent 1 receives the first and fourth conversations, Agent 2 receives the second, Agent 3 gets the third.
  </Accordion>

  <Accordion title="Scenario 1">
    Source table:

    ```txt theme={null}
    | Agent name | Limit | Status  |
    |------------|-------|---------|
    | Agent 1    | 2     | Online  |
    | Agent 2    | 2     | Busy    |
    | Agent 3    | 2     | Away    |
    | Agent 4    | 2     | Offline |
    ```

    Common conditions:

    * The limit is 2 conversations per agent.
    * 4 conversations are created.

    **Example 1**

    Additional conditions:

    * Auto Assign Conversations to a Busy Agent is deactivated.
    * Automatic Conversation Assignment is activated.
    * Automatic Conversation Reassignment is activated.

    Result: The first two will be assigned to Agent 1, and the next two remain unassigned.

    **Example 2**

    Additional conditions:

    * Automatic Conversation Assignment is activated.
    * Automatic Conversation Reassignment is activated.
    * Auto Assign Conversations to a Busy Agent is activated.

    Result: Agent 1 receives the first conversation, Agent 2 receives the second, Agent 1 gets the third, and Agent 2 gets the fourth.

    **Example 3**

    Additional conditions:

    * Automatic Conversation Assignment is activated.
    * Automatic Conversation Reassignment is deactivated.
    * Auto Assign Conversations to a Busy Agent is deactivated.
    * Agent 1 goes offline.

    Result: The first two conversations will remain assigned to Agent 1, and the next two remain unassigned.

    **Example 4**

    Additional conditions:

    * Automatic Conversation Assignment is deactivated.
    * Automatic Conversation Reassignment is activated.
    * Auto Assign Conversations to a Busy Agent is activated.

    Result: All conversations remain unassigned.
  </Accordion>

  <Accordion title="Scenario 2">
    Source table:

    ```txt theme={null}
    | Agent name | Status |
    |------------|--------|
    | Agent 1    | Online |
    | Agent 2    | Online |
    | Agent 3    | Online |
    | Agent 4    | Online |
    ```

    Common conditions:

    * Activated:
      * Automatic Conversation Reassignment.
      * Automatic Conversation Assignment.
    * Deactivated:
      * Consider Conversation Count.
    * Auto Assign Conversations to a Busy Agent is activated or deactivated.

    **Example 1**

    Additional conditions:

    * The limit is 2 conversations per agent.
    * 4 conversations are created.

    Result: Each agent receives one conversation.

    **Example 2**

    * The limit is 2 conversations per agent.
    * 8 conversations are created.
    * Agent 1 has changed the status to `Away` after the conversation assignment.

    Result: Each agent has two ongoing conversations. However, when Agent 1 goes offline or is marked as away, their two conversations will remain unassigned.
  </Accordion>

  <Accordion title="Scenario 3">
    Source table:

    ```txt theme={null}
    | Agent name | Limit | Status  |
    |------------|-------|---------|
    | Agent 1    | 2     | Offline |
    | Agent 2    | 2     | Away    |
    | Agent 3    | 2     | Offline |
    | Agent 4    | 2     | Busy    |
    ```

    Common conditions:

    * The limit is 2 conversations per agent.
    * 4 conversations are created.
    * Activated:
      * Automatic Conversation Reassignment.
      * Automatic Conversation Assignment.
    * Deactivated:
      * Consider Conversation Count.

    **Example 1**

    Additional conditions:

    * Auto Assign Conversations to a Busy Agent is deactivated.

    Result: all conversations will be unassigned.

    **Example 2**

    Additional conditions:

    * Auto Assign Conversations to a Busy Agent is deactivated.
    * Agent 1 has changed the status to `Online`.

    Result: When Agent 1 returns online, two conversations will be assigned to an agent, and two conversations will remain unassigned.

    **Example 3**

    Additional conditions:

    * Auto Assign Conversations to a Busy Agent is activated.

    Result: The first two conversations will be assigned to Agent 4, and the next two remain unassigned.
  </Accordion>

  <Accordion title="Scenario 4">
    Source table:

    ```txt theme={null}
    | Agent name | Limit | Status  |
    |------------|-------|---------|
    | Agent 1    | 2     | Offline |
    | Agent 2    | 2     | Offline |
    | Agent 3    | 2     | Offline |
    | Agent 4    | 2     | Offline |
    ```

    Common conditions:

    * The limit is 2 conversations per agent.
    * Activated:
      * Automatic Conversation Reassignment.
      * Automatic Conversation Assignment.
    * Deactivated:
      * Consider Conversation Count.

    **Example 1**

    Additional conditions:

    * 5 conversations are created.
    * Auto Assign Conversations to a Busy Agent is activated or deactivated.

    Result: All conversations are unassigned.

    **Example 2**

    Additional conditions:

    * Auto Assign Conversations to a Busy Agent is deactivated.
    * Agent 1 has changed the status to `Busy`.

    Result: All conversations are unassigned.

    **Example 3**

    Additional conditions:

    * Auto Assign Conversations to a Busy Agent is activated.
    * Agent 1 has changed the status to `Busy`.

    Result: When Agent 1 has changed the status from `Offline` to `Busy`, two conversations are assigned to this human agent.
  </Accordion>

  <Accordion title="Scenario 5">
    Source table:

    ```txt theme={null}
    | Agent name | Limit | Status  | Conversations |
    |------------|-------|---------|---------------|
    | Agent 1    | 5     | Online  | 2             |
    | Agent 2    | 5     | Online  | 3             |
    | Agent 3    | 5     | Online  | 1             |
    | Agent 4    | 5     | Online  | 4             |
    ```

    Common conditions:

    * The limit is 5 conversations per agent.
    * Each agent has assigned conversations.
    * Activated:
      * Automatic Conversation Assignment.
      * Automatic Conversation Reassignment.
      * Consider Conversation Count.

    **Example 1**

    Additional conditions:

    * 3 new incoming conversations.
    * Deactivated: Auto Assign Conversations to a Busy Agent.

    Result: The first conversation is assigned to Agent 3, while the second and third conversations are randomly assigned to either Agent 1 or Agent 3.

    **Example 2**

    Additional conditions:

    * 4 new incoming conversations.
    * Auto Assign Conversations to a Busy Agent is deactivated.
    * Agent 3 has changed the status to `Offline`, `Away`, or `Busy`.

    Result: The first conversation will be assigned to Agent 1. The second conversation will be assigned to either Agent 1 or Agent 2. The third conversation will be assigned to the Agent who is left with three conversations. The fourth conversation will be randomly assigned to one of the three agents.
  </Accordion>

  <Accordion title="Scenario 6">
    Source table:

    ```txt theme={null}
    | Agent name | Limit | Status | Conversations | Skills            |
    |------------|-------|--------|---------------|-------------------|
    | Agent 1    | 5     | Online | 1             | Technical support |
    | Agent 2    | 5     | Online | 1             | Billing           |
    | Agent 3    | 5     | Online | 1             | Customer Support  |
    | Agent 4    | 5     | Online | 1             | Technical support |
    ```

    Common conditions:

    * The limit is 5 conversations per agent.
    * Each agent is assigned per conversation.
    * Activated:
      * Automatic Conversation Assignment.
      * Automatic Conversation Reassignment.
      * Consider Conversation Count.
      * Consider Conversation Skills.

    **Example 1**

    Additional conditions:

    * 1 new incoming conversation with the `Technical support` skill.
    * 1 new incoming conversation with the `Technical support`, `Billing` skills.
    * Deactivated: Auto Assign Conversations to a Busy Agent.

    Result: The first conversation is assigned to either Agent 1 or Agent 4, both of whom possess the required skills. The second conversation is assigned to either Agent 2 or the agent left with one conversation from the previous assignment.

    **Example 2**

    Additional conditions:

    * 1 new incoming conversation with the `Technical support` skill.
    * 1 new incoming conversation with the `Technical support`, `Billing` skills.
    * Agent 2 is going to be offline.
    * Deactivated: Auto Assign Conversations to a Busy Agent.

    Result: The first incoming conversation will be assigned to Agent 1,
    the second to Agent 4.

    **Example 3**

    Additional conditions:

    * 1 new incoming conversation with the `Technical support` skill.
    * 1 new incoming conversation with the `Technical support`, `Billing` skills.
    * Agent 1 and Agent 2 are going to be offline.
    * Deactivated: Auto Assign Conversations to a Busy Agent.

    Result: The first and the second incoming conversations will be assigned to Agent 4.
  </Accordion>

  <Accordion title="Scenario 7">
    Source table:

    ```txt theme={null}
    | Agent name | Limit | Status | Conversations | Skills            | Languages        |
    |------------|-------|--------|---------------|-------------------|------------------|
    | Agent 1    | 5     | Online | 1             | Technical support | English          |
    | Agent 2    | 5     | Online | 1             | Billing           | Spanish          |
    | Agent 3    | 5     | Online | 1             | Customer Support  | English, Spanish |
    | Agent 4    | 5     | Online | 1             | Technical support | Greek            |
    ```

    Common conditions:

    * The limit is 5 conversations per agent.
    * Each agent is assigned to one conversation.
    * Activated:
      * Automatic Conversation Assignment.
      * Automatic Conversation Reassignment.
      * Consider Conversation Count.
      * Consider Conversation Skills.
      * Consider Conversation Languages.

    **Example 1**

    Additional conditions:

    * 1 new incoming conversation with the `Technical support`, `English` skills.
    * 1 new incoming conversation with the `Technical support`, `Billing`, `Spanish` skills.
    * Deactivated: Auto Assign Conversations to a Busy Agent.

    Result: The first incoming conversation will be assigned to Agent 1, the second conversation will be assigned randomly except Agent 1.

    **Example 2**

    Additional conditions:

    * 3 new incoming conversations with the `Technical support`, `English` skills.
    * Deactivated: Auto Assign Conversations to a Busy Agent.

    Result: 3 incoming conversations will be assigned to Agent 1.
  </Accordion>

  <Accordion title="Scenario 8">
    Source table:

    ```txt theme={null}
    | Agent name | Limit | Status | Conversations | Skills            | Languages        |
    |------------|-------|--------|---------------|-------------------|------------------|
    | Agent 1    | 5     | Online | 1             | Technical support | English          |
    | Agent 2    | 5     | Online | 1             | Billing           | Spanish          |
    | Agent 3    | 5     | Online | 1             | Customer Support  | English, Spanish |
    | Agent 4    | 5     | Online | 1             | Technical support | Greek            |
    ```

    Common conditions:

    * The limit is 5 conversations per agent.
    * Each agent is assigned per conversation.
    * Activated:
      * Automatic Conversation Assignment.
      * Automatic Conversation Reassignment.
      * Consider Conversation Count.
      * Consider Conversation Skills.
      * Consider Conversation Languages.
      * Consider Conversation Priority.

    **Example 1**

    Additional conditions:

    * 1 new incoming conversation with the `Technical support`, `English` skills, and **Low** priority.
    * 1 new incoming conversation with the `Technical support`, `Billing`, `Spanish` skills, and **Urgent** priority.
    * Deactivated: Auto Assign Conversations to a Busy Agent.

    Result: The first incoming conversation will be assigned to Agent 1, while the second one will be randomly assigned to any available agent, except Agent 1.

    **Example 2**

    Additional conditions:

    * 1 new incoming conversation with the **Low** priority.
    * 1 new incoming conversation with the `Technical support`, `Billing`, `Spanish` skills, and **Urgent** priority.
    * Deactivated: Auto Assign Conversations to a Busy Agent.

    Result: The first conversation will be assigned randomly,
    and the second one will also be assigned randomly between the remaining agents with skill and language matches.
  </Accordion>

  <Accordion title="Scenario 9">
    Source table:

    ```txt theme={null}
    | Agent name | Limit | Status | Conversations | Skills                              | Languages        |
    |------------|-------|--------|---------------|-------------------------------------|------------------|
    | Agent 1    | 5     | Online | 5             | Technical support                   | English          |
    | Agent 2    | 5     | Online | 4             | Billing                             | Spanish          |
    | Agent 3    | 5     | Online | 4             | Customer Support, Technical support | English, Spanish |
    ```

    Basic conditions:

    * The limit is 5 conversations per agent.
    * Each conversation is assigned to a single agent.
    * Activated:
      * Automatic Conversation Assignment.
      * Automatic Conversation Reassignment.
      * Consider Conversation Skills.
      * Consider Conversation Languages.
    * Deactivated: Auto Assign Conversations to a Busy Agent.

    Additional conditions:

    * Agent 1 snoozes the conversation until the next reply.
    * Agent 1 has changed their status to `Busy`.

    Result: the end user sends a reply, and the conversation is assigned to Agent 3, who shares the same skill and language as Agent 1.
  </Accordion>
</AccordionGroup>

## More Information

* [Manual Conversation Mode](/live-agent/conversation/conversation-routing/manual-mode)
* [Conversation Routing](/live-agent/conversation/conversation-routing/overview)
* [Inboxes](/live-agent/settings/inboxes)
* [Account Settings](/live-agent/settings/account-settings)


# Manual Mode
Source: https://docs.cognigy.com/live-agent/conversation/conversation-routing/manual-mode

Configure manual routing in Live Agent to assign or escalate conversations to specific agents, supervisors, or administrators based on case needs.

<a href="/release-notes/2025.25"><Badge>Updated in 2025.25</Badge></a>

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

*Manual mode* can be helpful in scenarios requiring specialized expertise, urgent or high-priority cases, complex or sensitive issues, and/or key customers. It enables organizations to provide personalized support, optimize customer experience, and efficiently handle conversations.

In manual mode, conversation can be manually assigned or reassigned to any users with the Live Agent [roles](/live-agent/roles): `Agent`, `Supervisor`, or `Administrator`.

## Manual Assignment

You can manually assign a conversation to an agent or team by following these steps:

1. In the left-side menu of Live Agent, click **Conversations**.
2. Go to the relevant conversation.
3. In the **Details** menu, go to the **Conversation Actions** section. Choose how you want to manage the conversation:

   <Tabs>
     <Tab title="Assign to an Agent">
       Assign the conversation to a human agent by selecting an agent from the **Assigned Agent** list. If you want to assign the conversation to yourself, click **→ Assign to me** above the list.
     </Tab>

     <Tab title="Escalate to a Supervisor">
       Escalate the conversation to a human agent with the `Supervisor` role by clicking the **↑ Escalate to Supervisor** button, which appears in the **Conversation Actions** section. The conversation will be automatically assigned to the available supervisor in the [Inbox](/live-agent/settings/inboxes#escalate-conversations-to-supervisors).

       <Accordion title="Supervisor Assignment Rules">
         A conversation remains unassigned if:

         * No supervisor is available in the Inbox.
         * An assigned supervisor changed their status to `Away` or `Offline`.
         * A supervisor is busy and [Assign conversation to a busy agent](/live-agent/settings/account-settings#auto-assign-conversations-to-busy-agents) is turned off.

         A conversation remains assigned if the supervisor was already assigned to it and then changed their status to `Busy`.
       </Accordion>

       <Note>
         Each conversation can be escalated via **↑ Escalate to Supervisor** button only once. Reassigning back to an agent disables the **Escalate to Supervisor** button. Use the **Assigned Agent** list to select a supervisor manually if needed.
       </Note>
     </Tab>
   </Tabs>
4. *(Optional)* From the **Assigned Team** list, select a team that can be assigned to this conversation. The previously selected agent must be a part of this team.

After the assignment, the agent or a team can receive a [notification](/live-agent/notifications).

## Manual Reassignment

You can manually reassign a conversation to an agent or team by following these steps:

1. In the left-side menu of Live Agent, click **Conversations** and go to the relevant conversation.
2. In the **Details** menu, go to the **Conversation Actions** section.
3. From the **Assigned Agent** list, select a new agent that should be assigned to this conversation.
4. *(Optional)* From the **Assigned Team** list, select a new team to be assigned to this conversation. The previously selected agent must be a part of this team.

If the conversation is reassigned, an agent can receive a [notification](/live-agent/notifications).

## Manual Unassignment

You can manually unassign a conversation to an agent or team by following these steps:

1. In the left-side menu of Live Agent, click **Conversations** and go to the relevant conversation.
2. In the **Details** menu, go to the **Conversation Actions** section.
3. Above the **Assigned Agent** list, click **→ Unassign**.
4. Above the list **Assigned Team**, click **→ Unassign**.

If the **Unassign** value appears in the **Assigned Agent** list, the conversation will be displayed in the list of unassigned conversations.

## More Information

* [Automatic Conversation Mode](/live-agent/conversation/conversation-routing/automatic-mode)
* [Conversation Routing](/live-agent/conversation/conversation-routing/overview)


# Conversation Routing
Source: https://docs.cognigy.com/live-agent/conversation/conversation-routing/overview

Learn how manual and automatic routing distributes conversations to the right human agents and teams.

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

*Conversation routing* assigns and distributes incoming conversations to the appropriate human agents. It involves intelligent routing algorithms that consider factors like agent availability and workload to determine a recipient for each conversation. In Cognigy Live Agent, conversation routing ensures that conversations are efficiently directed to the right agents or teams, enabling timely and effective customer support. This feature optimizes resource utilization and enhances the productivity of human agents.

In Live Agent, there are two routing modes:

* [Manual](/live-agent/conversation/conversation-routing/manual-mode)
* [Automatic](/live-agent/conversation/conversation-routing/automatic-mode)

Conversation routing can consider a [limit per agent](/live-agent/settings/account-settings#conversation-limit-per-agent) applied to both Manual and Automatic modes and can only be set up for all Inboxes.


# Other Operations
Source: https://docs.cognigy.com/live-agent/conversation/other-operations

Learn about other operations in Live Agent, such as muting, sending transcripts by email, and enabling agent masking for efficient support.

<a href="/release-notes/4.72"><Badge>Updated in 4.72</Badge></a>

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

In Live Agent, you can mute or unmute, send transcripts via email, clear chat history, and enable agent masking. These actions make managing conversations more efficient, organized, and user-friendly.

## Mute Conversations

To mute a conversation, do the following:

1. In the left-side menu of the Live Agent interface, select **Conversations**.
2. In the **Inboxes** column, select an Inbox.
3. Select a conversation.
4. In the upper-right corner of the Conversation page, click <img alt="mute conversation" />.

No audio messages can then be sent or received in this chat.

## Unmute Conversations

To unmute a conversation, do the following:

1. In the left-side menu of the Live Agent interface, select **Conversations**.
2. In the **Inboxes** column, select an Inbox.
3. Select a conversation.
4. In the upper-right corner of the Conversation page, click <img alt="unmute conversation" />.

Audio messages can then be sent or received again in this chat.

## Send Conversation Transcripts by Email

To send a copy of the current conversation by email, do the following:

1. In the left-side menu of the Live Agent interface, select **Conversations**.
2. In the **Inboxes** column, select an Inbox.
3. Select a conversation.
4. In the upper-right corner of the Conversation page, click <img alt="send transcript" />. A screen opens up which allows sending a conversation transcript as email to human agents in a team. The chat transcript can also be sent to the customer or another selected email address.
5. Select your desired option and click **Submit**.

A confirmation message appears: `The chat transcript was sent successfully`. The chosen recipient receives the chat transcript in their email inbox.

## Clear Conversation History

To clear a chat history, the `Administrator` or `Supervisor` role is required.

To send a copy of the current conversation by email, do the following:

1. In the left-side menu of the Live Agent interface, select **Conversations**.
2. In the **Inboxes** column, select an Inbox.
3. Select a conversation.
4. In the upper-right corner of the Conversation page, click <img alt="clear conversation history" />.

Live Agent deletes the selected conversation history.

## Enable Agent Masking in Conversations

If you don't want the end user to see the name and image of the human agent in the chat, ask your administrator to enable the **Agent masking** option in the Account Settings.

Activating the **Enable Agent masking** setting masks all agent's personal information, such as name, email, and profile images. For example, the agent name **Max Mustermann** would turn into **Agent x4gH5tZ**. Only the agent himself can see the clear name.
Only the agent can see the original name, so there's no way for anyone else in your organization or the end user to link the masked name to a real human agent.

For more information, read the [Account Settings](/live-agent/settings/account-settings) documentation.

## Expand or Collapse the Chat Area

To change the size of the conversation chat area, select one of the following options:

* To expand the chat area, click <img alt="expand" /> in the upper-left corner of the chat screen.
* To collapse the chat area, click <img alt="collapse" /> in the upper-left corner of the chat screen.

This feature can be particularly useful if you want to use the [Agent Copilot](/live-agent/assistants/agent-copilot) workspace within your conversation.


# Conversations
Source: https://docs.cognigy.com/live-agent/conversation/overview

Learn about Live Agent conversations, including life cycle, statuses, and handover from AI Agents.

<a href="/release-notes/4.84"><Badge>Updated in 4.84</Badge></a>

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

*Conversations* are chats between a human agent and a customer, initiated by the customer via a messaging application, such as Cognigy Webchat, WhatsApp, Slack, and so on.
The conversation is handed over to a human agent when the AI Agent cannot assist with a question due to its complexity or lack of information. During the conversation, the human agent may use additional virtual assistants to help answer the question.

<Frame>
  <img alt="Live Agent conversation panel showing the main chat interface between human agents and end users" />
</Frame>

## Conversation Life Cycle

The conversation life cycle involves starting a conversation, keeping it open for active engagement, possibly putting it on hold temporarily, and finally closing or resolving it.

<Frame>
  <img alt="Live Agent Conversation Statuses" />
</Frame>

Each conversation has statuses that are listed in the table.

| Status    | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   |
| --------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Opened    | This status indicates that a conversation is handed over to a human agent. The conversation can be opened by a user or reopened by a human agent. From this status, the conversation can transition in the following ways:<br />- Move to `Pending` and `Snoozed` statuses. Can be returned from these inactive statuses to the `Open` status.<br />- Move to `Resolved` or `Abandoned` statuses. If the conversation is marked as resolved or abandoned, it cannot be reopened.<br />                                                                                                                                                                                                                                                        |
| Pending   | This status indicates that a conversation is temporarily paused, awaiting a response from either the customer or a human agent.<br /> To mark a conversation as pending, in the upper-right corner of the Conversation page, next to the **Resolve** button, click <img alt="select" /> **> Mark as pending**. To reopen the conversation, click **Open**.                                                                                                                                                                                                                                                                                                                                                                                    |
| Snoozed   | This status indicates that a conversation is temporarily suspended until a time trigger or customer response trigger was activated.<br />To mark conversation as snoozed, in the upper-right corner of the Conversation page, next to the **Resolve** button, click <img alt="select" />. Select one of the following trigger options:<br />- **Snooze Until Next Reply** — the issue will be set aside for now, and the status will change to `Open` upon the next customer reply.<br />- **Snooze Until Tomorrow** — the issue will be set aside for now, and the status will change to `Open` the next day.<br />- **Snooze Until Next Week** — the issue will be set aside for now, and the status will change to `Open` next week.<br /> |
| Abandoned | The status indicates that a conversation is closed because the user left the conversation.<br />To mark a conversation as abandoned, in the top right corner of the Conversation page, click **Resolve** and select <img alt="select" /> **> Abandon and resolve**. Once a conversation is marked as abandoned, its state cannot be changed back to open or pending.<br /> Agents with the `Administrator` role can activate the following options in the [Account Settings](/live-agent/settings/account-settings):<br />- Automatically abandon conversations.<br />- Automatically delete abandoned conversations.                                                                                                                         |
| Resolved  | This status indicates that a conversation is closed because the user's problem has been fixed.<br /> To mark a conversation as resolved, in the upper-right corner of the Conversation page, click **Resolve**. Once a conversation is marked as resolved, its state cannot be changed back to open or pending.<br />Agents with the `Administrator` role can activate the following settings in the [Account Settings](/live-agent/settings/account-settings).<br />- Automatically resolve conversations.<br />- Automatically delete conversations that were resolved.<br />                                                                                                                                                               |

Conversations can be [selected and filtered by their statuses](/live-agent/conversation/search-and-filter) within the Conversation page.

## Conversation Assignment Statuses

A conversation with the `Opened` status in Live Agent can have one of the following assignment statuses:

* **Unassigned** — a conversation is not assigned to a human agent.
* **Assigned** — a conversation is assigned to a human agent. The assignment occurs either [manually](/live-agent/conversation/conversation-routing/manual-mode) or [automatically](/live-agent/conversation/conversation-routing/automatic-mode).

## Conversation Details

When you select a conversation, conversation details appear on the right side of the Live Agent interface.
These details include the following conversation information:

* [Contact Profile Details](#contact-profile-details)
* [Conversation Actions](#conversation-actions)
* [Conversation Information](#conversation-information)

<Frame>
  <img alt="Live Agent Conversation Details" />
</Frame>

### Contact Profile Details

The Contact Profile information is shared at the beginning of the handover process. Live Agent does not store private data to comply with GDPR regulations. Instead, Live Agent only retains a reference to Cognigy. When a human agent enters the conversation, Live Agent requests data from Cognigy.

If the human agent doesn't leave the conversation page, they may miss contact profile updates. To view profile updates that occur on the Cognigy side, the human agent must reenter the conversation. This can be done by switching to another page within Live Agent and then returning.

The following table represents data that can be edited within Live Agent and stored in the Contact Profile.

| Contact Profile Details | Description                                                                                                                                              |
| ----------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------- |
| First Name              | User's first name.                                                                                                                                       |
| Last name               | User's last name.                                                                                                                                        |
| Email                   | User's email address.                                                                                                                                    |
| Gender                  | User's gender.                                                                                                                                           |
| Age                     | User's age.                                                                                                                                              |
| Birthday                | User's birthday.                                                                                                                                         |
| Location                | User's location.                                                                                                                                         |
| Accepted GDPR           | Accepted General Data Protection Regulation.                                                                                                             |
| Accepted Privacy Policy | Accepted the privacy policy terms. This field is applicable to Webchat v3 when a user accepts the [Privacy Notice](/webchat/v3/features#privacy-notice). |
| Tasks                   | Tasks are the user's intention or task.                                                                                                                  |
| Profile Picture         | When available, the URL address of a user's picture.                                                                                                     |

#### Change Contact Profile Details

To change the Contact Profile, follow these steps:

1. In the left-side menu of the Live Agent interface, select **Conversations**.
2. Select a conversation from the list.
3. On the right side of the page, navigate to **Conversation Details**.
4. On the **Details** tab, click <img alt="edit" /> to change the Contact Profile.
5. In the **Contact Profile** window, fill in the fields and click **Submit** to save your changes.

This action will also update the [Contact Profile](/ai/agents/analyze/contact-profiles) on the Cognigy side. In Live Agent, cached Contact Profile data will persist for one hour. You can change this parameter if you have an on-premises environment. To do so, add the `COGNIGY_CONTACT_PROFILE_CACHE_TIME` variable in the Live Agent `values.yaml` file and specify the value in seconds.

#### View Hidden Fields in Contact Profile Details

By default, only fields that contain values appear in the Contact Profile.
To view empty fields, follow these steps:

1. In the left-side menu of the Live Agent interface, select **Conversations**.
2. Select a conversation from the list.
3. On the right side of the page, navigate to **Conversation Details**.
4. On the **Details** tab, click <img alt="hide password" /> to view the empty fields. To hide them again, click <img alt="show password" />.

### Conversation Actions

Within a conversation in Live Agent, a human agent can do the following:

* Assign other agents to the selected conversation.
* Assign a group of agents defined in a team.
* Define the priority for handling the customer's request in the conversation.
* Assign one or more labels to make the overview of the listed conversations easier for the agent at a glance.

For more information on these details, read the [Assign Conversations](/live-agent/conversation/assign-conversations) documentation.

### Conversation Information

Conversation information includes the following details:

* **Cognigy Session ID** — a unique identifier associated with a session in the Cognigy platform. In the context of Cognigy and the Live Agent interface, a session refers to the collection of user inputs and agent replies within a conversation.
* **Channel Type** — a messaging platform through which users can interact with the AI Agent. Each channel is associated with one or more Cognigy Endpoints, allowing users to communicate with the AI Agent through various interfaces. Cognigy channels include platforms like Webchat, Facebook Messenger, Slack, WhatsApp, and other messaging or communication platforms.
* **Cognigy User ID** — a unique identifier for authorized access to Cognigy and Live Agent.

To copy this information, click <img alt="copy" /> next to the item.

### Previous Conversations

After a conversation has been marked as resolved, the next time the same customer starts a new conversation, the previous conversation associated with this customer will be shown in the Previous Conversations menu.
This will help the agent understand the history of the customer's issue.

## Assistants

Cognigy provides three assistants for giving human agents support in their service conversations. You can choose the installation that best suits your requirements:

* [Agent Copilot Bot](/live-agent/assistants/agent-copilot-bot)
* [Agent Copilot Whisper](/live-agent/assistants/agent-copilot-whisper)
* [Agent Copilot Workspace](/live-agent/assistants/agent-copilot)

## View Conversations

To open a conversation in Live Agent, follow these steps:

1. In the left-side menu of the Live Agent interface, click **Conversations**.
2. In the upper-left corner, select **All Conversations**.
3. On the Conversations panel, choose one of the following tabs:
   * **Mine** — displays conversations assigned to you.
   * **Unassigned** — shows conversations without a designated owner.
   * **All** — shows all conversations in the system. Agents, by default, cannot view all conversations, but administrators can grant access through the **Inboxes** settings. For more information, read the [Configure Agent Access to all Inbox Conversations](/live-agent/settings/inboxes#configure-agent-access-to-all-inbox-conversations).
4. Select a specific conversation from the chosen tab. The conversation will open, providing additional details and actions in the right-side panel.

## More Information

* [Assign Conversations](/live-agent/conversation/assign-conversations)
* [Send Replies](/live-agent/conversation/send-reply)
* [Filter and Sort](/live-agent/conversation/search-and-filter)
* [Other Operations](/live-agent/conversation/other-operations)


# Search, Filter, and Sort Conversations
Source: https://docs.cognigy.com/live-agent/conversation/search-and-filter

Search, filter, and sort Live Agent conversations by keywords, status, assignee, or team.

<a href="/release-notes/4.85"><Badge>Updated in 4.85</Badge></a>

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

In Live Agent, human agents can easily search, filter, and organize conversations.
The agents can find messages by entering keywords,
apply filters for status, assignee name, and team name,
and sort conversations by last activity, creation date, or priority.

## Search Conversations

To search for messages in conversations, do the following:

1. Go to Live Agent.
2. On the left side of the screen, select **Conversations**.
3. From the **Inboxes** list, select an Inbox.
4. At the top of the **Conversations** list next to the magnifying glass icon, click the **search for messages in conversations** field.
5. Enter the message keywords in that field, for example, keywords describing a specific customer request. Once you enter a search keyword, the search process starts, and a list of search results will be displayed below the search field.
6. Select the conversation in the search result list you want to inspect. The conversation will be opened and displayed on the right-hand side of the window including the conversation details.

### Search for Conversations where you were Mentioned

You can find all conversations where other agents mentioned you. To do so, follow these steps:

1. Go to Live Agent.
2. On the left side of the screen, select **Conversations**.
3. At the top of the **Inboxes** list, click **@Mentions**.

## Filter Conversations

You can filter conversations using the filters mentioned in the table below. Multiple filters can be combined using the `AND` and `OR` operators.

| Filter Name             | Description                                                                                                         |
| ----------------------- | ------------------------------------------------------------------------------------------------------------------- |
| Status                  | Filters data based on the current conversation [status](/live-agent/conversation/overview#conversation-life-cycle). |
| Assignee Name           | Filters based on the human agent responsible for a conversation.                                                    |
| Inbox Name              | Filters based on conversations associated with a specific Inbox.                                                    |
| Team Name               | Filters data related to a particular team.                                                                          |
| Conversation Identifier | Filters based on a unique identifier for a conversation.                                                            |
| Cognigy Session ID      | Filters based on a unique identifier for a session within the Cognigy platform.                                     |
| Cognigy User ID         | Filters based on a unique identifier for a user within the Cognigy platform.                                        |
| Labels                  | Filters based on tags assigned to conversations for categorization and easy filtering.                              |
| Created At              | Filters based on the timestamp of when a conversation was created.                                                  |
| First Reply Created At  | Filters based on the timestamp of the first reply to a conversation.                                                |
| Last Activity           | Filters based on the timestamp of the last activity on a conversation.                                              |
| Priority                | Filters based on the priority level assigned to a conversation.                                                     |
| Cognigy Channel Type    | Filters based on the type of channel or communication platform within Cognigy.                                      |

### Save Filters

After applying a filter, you can save it for future use.

To save a filter, do the following:

1. Go to Live Agent.
2. On the left side of the screen, select **Conversations**.
3. From the **Inboxes** list, select an Inbox.
4. At the top of the **Conversations** list, click <img alt="filter" />.
5. In the **Filter Conversations** window, select the criteria by which you want to filter. For example, select **Status** equal to **Abandoned** to see only abandoned conversations.
6. Click **Apply Filters**.
7. At the top of the **Conversations** list, click <img alt="save filter" />.
8. In the opened window, enter a name for the filter. For example, `Abandoned Conversations`.
9. Click **Save Filter**. After that, the filter will appear under the **Folders** section.

Saved filters are visible only to the human agent who created them and are not accessible to other agents.

If you want to return to the unfiltered list, click **All Conversations >** <img alt="clear filters" />.

### Filter Conversations by Priority

To filter conversations by priority, do the following:

1. Go to Live Agent.
2. On the left side of the screen, select **Conversations**.
3. From the **Inboxes** list, select an Inbox.
4. At the top of the **Conversations** list, click <img alt="filter" />.
5. In the **Filter Conversations** window, select the following options:<br />
   1. In the **Filter** list, select **Priority**.<br />
   2. In the **Operator** list, select **Equal to** or **Not Equal to** as the filter condition. These operators indicate whether you want to filter conversations that match or do not match the specified priority level.<br />
   3. In the **Select** list, choose a priority level by either clicking it or hovering your cursor over it and then pressing <kbd>Enter</kbd>. You can select multiple priorities from the list.<br />
6. Click **Apply filters** to activate the filter settings.

The filtered conversations, matching to the activated priority filter settings, will be displayed in the list of conversations.

## Sort Conversations

To sort conversations by status and specified characteristics, follow these steps:

1. Go to Live Agent.
2. On the left side of the screen, select **Conversations**.
3. From the **Inboxes** list, select an Inbox.
4. At the top of the **Conversations** list, select one of the following tabs:
   * **Mine** — shows conversations assigned to you.
   * **Unassigned** — shows conversations without a designated owner.
   * **All** — shows all conversations in the system. Agents, by default, cannot view all conversations, but administrators can grant access through the **Inboxes** settings. For more information, read the [Configure Agent Access to all Inbox Conversations](/live-agent/settings/inboxes#configure-agent-access-to-all-inbox-conversations).
5. Above the tabs, click <img alt="sort conversations" />.
6. In the **Sort Conversations** window, select the following options:
   * **Status** — select a conversation [status](/live-agent/conversation/overview#conversation-life-cycle) you want to search for. Available statuses are `Open`, `Pending`, `Resolved`, `Snoozed`, and `Abandoned`.<br />
   * **Order by** — select one of the available options:
     * **Oldest** — sorts conversations by the longest active ones. By default, the **Unassigned** tab shows the oldest at the top of the list.
     * **Last activity** — sorts conversations by the most recent user or human agent interaction, displaying the most recently active conversations at the top of the list.
     * **Created at** — sorts conversations by creation date, starting with the latest. By default, the **Mine** and **All** tabs show the latest conversations at the top of the list.
     * **Priority** — sorts conversations by assigned priority, displaying high-priority conversations at the top of the list.

As soon as you select a sort option, it becomes active, and the listed conversations are immediately updated according to the selected search options.

## More Information

* [Conversations](/live-agent/conversation/overview)
* [Conversation Assignment](/live-agent/conversation/assign-conversations)
* [Conversation Replies](/live-agent/conversation/send-reply)


# Send Replies
Source: https://docs.cognigy.com/live-agent/conversation/send-reply

Send replies with canned responses and variables in Live Agent for quick, efficient customer support messages.

<a href="/release-notes/4.87"><Badge>Updated in 4.87</Badge></a>

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

Within a conversation, as a human agent, you can send a reply to the following actors:

* [end users](#send-replies-to-users)
* [other human agents](#send-replies-to-other-human-agents)

You can also [track typing activity](#track-human-agent-typing-events) via a WebSocket connection to determine when the human agent starts typing.

<Frame>
  <img alt="Live Agent Conversations Reply" />
</Frame>

## Send Replies to Users

To respond to a user in a conversation, type your message in the **Reply** section and click **Send**.

For quick replies, you can use canned responses, variables, or a combination of both.

By default, you can send messages by pressing <kbd>Cmd</kbd>+<kbd>Enter</kbd> (Mac) or <kbd>Ctrl</kbd>+<kbd>Enter</kbd> (Windows/Linux) instead of clicking the **Send** button.
The <kbd>Enter</kbd> key will be used for line breaks. If you want to change hotkey options, go to [Profile Settings](/live-agent/profile-settings).

If you don't send your reply right away but have already composed it, the message will be saved for 24 hours. This means that when you switch to another conversation and return to the previous one, you'll find the initial draft of your reply.

### Reply Options

Human agents can access a range of reply options through the following buttons:

| Button                     | Description                                                                                            |
| -------------------------- | ------------------------------------------------------------------------------------------------------ |
| <img alt="emoji" />        | The **Show emoji selector** button allows a human agent to select and include an Emoji in the message. |
| <img alt="attach files" /> | The **Attach files** button allows a human agent to select and include attachments in the message.     |
| <img alt="request file" /> | The **Request file** button allows a human agent to request a file from the end user.                  |
| <img alt="record audio" /> | The **Record audio** button allows a human agent to send voice messages to the end user.               |
| <img alt="insert link" />  | The **Insert link** button allows a human agent to send hyperlinked text to the end user.              |

#### Request File Uploads

By default, human agents cannot request files from users. To enable this feature, you need to install the [File Upload](https://github.com/Cognigy/WebchatPlugins/tree/master/plugins/file-upload) plugin in the [Webchat Endpoint](/webchat/v2/configuration#persistent-menu).

File requirements:

* Supported file types are jpeg, jpg, png, docx, wav, ogg, pdf.
* Maximum file size is 40 MB.

To request a file, follow these steps:

1. At the bottom of the **Reply to User** section, click <img alt="request file" />. A window for the file request appears and prompts you to confirm.
2. Click **Send** to confirm that you want to send the file request to the end user to enable the file upload. A message should appear in the chat indicating that the file has been requested.

The end user will see a button in the chat to open a file upload dialog.

<Frame>
  <img alt="Live Agent request file upload webchat" />
</Frame>

After uploading the file, it should be reflected in both the Webchat and the Live Agent interfaces, making the file ready for download on the Live Agent side.

<Frame>
  <img alt="Live Agent request file upload webchat completed" />
</Frame>

The successful upload is displayed in Live Agent.

<Frame>
  <img alt="Live Agent request file upload completed" />
</Frame>

### Simple Quick Reply

For a simple quick reply to a user in the conversation chat, you can use canned responses, variables, or a combination of both.

#### Send Responses with Variables

Live Agent offers a limited list of variables for user contact profile data or human agent profile data that you can re-use in the **Reply** section.

To include a variable, follow these steps:

1. Begin typing <code>\{\{</code> in the field. After that, a list of available variables will appear, allowing you to select the desired variable.
2. Select a variable by using one of these methods:
   * Click the desired variable.
   * Use the arrow keys on your keyboard to navigate to the variable, and then press **Enter**.<br />
     Alternatively, you can type out variables instead of selecting them from the suggestions.

The variable will be replaced with a value relevant to the current conversation.
For example, <code>\{\{contact.first\_name}}</code> → `David`,
<code>\{\{conversation.id}}</code> → `3705`. If any of the variables do not have values, you will receive a corresponding warning.

Note that you cannot create a custom variable.

There is a list of available variables:

| Variable                | Description                                                                                                                                        |
| ----------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------- |
| `conversation.id`       | The ID of the current conversation.                                                                                                                |
| `contact.id`            | The ID belonging to a user who creates a conversation.                                                                                             |
| `contact.name`          | The user's first and last names. If both are not specified in the profile, the default value `end user` will be used.                              |
| `contact.first_name`    | The user's first name. If it is not specified in the profile, the last name will be used. Otherwise, the default value `end user` will be applied. |
| `contact.last_name`     | The user's last name. If it is not specified in the profile, the first name will be used. Otherwise, the default value `end user` will be applied. |
| `contact.email`         | The user's email address.                                                                                                                          |
| `contact.age`           | The user's age.                                                                                                                                    |
| `contact.gender`        | The user's gender.                                                                                                                                 |
| `contact.birthday`      | The user's date of birth.                                                                                                                          |
| `contact.location`      | The user's current location.                                                                                                                       |
| `contact.accepted_gdpr` | The user's acceptance of the General Data Protection Regulation (GDPR) policy.                                                                     |
| `agent.name`            | The first and last names of the human agent.                                                                                                       |
| `agent.first_name`      | The human agent's first name. If it is not specified in the profile, the last name will be used.                                                   |
| `agent.last_name`       | The human agent's last name. If it is not specified in the profile, the first name will be used.                                                   |
| `agent.email`           | The human agent's email address.                                                                                                                   |

The same variables can be used multiple times within a single reply.

Example:

<code>
  Hi \{\{contact.first\_name}},
  Thank you for your recent message. We appreciate your inquiry.
  Your conversation with the ID \{\{conversation.id}} has been received.
  Our team will review it and provide a response as soon as possible.
  \{\{contact.first\_name}}, please let us know if you have any additional questions or need further assistance.
</code>

### Advanced Quick Reply

To include a [canned response](/live-agent/settings/canned-responses) as a reply to the user, follow these steps:

1. In the **Reply to User** or **Reply to Bot** sections, type `/`. A list of available canned responses will be displayed.
2. Select the desired canned response from the list.

The selected canned response will be inserted into your reply, allowing you to easily include prewritten messages in your conversation.

<Frame>
  <img alt="Live Agent Canned Responses" />
</Frame>

#### Send Canned Responses with Variables

You can create text including [variables](#send-responses-with-variables) for [canned responses](/live-agent/settings/canned-responses).

A canned response with variables is called the same way as a [simple canned response](#advanced-quick-reply), by using the symbol `/`.
When you select the canned response, variables in the response will be replaced with values relevant to the current conversation.
If any of the variables do not have values, you will receive a corresponding warning.

## Send Replies to Other Human Agents

As a human agent, you can discuss the end user's conversation with another agent if you need additional clarification or assistance.

To send a reply to another human agent, follow these steps:

1. On the Private Note tab, type `@`. All agents who have been added to the team will be listed.
2. Select an Agent and type in a message to consult with the selected Agent.
3. Click **Add Note** to send the message. You can see a locked sign <img alt="private note" /> in the chat next to the message that has been sent. This message will not be visible to the end user but only to you and your team.

<Frame>
  <img alt="Live Agent Private Message" />
</Frame>

The second agent addressed in the private note will receive a notification in their account.
Subsequent replies by the second agent will be visible on the **Private Note** tab.

You can find all conversations where other agents mentioned you. To do so, use the **@Mentions** option at the top of the Inboxes list.

If you don't send your reply right away but have already composed it, the message will be saved for 24 hours. This means that when you switch to another conversation and return to the previous one, you'll find the initial draft of your private note.

### Delete Private Notes

To delete a private note you have already sent, follow these steps:

1. Click <img alt="menu" /> icon beside the sent message.
2. Select **Delete**. The private note will be deleted.

## Track Human Agent Typing Events

To monitor when human agents start and stop typing, you can set up a WebSocket connection. This connection allows you to monitor typing events in real time using a webhook, which enables immediate updates.
In this implementation, Live Agent utilizes [CloudEvents](https://github.com/cloudevents/spec), which serve as a standardized way to describe and transport event data across different systems and services.

By analyzing typing patterns during conversations, you can identify peak agent activity and pinpoint any bottlenecks in response times. This information can be used to enhance staffing and streamline workflows.

To track human agent typing events, you'll need to update your Live Agent `.yaml` file by adding the necessary configuration variables.

* `EVENT_GATEWAY_ACTIVE=true`
* `EVENT_GATEWAY_HTTP_ENDPOINT="your-webhook-url"`, replace `your-webhook-url` with the URL to which you want to send events for tracking. For example, `https://webhook-test.com/e183f157-5703-4258-a519-fba23ca15b09`.

After configuration, your webhook will receive the event payload in the following format:

```json theme={null}
{
  "specversion": "1.0",
  "type": "ai.cognigy.live-agent.conversation.typing_off",
  "source": "http://127.0.0.1:3000",
  "subject": "conversation.2",
  "id": "01J9KKW5VS6A0FPWDQ11BHRVWX",
  "time": "2024-10-07T13:49:35Z",
  "datacontenttype": "application/json",
  "data": {
    "cognigy_session_id": "c101cbc0-807d-426f-bd81-tb756dbdec44"
  }
}
```

| Parameter                 | Type   | Description                                                                                                                                                                                                                                                                |
| ------------------------- | ------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| specversion               | String | Indicates the version of the [CloudEvents](https://github.com/cloudevents/spec) specification being used. For example, `1.0`.                                                                                                                                              |
| type                      | String | Specifies the type of event. The following events are available: <br />- `ai.cognigy.live-agent.conversation.typing_off` — indicates when a human agent stops typing. <br />- `ai.cognigy.live-agent.conversation.typing_on` — indicates when a human agent starts typing. |
| source                    | String | The source of the event, typically a URL identifying the origin. For example, `http://127.0.0.1:3000`.                                                                                                                                                                     |
| subject                   | String | The context for the event, specifying the entity. For example, `conversation.2`.                                                                                                                                                                                           |
| id                        | String | A unique identifier for the event for distinct recognition. For example, `01J9KKW5VS6A0FPWDQ11BHRVWX`.                                                                                                                                                                     |
| time                      | String | A timestamp indicating when the event occurred, formatted in the ISO 8601 standard. For example, `2024-10-07T13:49:35Z`.                                                                                                                                                   |
| datacontenttype           | String | The media type of the data being sent. For example, `application/json`.                                                                                                                                                                                                    |
| data                      | Object | Contains the actual event data, which holds additional information relevant to the event.                                                                                                                                                                                  |
| data.cognigy\_session\_id | String | A unique identifier for the Cognigy session associated with this event. For example, `c101cbc0-807d-426f-bd81-tb756dbdec44`.                                                                                                                                               |

## More Information

* [Conversation](/live-agent/conversation/overview)
* [Conversation Assignment](/live-agent/conversation/assign-conversations)


# Add Another Inbox for your Project
Source: https://docs.cognigy.com/live-agent/getting-started/live-agent-setup/live-agent-setup-additional-inbox

Create additional Inboxes in Live Agent and configure them for handover within your Cognigy Flows.

<a href="/release-notes/2025.10"><Badge>Updated in 2025.10</Badge></a>

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

You can create extra Inboxes in Live Agent and configure them for handover within your Flows.

## Add an Additional Inbox

To add another Inbox in Live Agent, follow these steps:

1. In the sidebar of the Live Agent dashboard, click **Inboxes**.

2. In the upper-right corner, click the **+ Add Inbox** button. Live Agent will show you a screen to set up your Inbox.

   <Frame>
     <img alt="Live Agent Inbox Setup" />
   </Frame>

3. On the Inbox setup page, fill in the following fields:
   * **Inbox Name** — create a unique Inbox name. For example, `Cognigy Handover`.
   * **Webhook URL** — enter the Cognigy Endpoint Config URL domain name, extended by `/handover/liveagent`. For example, `https://endpoint-trial.cognigy.ai/handover/liveagent`. You can find the Endpoint Config URL in the **Configuration Information** section of your Endpoint.

     <Frame>
       <img alt="Endpoint Configuration Information section showing the Config URL" />
     </Frame>

4. Click the **Create Cognigy Inbox** button. Once you have successfully set up the Inbox, you are prompted in the next step to assign your Agents to the new Inbox.

5. Select your required Agents in the selection list and click **Add Agents**.

A new Live Agent Inbox has been created.

### Find Account and Inbox IDs

The new Inbox is defined by an Account ID and an Inbox ID. The Inbox ID is required to configure the **Handover to Human Agent** Node.

Both IDs can be found in the Live Agent URL after creating the Inbox.

To locate them, follow these steps:

1. Go to **Settings > Inboxes**.
2. In the **Inboxes** list, find the Inbox that you have created before, and click the **Settings** icon next to the Inbox.

The Inbox settings will be displayed,
and you will see the address line of your browser as follows:
`https://liveagent-<your-environment>/app/accounts/<account-id>/settings/inboxes/<inbox-id>`.
For example, in the URL `https://liveagent-dev.cognigy.ai/app/accounts/2/settings/inboxes/32852`, the Account ID is `2`,
and the Inbox ID is `32852`.

<Frame>
  <img alt="Live Agent browser URL showing the Account ID and Inbox ID" />
</Frame>

## Configure a Flow for Handover to the Additional Inbox

To configure a Flow for a Handover to a Live Agent, follow these steps:

1. In the Cognigy interface, open the Flow that contains the Handover Node.

   <Frame>
     <img alt="Cognigy Flow editor with a Handover to Human Agent Node" />
   </Frame>

2. Click the **Handover to Human Agent** Node to open the Node Editor.

3. In the **Live Agent Settings** section, enter the Inbox Id of your new Live Agent Inbox.

4. Click **Save Node**.

Your Flow is now configured for a Live Agent Handover.

## What's Next?

* [Explore Conversation Workflow](/live-agent/conversation/overview)
* [Configure Conversation Routing](/live-agent/conversation/conversation-routing/overview)
* [Customize your Live Agent by changing the Account Settings](/live-agent/settings/overview)
* [Group human agents in Teams](/live-agent/settings/teams)
* [Organize your work by adding Labels](/live-agent/settings/labels)


# Add Agents and Admins to Live Agent
Source: https://docs.cognigy.com/live-agent/getting-started/live-agent-setup/live-agent-setup-agents

Create users for human agents and admins in Live Agent, assign user roles, and connect them to inboxes.

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

Once you have [set up your Endpoint and Flow for Handover to Live Agent](/live-agent/getting-started/live-agent-setup/live-agent-setup-handover) and tested a conversation, you can then proceed to add new Agents and Admins and assign them to Inboxes in Live Agent.

## Create Users

To create users, follow these steps:

1. Open the Cognigy interface. If the Agent or Admin you want to add already exists as a User in Cognigy, proceed to [Adding Live Agent Roles to Users](#add-live-agent-roles-to-users).
2. On the **Main** page, in the upper-right corner, click **User Menu <img alt="user menu" /> > Admin Center**.
3. In the Admin Center interface, select **Access Control** from the left-side menu, then click **+ Create new User**.
4. In the **New User** window, fill in the following fields:
   * **Display Name** — enter the name that you want to be displayed for this user. For example, `John Doe`.
   * **Email Address** — enter the user's email address. For example, `johndoe@example.com`.
   * **Password** — set a password for the user's account. A user can change a password later.
5. Click **Save**.

Once a user has been created, you can assign Live Agent roles to this user.

## Add Live Agent Roles to Users

Cognigy Live Agent has three different user roles:

* `Administrator`
* `Agent`
* `Supervisor`

For more information, see [Live Agent Roles](/live-agent/roles).

To add one of these roles to users, follow these steps:

1. On the **Main** page of the Cognigy interface, click **User Menu <img alt="user menu" /> > Admin Center** in the upper-right corner.
2. In the Admin Center interface, select **Access Control** from the left-side menu, then click **+ Create new User**.
3. Select a user you would like to assign roles to.
4. On the User page, go to the **Global Roles** section.
5. Click **+** and select the desired role.
6. Assign one of the following roles:
   * `liveAgentAdmin` — for users intended to have Admin rights in Live Agent.
   * `liveAgentAgent` — for users intended to have Agent rights in Live Agent.

After assigning the Live Agent Role to the user, you can add them to an Inbox.

## Add Agents to an Inbox

To add an Agent to an Inbox, follow these steps:

1. In the left-side menu of the Live Agent interface, select **Settings**.
2. Go to **Inboxes**, select an Inbox and click the gear icon next to your Inbox.
3. Open the **Collaborators** tab.
4. Go to the **Agents** section, click the Agent list and select Agents from this list.
5. Click **Update**.

When the Agents have been included in the Inbox, you can [add another Inbox for your Project](/live-agent/getting-started/live-agent-setup/live-agent-setup-additional-inbox).


# Set up Handover for Live Agent
Source: https://docs.cognigy.com/live-agent/getting-started/live-agent-setup/live-agent-setup-handover

Set up handover for Live Agent by configuring your Cognigy Endpoint and Flow, enabling smooth transitions from AI to human agent support.

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

After configuring your Project Inbox for Live Agent, proceed to set up your Cognigy Agent for Handover to Live Agent.

Configuring the Handover for Live Agent involves two steps:

1. [Set up your Endpoint for Handover to Live Agent](/live-agent/getting-started/live-agent-setup/live-agent-setup-handover-endpoint)
2. [Set up your Flow for Handover to Live Agent](/live-agent/getting-started/live-agent-setup/live-agent-setup-handover-flow)


# Set up a Cognigy Endpoint for Handover
Source: https://docs.cognigy.com/live-agent/getting-started/live-agent-setup/live-agent-setup-handover-endpoint

Configure a Live Agent Endpoint for Live Agent in Cognigy to hand over conversations from AI Agents to humans.

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

When you create a new Agent in Cognigy you are prompted to select an Endpoint. You can select, for example, **Cognigy Webchat** as a specific Endpoint to the Cognigy platform.
For more information, see [Endpoints](https://docs.cognigy.com/ai/endpoints/overview/).

If you want to integrate a Handover process in your Flow, you need to configure your Endpoint for Handover.

## Prerequisites

Before setting up the Endpoint, check whether these steps are done:

* [Live Agent is set up for your Organization](/live-agent/getting-started/live-agent-setup/live-agent-setup-org)
* [The Inbox is created for your Live Agent account](/live-agent/getting-started/live-agent-setup/live-agent-setup-inbox)

## Handover Configuration

To configure the Handover Settings of your Endpoint, follow these steps:

1. In the left-side menu of the Cognigy interface, select **Deploy > Handover Providers**.
2. Click **+ New Handover Provider** and select **Live Agent**.
3. Enter a unique name and click **Save**.

To configure your Live Agent handover, you have two options:

<AccordionGroup>
  <Accordion title="Handover to pre-configured Live Agent">
    If the **Handover to pre-configured Live Agent account** option is activated, conversations will be transferred to the Live Agent account mapped to the current Cognigy organization.

    For on-premises installation, make sure the following environment variables are set in the [value.yaml](https://github.com/Cognigy/cognigy-ai-helm-chart/blob/main/values.yaml) configuration file.

    ```txt theme={null}
    FEATURE_USE_COGNIGY_LIVE_AGENT: true
    COGNIGY_LIVE_AGENT_API_BASE_URL_WITH_PROTOCOL: <https://sub-domain.domain.com>
    COGNIGY_LIVE_AGENT_PLATFORM_TOKEN: <app-platform-token>
    ```
  </Accordion>

  <Accordion title="Handover to manually configured Live Agent">
    To configure your Live Agent manually, follow these steps:

    1. In the handover provider settings, deactivate the **Handover to pre-configured Live Agent account** option.
    2. Enter the **Base URL** for the Live Agent instance. The Base URL is your Live Agent installation URL, extended by `/api/v1`. For example, in the Live Agent trial environment `https://liveagent-trial.cognigy.ai/api/v1` is the Base URL where `https://liveagent-trial.cognigy.ai` is the Live Agent URL.
    3. Enter the **Account ID** and the **API Key**. How to find the **Account ID** and the **API Key**, see sections below.
    4. Click **Save** to confirm your Live Agent settings.

    Your Live Agent has been configured.
  </Accordion>
</AccordionGroup>

## Find an Account ID

You can find the account ID in the Live Agent URL of your Live Agent **Inbox**.

To locate the ID, follow these steps:

1. Open your Live Agent account and go to **Settings > Inboxes**.
2. In the Inboxes list, find the Inbox of your project, and click the **Settings** icon next to the Inbox.
3. The Inbox settings will be displayed, and you will see the address line of your browser as follows: `https://liveagent-<your-environment>/app/accounts/<account-id>/settings/inboxes/<inbox-id>`. For example, in the URL `https://liveagent-dev.cognigy.ai/app/accounts/2/settings/inboxes/32852`, the account ID is 2.

<Frame>
  <img alt="Live Agent browser URL highlighting the account ID in the Inbox settings" />
</Frame>

When you have found the account ID, you can enter it in the **Handover Settings** of your Endpoint in Cognigy.

## Find API Keys

1. Click the **Profile Settings** icon of your account in the bottom-left corner of the Live Agent dashboard, then select **Profile Settings** in the selection menu.
2. Find the **Access Token** section, showing your API Key.
3. Copy the API Key and paste it into the **API Key** field in the **Live Agent Authentication** section of your handover provider settings in Cognigy.

## What's Next?

When the Endpoint is configured, [set up your Flow for Handover to Live Agent](/live-agent/getting-started/live-agent-setup/live-agent-setup-handover-flow)


# Set up a Cognigy Flow for Handover
Source: https://docs.cognigy.com/live-agent/getting-started/live-agent-setup/live-agent-setup-handover-flow

Configure a Cognigy Flow with a Handover Node to route conversations from an AI Agent to a human agent in Live Agent.

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

Once you've set up your Endpoint for handover to a Live Agent, the next step is configuring your Flow for this purpose within Cognigy.

## Set up Flows for Handover

Once you have [Set up your Endpoint for Handover to Live Agent](/live-agent/getting-started/live-agent-setup/live-agent-setup-handover-endpoint), you can set up your [Flow](/live-agent/getting-started/live-agent-setup/live-agent-setup-handover-flow) for handover to Live Agent.

To create a Live Agent Flow, follow these steps:

1. In the Cognigy interface, open the Project.
2. In the left-side menu of the Project, go to **Build > Flows**.
3. Click **+ New Flow**.
4. Specify a unique name and provide a relevant description. For example, `Customer Support Flow`.
5. Click **Create**.
6. In the Flow editor, add a Say Node.
7. Open the Node editor by clicking the Say Node.
8. In the **Text** field, enter the following text: `Hi, let's hand you over to the humans`.
9. Click **Save Node**.
10. Below the Say Node, add a Handover to Human Agent Node.
11. Open the Node editor by clicking the Handover to Human Agent Node.
12. In the **Handover Accepted Message** field, enter the following text: `Performing the Handover...`
13. In the **Live Agent Settings** section, fill in the following fields:
    * **Live Agent Inbox Id** — enter the Inbox ID that you obtained in the [previous step](/live-agent/getting-started/live-agent-setup/live-agent-setup-handover-endpoint).
    * **Skills** — enter a list of skills that should be used to filter the available agents. For example, `technical support, billing`.
    * **Languages** — enter a list of languages that should be used to filter the available agents. For example, `english, spanish`.
    * **Priority** — select a priority for the conversation from the list. For example, `Urgent`.
14. Click **Save Node**.

## Test your Connection

1. Go to the Endpoint that you created in the Cognigy interface.

2. Click **Open Demo Webchat**.

3. Start the conversation by typing `Hi` and clicking the **Send** button.

   <Frame>
     <img alt="Demo Webchat with the initial Hi message sent by the user" />
   </Frame>

4. In the bottom-left corner of your Project, click **User Menu > Live Agent**.

5. In the left-side menu of the Live Agent interface, select **Conversations**.
   On the **All** tab, the conversation that you started should appear in the conversation list.

6. Open the conversation.

7. In the **Reply** section, send a message to the user as a human agent.

   <Frame>
     <img alt="Live Agent conversation panel showing the human agent's reply in the Reply section" />
   </Frame>

8. This message should also appear in the Cognigy Webchat window.

   <Frame>
     <img alt="Cognigy Webchat showing the human agent's reply message received from Live Agent" />
   </Frame>

You have successfully completed your first conversation in Live Agent.

## What's Next?

Next, you can [add more users](/live-agent/getting-started/live-agent-setup/live-agent-setup-agents) with `Agent` and `Administrator` roles, and assign them to Inboxes in Live Agent:


# Set up the Inbox for Live Agent
Source: https://docs.cognigy.com/live-agent/getting-started/live-agent-setup/live-agent-setup-inbox

Create and configure the default Live Agent inbox in Cognigy.

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

After [setting up Live Agent for your organization](/live-agent/getting-started/live-agent-setup/live-agent-setup-org), configure your default Live Agent Inbox.

## Agent Settings

To create a new Inbox for the Agent, follow these steps:

1. In the Cognigy interface, open an existing [Agent](/ai/agents/develop/projects-and-flows/projects).
2. In the left-side menu of the Agent, click **Manage > Settings**.
3. Scroll down and find **Live Agent Settings**.
4. Click **Create Project Inbox**.

Once the Inbox is created, the **Default Inbox ID** button will be disabled.

<Frame>
  <img alt="Live Agent settings showing the Default Inbox ID after Inbox creation" />
</Frame>

## Troubleshooting

### Create Project Inbox disabled

The **Create Project Inbox** button could also be disabled for the following reasons:

* You do not have the Cognigy Live Agent feature enabled.
* The organization does not have a Live Agent account set up.
* You do not have permissions to create an Inbox for the Agent.

**Example 1**

The Cognigy Live Agent feature is not enabled, or a Live Agent account is not set up at the organizational level.

<Frame>
  <img alt="Disabled Create Project Inbox button when Live Agent is not enabled or configured" />
</Frame>

**Example 2**

You don't have permissions to create an Inbox for Live Agent.

<Frame>
  <img alt="Disabled Create Project Inbox button when user lacks permissions" />
</Frame>

If you have both **Theming** enabled and the Cognigy Live Agent feature disabled, The **Live Agent Settings** section on both the **Admin Center** and **Agent Settings** pages will not be displayed.

## What's Next?

Now you need to set up a [Handover](/live-agent/getting-started/live-agent-setup/live-agent-setup-handover) to Live Agent.


# Set up Live Agent for your Organization
Source: https://docs.cognigy.com/live-agent/getting-started/live-agent-setup/live-agent-setup-org

Activate and configure Cognigy Live Agent for your organization via the Admin Center.

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

To set up the Live Agent Account for your organization, follow these steps:

1. On the **Main** page of the Cognigy interface, click the **User Menu** in the upper-right corner.
2. Select the **Admin Center** from the menu.
3. Click **Set up Live Agent**.

<Frame>
  <img alt="Set up Live Agent button in the Cognigy Admin Center" />
</Frame>

When you set up Live Agent for your Organization, you can [create your Project Inbox for Live Agent](/live-agent/getting-started/live-agent-setup/live-agent-setup-inbox).

## Troubleshooting

If the **Set up Live Agent** button is disabled, it could be for the following reasons:

* You do not have the Cognigy Live Agent feature enabled.
* You do not have permissions to create a Live Agent account.
* There is already a Live Agent account created for your organization.

**Example 1**

Cognigy Live Agent feature is not enabled.

<Frame>
  <img alt="Disabled setup button when the Live Agent feature is not enabled" />
</Frame>

**Example 2**

Live Agent Account is already set up for the Organization.

<Frame>
  <img alt="Disabled setup button when Live Agent is already configured for the organization" />
</Frame>


# Getting Started with Live Agent
Source: https://docs.cognigy.com/live-agent/getting-started/overview

Get started with Cognigy Live Agent by following the setup steps to integrate it with Cognigy, configure handover, and add human agents.

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

Cognigy Live Agent enhances your customer service by extending your Conversational.AI platform, adding a Handover service to a Human Agent.

To integrate Live Agent with the Cognigy platform, follow these steps:

1. [Set up Live Agent for your Organization](/live-agent/getting-started/live-agent-setup/live-agent-setup-org)
2. [Set up your Project Inbox for Live Agent](/live-agent/getting-started/live-agent-setup/live-agent-setup-inbox)
3. [Set up your Endpoint and Flow for Handover to Live Agent](/live-agent/getting-started/live-agent-setup/live-agent-setup-handover)
4. [Add Agents and Admins to Live Agent](/live-agent/getting-started/live-agent-setup/live-agent-setup-agents)
5. *(Optional)* [Add another Inbox for your Project](/live-agent/getting-started/live-agent-setup/live-agent-setup-additional-inbox)


# Live Agent Glossary
Source: https://docs.cognigy.com/live-agent/glossary

Live Agent terminology, defining key concepts such as conversations, inboxes, agents, labels, and automation rules.

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

Live Agent terms listed in alphabetical order. All terms in the glossary are given for informational purposes.

## A

**Abandoned (conversations)**

The status that indicates that a conversation is closed because the user left the conversation. It is configurable under the Account Settings section.

**Account**

A separate Live Agent instance that represents the entire organization or company. In this context, all the resources, data, and users related to the organization are managed centrally at the org level. In other alternative Live Agent systems, this entity may also be referred to as Tenant.

**Agent Copilot Bot**

A tool for assisting the Agents while talking to the customers. It consists of a dedicated Flow to provide information to the Agent. Therefore, it needs to be configured as a separate Flow in Cognigy.
For more information, read the [Agent Copilot Bot](/live-agent/assistants/agent-copilot-bot) documentation.

**Agent Copilot Whisper**

A virtual assistant that helps human agents improve their performance. It displays AI-powered suggested replies and next-action suggestions directly in the chat box. When a user sends a message, a suggested reply will appear in a centered frame inside the chat. A human agent can ignore it or click the reply arrow icon to send. For more information, read the [Agent Copilot Whisper](/live-agent/assistants/agent-copilot-whisper) documentation.

**Agent Copilot Workspace**

A virtual assistant that helps human agents improve their performance. It consists of an embedded or standalone collection of interactive widgets arranged in a customizable grid. The widgets actively support the human agent by displaying various information, including user data, user sentiment, suggesting next action, providing chat transcription, and more. For more information, read the [Agent Copilot](/agent-copilot/index) documentation.

**Account Settings**

You can change the general settings for an account, such as name and language. You will find access via the dashboard sidebar of Live Agent. For more information, read the [Account Settings](/live-agent/settings/account-settings) documentation.

**Agents**

Members of your Customer Support team. Agents will be able to view and reply to messages from your users. On the Live Agent sidebar you can click Settings > Agents to configure it. To differentiate between AI Agents and human agents in the Live Agent documentation, we primarily use the [Human Agent](#h) term to denote members of the support team. For more information, read the [Agent](/live-agent/settings/agents) documentation.

**Agent Queue**

A list of conversations waiting for attention from human agents. In Live Agent, a queue is similar to the waiting area, where conversations are held until the human agent has sent the first reply. The queue is visible to human agents on the Mine tab of the Conversation page, where they can see the conversations assigned to them and wait for their attention. For more information, read the [Agent Queue](/live-agent/conversation/conversation-queue/overview) documentation.

**Agent Status**

Agent statuses help manage the availability and workload of human agents, ensure consistent and efficient customer service, and provide transparency regarding the agent's availability to handle chat requests. For more information, read the [Agent Statuses](/live-agent/settings/agents) documentation.

**API Access Token**

A data structure that contains authentication and authorization information to use an application. It can be used if you are building an API-based integration and its actions are limited by the user role. For more information, read the [Profile Settings](/live-agent/profile-settings) documentation.

**Assigned (conversations)**

Every Conversation in Live Agent can have an assignee. An assignee can be manually added to the Conversation via the UI, the API, or automatically via an available round-robin assignment system.
For more information, read the [Conversation Routing](/live-agent/conversation/conversation-routing/overview) documentation.

**Attachments**

In an ongoing Live Agent conversation, both the user and the agent can send files such as images, videos, documents, and forms directly during the conversation. All attachments within account conversations are accessible in the Settings menu.
For more information, read the [Attachments](/live-agent/settings/attachments) documentation.

**Audit Logs**

A tool for maintaining transparency and accountability within your Live Agent resources. These logs record important user activities, system events, and configuration changes, providing details for security and auditing purposes. With audit logs, you can answer "Who did what, where, and when?" questions within your Live Agent environment. The logs include events such as create, update, and delete for Inboxes and Automation Rules. Only users with the Administrator role can access these logs.
For more information, read the [Audit Logs](/live-agent/settings/audit-logs) documentation.

**Automation**

Automation rules can replace and automate existing processes that require manual effort. You can do many things with automation, including adding labels and assigning conversations to the best agent. So the team focuses on what they do best and spends less time on manual tasks.
For more information, read the [Automation](/live-agent/settings/automation-rules) documentation.

**Away (Agents)**

One of four possible Live Agent statuses.
This status indicates that the Agent is temporarily unavailable or not actively monitoring the chat system. Agents in this status cannot handle customer conversations.
For more information, read the [Agent Statuses](/live-agent/settings/agents) documentation.

## B

**Busy (Agents)**

One of four possible Live Agent statuses.
This status indicates that the Agent is currently engaged in productive activities or actively assisting a customer. Agents in the Busy status temporarily defer new customer interactions.
For more information, read the [Agent Statuses](/live-agent/settings/agents) documentation.

## C

**Canned Responses**

Pre-saved reply templates allow human agents to respond to a conversation quickly.
Canned Response templates provide support for HTML formatting.
For creating a Canned Response, just click the Add Canned Response.
Canned responses are used with the help of Short Codes. Agents can access canned responses while on a chat by typing `/` followed by the short code.
For more information, read the [Canned Responses](/live-agent/settings/canned-responses) documentation.

**Cognigy Live Agent OAuth**

You can log in to Live Agent using your Cognigy account by utilizing the OAuth2 protocol. To configure this, follow the instructions provided in the Helm Values section.
For more information, read the [Live Agent OAuth](/live-agent/installation/deployment/cognigy-live-agent-oauth) documentation.

**Conversation**

Your chats with various customers are referred to as conversations in Live Agent.
This includes all session messages, including the customer, bot or human agent responses.
A Conversation is the communication channel opened between an Agent and a customer.
For more information, read the [OData Analytics Endpoint](/live-agent/tools/odata-endpoint#conversation) and the [Conversation workflow](/live-agent/conversation/overview) documentation.

**Conversation Routing**

The process of directing customer or user inquiries to the right agent or team. It matches inquiries with agent skills, languages, and workload, and ensures efficient issue resolution. How you can assign an Agent to a Conversation is described in the [Conversation Routing](/live-agent/conversation/conversation-routing/overview) documentation.

**Conversation Status**

Communication between customer and operator in Live Agent can get different statuses:
Unassigned, Assigned, Pending, Snoozed, Open, Resolved and Abandoned.

**Conversation Workflow**

A structured sequence of steps that guide a conversation, from initiation to conclusion. It outlines how interactions progress, often involving greeting, information gathering, responses, problem resolution, and closing. For more information, read the [Conversation Workflow](/live-agent/conversation/overview) documentation.

## H

**Human Agents**

Human members of your Customer Support team. They will be able to view and reply to messages from your users directly or, for example, when an AI Agent is asked to forward a customer request to a human support member.

## I

**Inboxes**

Workspaces where human agents manage real-time chat interactions with customers and handle related tasks. In your Live Agent account, you can have unlimited inboxes, and you can view conversations specific to an inbox by clicking on it. Additionally, you can modify settings for already existing inboxes. For more information, read the [Inboxes](/live-agent/settings/inboxes) documentation.

**Inbox ID**

A unique identifier that represents the connection between Cognigy and Live Agent. It is used to link and manage the integration between these two systems, enabling seamless communication and data exchange.

## L

**LA Copilot**

LA Agent Copilot allows human agents using Cognigy Live Agent to be assisted by AI-powered agents when conversing with the customer. Using this functionality, during an active handover, fully configurable AI Agent messages can be displayed to the human agent based on customer messages to best assist the customer. For more information, read the [LA Agent Copilot](/live-agent/assistants/overview) documentation.

**Labels**

Tags that help you to categorize and prioritize conversations, providing additional context and information about the content or status of each conversation. For more information, read the [Labels](/live-agent/settings/labels) documentation.

**Live Agent**

Cognigy Live Agent is a product that combines conversational AI and human agents to enhance customer experiences and improve support team productivity. It allows a smooth transition between customer-AI Agent conversations and human support, making it easier to handle customer inquiries effectively. For more information, read the [Live Agent](/live-agent/overview) documentation.

**Live Agent Settings**

Cognigy Live Agent settings offer many features, such as team and inbox management, automation rules, canned responses, attachments, audit logs, and account settings. These capabilities can significantly improve customer interactions and enhance support efficiency. Note that these settings overwrite all Profile settings. For more information, read the [Live Agent settings](/live-agent/settings/overview) documentation.

## M

**Mentions**

You have the option to tag a teammate in private notes to bring their attention to a specific topic for discussion. Additionally, you can access conversations that specifically mention you through the Mentions view.

## N

**Notifications**

Messages or alerts that are generated to inform human agents about new events, activities, or interactions within a conversation chat system in Live Agent. For more information, read the [Notifications](/live-agent/notifications) documentation.

## O

**OData Analytics Endpoint**

Live Agent exposes an OData 4 analytics Endpoint to retrieve analytics records. OData, the best way to REST, is a powerful API framework. The OData Endpoints allow you to retrieve the most relevant data models from Live Agent by using GET requests. It covers all your enterprise analytics needs to make fine-grained queries in your spreadsheets or build rich dashboards for your bots with your favorite data visualization tool. Note that the Live Agent OData Endpoint only supports GET requests and does not support any other request types, such as `PATCH`, `DELETE`, or `POST`. For more information, read the [OData Endpoint](/live-agent/tools/odata-endpoint) documentation.

**Offline (Agents)**

One of four possible Live Agent statuses. This status indicates that the agent is not currently logged in to the Live Agent system. Agents in this status cannot handle customer interactions. For more information, read the [Agent Statuses](/live-agent/settings/agents) documentation.

**Online (Agents)**

One of four possible Live Agent statuses. This status indicates that the Agent is available and actively handling customer interactions. For more information, read the [Agent Statuses](/live-agent/settings/agents) documentation.

**Open (conversations)**

This conversation status indicates that a conversation is prepared for a handover to a human agent.

## P

**Pending (conversations)**

This status reflects the quantity of conversations that were temporarily paused, awaiting a response from either the customer or a human agent.

**Profile Settings**

A group of settings defined by an end user that are important for configuration in Live Agent. Basic User Profile Settings can be changed or updated. For more information, read the [Profile Settings](/live-agent/profile-settings) documentation.

## R

**Reports**

The conversation reports which have occurred over time. They can be viewed in the Reports section. Click Reports in the Live Agent Dashboard sidebar to get an overview displayed. A graph will show the Conversations that have occurred during the selected days. For more information, read the [Reports](/live-agent/reports/overview) documentation.

**Resolved (conversations)**

This conversation status means that a conversation is closed because the user's problem has been fixed.

**Roles**

A method of granting permissions to specific users for data access. These permissions are based on a user role profile. In Live Agent users can have three different Single Roles: `Agents`, `Supervisors`, and `Administrators`. You can also assign more than one role to a human agent. For more information, read the [Roles](/live-agent/roles) documentation.

## S

**Skills**

Specific abilities possessed by Agents, such as language proficiency or expertise in a particular domain. You can add Skills using the Agents dashboard of Live Agent.

**Snoozed (conversations)**

This conversation status is applied when you need to wait for a specific period or require a reply from the customer to continue the conversation. In a conversation workflow, it indicates that the issue will be set aside for now, and the status will change to Open when a time trigger or customer response trigger was activated.

## T

**Teams**

Groups based on human agents' responsibilities. A user can be part of multiple Teams. You can assign conversations to a team when you are working collaboratively. For more information, read the [Teams](/live-agent/settings/teams) documentation.

## U

**Unassigned (conversations)**

This conversation status is applied when a conversation has not yet been assigned to a human agent.


# Cognigy Live Agent
Source: https://docs.cognigy.com/live-agent/index



<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

<div>
  <div>
    *Cognigy Live Agent* is a contact center solution that is designed to work exclusively with Cognigy.

    <p>
      Live Agent is built for human agents to handle escalated conversations from AI Agents.
    </p>

    <p>
      The system automatically assigns human agents based on their expertise and availability, which helps businesses optimize efficiency, scalability, and deliver exceptional customer support.
    </p>
  </div>

  <div>
    <video aria-label="Cognigy Live Agent hero animation showing a group of three human agents.">
      <source type="video/mp4" />
    </video>

    <video aria-label="Cognigy Live Agent hero animation showing a group of three human agents.">
      <source type="video/mp4" />
    </video>

    <button type="button" aria-label="Pause video">
      ⏸
    </button>
  </div>
</div>

<CardGroup>
  <Card title="About Live Agent" icon="info" href="/live-agent/overview">
    Learn more about Live Agent features and set up Live Agent to your needs by creating your first inbox and adding agents to it.
  </Card>

  <Card title="Conversations" icon="mail" href="/live-agent/conversation/overview">
    Conversation is a core component of Live Agent. Explore conversation's interface and start implementing conversation routing.
  </Card>

  <Card title="Build Processes" icon="hammer" href="/live-agent/settings/agents">
    Organize your human agents' work. Create inboxes, add agents, and group them into teams.
  </Card>

  <Card title="Assistants" icon="bot" href="/live-agent/assistants/overview">
    Use virtual assistants to support your human agents, enabling them to respond to user queries more accurately and efficiently.
  </Card>

  <Card title="Analytics Tools" icon="chart-column" href="/live-agent/reports/overview">
    Use to enhance and optimize human agent performance, maximizing productivity and efficiency.
  </Card>

  <Card title="Automation and Logs" icon="workflow" href="/live-agent/settings/automation-rules">
    Automate routine tasks and track user actions in Live Agent.
  </Card>

  <Card title="Administer" icon="user-cog" href="/live-agent/settings/account-settings">
    Manage your account settings and assign roles such as Agent, Supervisor, and Administrator to different users based on their responsibilities and access level.
  </Card>

  <Card title="User Profile and Notifications" icon="bell-dot" href="/live-agent/profile-settings">
    Live Agent users can manage their profile settings and configure notifications based on their preferences.
  </Card>
</CardGroup>


# Browser Requirements
Source: https://docs.cognigy.com/live-agent/installation/browser-requirements

Live Agent supports Firefox, Chrome, Chromium, Safari, and Edge. Check the minimum supported browser version requirements for full compatibility.

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

The Live Agent web frontend is compatible with most modern web browsers. The table below provides a list of supported web browsers.

| Browser         | Version        |
| --------------- | -------------- |
| Mozilla Firefox | 110 or higher  |
| Google Chrome   | 110 or higher  |
| Chromium        | 86 or higher   |
| Apple Safari    | 14.1 or higher |
| Microsoft Edge  | 110 or higher  |

<Note>
  Push notifications work as expected in Google Chrome and Mozilla Firefox. For the other browsers, Cognigy cannot provide a guarantee.
</Note>


# Notifications
Source: https://docs.cognigy.com/live-agent/notifications

Set up email, push, and audio notifications in Live Agent to keep agents informed about new conversations, assignments, and customer messages in real time.

<a href="/release-notes/2025.25"><Badge>Updated in 2025.25</Badge></a>

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

*Notifications* are messages or alerts that are generated to inform human agents about new events, activities, or interactions within a conversation chat system in Live Agent. These notifications serve as a way to keep agents updated and responsive to customer inquiries in a timely manner.

## Restrictions

* Modern browsers block automatic audio playback for security reasons. As a result, audio notifications in Live Agent only trigger after a human agent interacts with the interface, for example, by clicking a tab, pressing a key, or moving the mouse.
* If a human agent has multiple inboxes and is not in the inbox specified in the [Handover to Human Agent](/ai/agents/develop/node-reference/service/handover-to-agent) Node, audio notifications for that inbox may be delayed. Alternatively, the human agent can check [unread notifications](#view-unread-notifications).

## Notification types

In Live Agent, you can configure email and push notifications, and also add audio alerts for certain types of actions that trigger notifications.

The table below lists the actions in Live Agent for which notifications can be configured.

| Action / Notification Type                           | Email | Push | Audio |
| ---------------------------------------------------- | ----- | ---- | ----- |
| A new conversation is created                        | +     | +    | +     |
| A conversation is assigned to a user                 | +     | +    | +     |
| A user is mentioned in a conversation                | +     | +    | -     |
| A new message is created in an assigned conversation | +     | +    | -     |

### Push Notification Requirements

For push notifications, the following requirements exist:

* Push notifications must be enabled in Live Agent and in your browser. Otherwise, they are not available.
* Push notifications are not supported by the Microsoft Edge browser.
* Push notifications work as expected in Google Chrome and Mozilla Firefox. For the other browsers, Cognigy cannot provide a guarantee.

### Customization of Email Notifications

For on-premises installations, you can customize your [email template](/live-agent/installation/helm-values/email-templates).

## Configure Notifications

Notifications can be configured at three levels:

* [inbox level](#configure-notifications-at-the-inbox-level) (only for users with the `Supervisor` role)
* [user profile level](#configure-notifications-at-the-user-profile-level)
* [account level](#configure-notifications-at-the-account-level)

### Configure Notifications at the Inbox Level

You can configure email and push notifications for supervisors at the Inbox level. This configuration applies to all supervisors within the selected Inbox and overrides the [Profile Settings](#configure-notifications-at-the-user-profile-level).

1. In the Live Agent interface, go to **Settings > Inboxes**.
2. On the **Inboxes** page, to the right of the Inbox you want to change, click **Settings** and open the **Collaborators** tab.
3. In the **Supervisors** section, activate the following notifications:
   * **Send push notifications to supervisors** — activate the option to let supervisors receive real-time notifications when a conversation is escalated.
   * **Send email notifications to supervisors** — activate the option to let supervisors receive email notifications when a conversation is escalated.
4. Click **Update**.

For more information, refer to [Escalate Conversations to Supervisors](/live-agent/settings/inboxes#escalate-conversations-to-supervisors).

### Configure Notifications at the User Profile Level

In Live Agent, a user with any role can configure audio, email, and push notifications within their profile.

<Note>
  The email and push notification settings in the [Account Settings](#configure-notifications-at-the-account-level) will take precedence over the Profile Settings.
  If the administrator has not yet changed the push and email notification settings, a user still has the opportunity to change them in the Profile Settings.
</Note>

To configure notifications at the user profile level, follow these steps:

1. In the bottom-left corner of the Live Agent screen, click the **Profile Settings** icon.
2. From the list, select **Profile Settings**.
3. On the **Profile Settings** setting, activate the following settings:

* **Audio Notifications** — enable audio notifications. Select one of the following options:
  * **None** — the audio notifications are turned off.
  * **Assigned Conversations** — the audio notifications work only when the conversation is assigned to a user.
  * **All Conversations** — the audio notifications work for all created conversations.
* **Email Notifications** — enable desired event options for your email notifications.
* **Push Notifications** — enable desired event options for your push notifications.

The settings will be applied automatically.

Users with the `Administrator` role can configure the notification sound in the [Account Settings](/live-agent/settings/account-settings).

### Configure Notifications at the Account Level

Users with the `Administrator` role can configure email and push notifications at the Account level.

This configuration applies to all user profiles within your account (organization) and overrides the [Profile Settings](#configure-notifications-at-the-user-profile-level).

To configure email and push notifications at the Account level, follow these steps:

1. On the left side of the Live Agent screen, select **Settings**.
2. In the settings column, select **Account Settings**.
3. On the **Account Settings** page, activate the following notifications:
   * **Email Notifications** — activate **Enable overriding email notification preferences** and enable desired event options for your email notifications.
   * **Push Notifications** — activate **Enable overriding push notification preferences** and enable desired event options for your push notifications.
4. In the upper-right corner, click **Update Settings** to apply the settings.

## View Unread Notifications

To view your unread notifications in Live Agent, do the following:

1. In the bottom-left corner of the Live Agent screen, click <img alt="notifications" />.
2. To see all the notifications in a window, click **View all Notifications**.
3. Click a notification item to see the corresponding conversation.

## More Information

* [Profile Settings](/live-agent/profile-settings)
* [Account Settings](/live-agent/settings/account-settings)


# About Live Agent
Source: https://docs.cognigy.com/live-agent/overview

Learn about Cognigy Live Agent, a contact center solution that works with Cognigy to route conversations to human agents based on expertise and availability.

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

*Cognigy Live Agent* is a contact center solution that specifically works with the Cognigy platform.
Within Live Agent, human agents handle incoming conversations that are escalated by AI Agents based on the end user's preference.
This solution allows users to connect with live customer support representatives when they wish to address specific or critical issues.
The Live Agent system is built around role-based assignments for human agents, supervisors, and admins, as well as conversation routing to ensure that the right agents are assigned to each conversation based on their expertise and availability.
Cognigy Live Agent helps businesses deliver exceptional customer support experiences while optimizing operational efficiency and scalability. It empowers businesses to better meet the needs of their customers and drive satisfaction and loyalty in the process.

<Frame>
  <img alt="Live Agent Overview" />
</Frame>


# Profile Settings
Source: https://docs.cognigy.com/live-agent/profile-settings

Manage your Live Agent Profile settings, including notification preferences, Agent Copilot options, message hotkeys, and personal API access token.

<a href="/release-notes/2026.17"><Badge>Updated in 2026.17</Badge></a>

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

*Profile Settings* are a set of customizable options and configurations available to individual users within the Live Agent system. These settings include the customization of profile appearance, notification preferences, activation of Agent Copilot, configuration of hotkeys for sending messages, and access to API functionality.

To open the **Profile Settings** in Live Agent, do the following:

1. In the bottom-left corner of the Live Agent interface, click **Profile Settings**.
2. In the dropdown list, select **Profile Settings**. The **Profile Settings** page will be displayed.

The following settings are available within the user profile:

* [User Profile Appearance](#user-profile-appearance)
* [Profile Language](#user-profile-appearance)
* [Notifications](#notifications)
* [Agent Copilot](#agent-copilot)
* [Hotkey to send messages](#hotkey-to-send-messages)

Additionally, you can copy the [API Token](#api-access-token) to access the [Live Agent API](https://liveagent-trial.cognigy.ai/openapi).

## User Profile Appearance

To personalize your profile settings, you can edit your profile image and the name displayed in conversations.

| Setting       | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  |
| ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Profile image | The image will be displayed in the conversations, agent lists, and user profile menu.<br /> For [Webchat v3](/webchat/v3/configuration#webchat-layout), when the Profile Image is not available, the system will use the value from the **Human Agent Avatar Logo URL** parameter from the **Webchat Layout** section if it is provided. Otherwise, a default human agent icon will be displayed.                                                                                            |
| Display name  | This name will be displayed in the conversations and is used as the sender name when human agents send conversation transcripts in emails.<br /> For [Webchat v3](/webchat/v3/configuration#webchat-layout), when the Display name is not available, the system will use the value from the **Human Agent Avatar Name** parameter from the **Webchat Layout** section if it is provided. Otherwise, the value of the **Name** parameter in the **Webchat Layout** section will be displayed. |

The other user data, such as email address, full name, and password, needs to be updated in the Cognigy account. The changes will automatically apply to Live Agent.

## Profile Language

Choose your preferred language for the Live Agent interface from the dropdown menu.
By default, your [account language](/live-agent/settings/account-settings) is applied, but you can set a different language for your profile. For example, if your account is set to English but you prefer Japanese, you can select Japanese (ja) to make navigation easier.
Setting a profile language also translates system messages, push notifications, and email notifications.

## Notifications

In Live Agent, you can configure email and push notifications, and audio alerts for certain types of actions that trigger notifications.

The email and push notification settings in the [Account Settings](/live-agent/notifications#configure-notifications-at-the-account-level) will take precedence over the Profile Settings.
If the administrator has not yet changed the push and email notification settings, you can change them in the Profile Settings.

To configure notifications, refer to [Notifications](/live-agent/notifications#configure-notifications-at-the-user-profile-level).

## Agent Copilot

Cognigy Agent Copilot offers a variety of advanced features that empower human agents to provide faster and more accurate customer support. The workspace is fully customizable with widgets that display the most relevant information for any chat and voice use case. When embedded into your contact center, it enhances the overall efficiency of your contact center.

In the **Profile Settings**, a user can activate Agent Copilot if [it has been installed previously](/agent-copilot/getting-started/chat).

If an administrator has already preset the **Copilot Workspace** setting in the **Account Settings**,
this setting will take precedence over the user profile settings.
In this case, a user is not able to change the **Copilot Workspace** settings.

For more information, read the [Agent Copilot](/live-agent/assistants/agent-copilot) and [Agent Copilot Whisper](/live-agent/assistants/agent-copilot-whisper) documentation.

## Hotkey to send messages

To send messages in the [conversation chat](/live-agent/conversation/overview), you can select one of the following hotkey options:

* **None** — send messages by clicking the **Send** button. The <kbd>enter</kbd> key will be used for line breaks.
* **Enter (↵)** — send messages by pressing the <kbd>enter</kbd> key instead of clicking the **Send** button. The <kbd>shift+enter</kbd> key combination will be used for line breaks.
* **Cmd/Ctrl + Enter (⌘ + ↵)** — send messages by pressing <kbd>cmd+enter</kbd> or <kbd>ctrl+enter</kbd> instead of clicking the **Send** button. This option is selected by default. The <kbd>enter</kbd> key will be used for line breaks.

## API Access Token

The displayed Access Token serves as the authentication and authorization key for the [Live Agent API](https://liveagent-trial.cognigy.ai/openapi). It contains information specific to a user, and the actions it enables are restricted by the user's [assigned role](/live-agent/roles).

## More Information

* [Account Settings](/live-agent/settings/account-settings)
* [User Roles](/live-agent/roles)
* [Agent Copilot](/live-agent/assistants/agent-copilot)


# Agents Report
Source: https://docs.cognigy.com/live-agent/reports/admin/agents

Analyze human agent performance in Live Agent with the Agents Report, including conversation counts, response time, resolution time, and resolved requests.

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

This report shows agent-specific metrics, such as the number of conversations, outgoing messages, average response time, average resolution time, and the number of successfully resolved user requests, to assess and optimize an agent's performance in handling conversations.

By analyzing these metrics over time, you can identify trends and make data-driven decisions to improve the efficiency and quality of your customer support or communication processes. For example, if you notice an increase in average response time, you may need to allocate more resources or provide additional training to your agents. If the number of resolved requests is declining, it may signal a need for process improvements or better tools.

To learn more about metrics, refer to [Report Metrics](/live-agent/reports/admin/overview#report-metrics).

<Frame>
  <img alt="Live Agent Agents report" />
</Frame>

## View Reports

To view a report, do the following:

1. In the Live Agent interface, go to **Reports > Agents**.
2. In the upper-left corner of the report page, select an agent and time period from the lists.

The chart will be updated with the corresponding data related to the selected filters.

## Download Reports

To download a report as a CSV file, click **Download Report** in the upper-right corner of the report page.

## More information

* [Admin reports](/live-agent/reports/admin/overview)
* [Reports](/live-agent/reports/overview)


# Conversations Report
Source: https://docs.cognigy.com/live-agent/reports/admin/conversations

Track conversation trends in your Live Agent account with the Conversations Report, including message counts, response time, and resolution time.

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

This report shows detailed conversation metrics within an account.

The report lets you easily track trends,
such as increases or decreases in the number of conversations, both incoming and outgoing messages,
the average response time from agents, the average resolution time for customer issues,
and the number of successfully resolved customer requests.

To learn more about metrics, refer to [Report Metrics](/live-agent/reports/admin/overview#report-metrics).

<Frame>
  <img alt="Live Agent conversations report" />
</Frame>

You can filter data by date. You can select the time period for measurement from the upper dropdown list.

For example, if **last 7 days** is selected, which is the default, the calculation period extends backward from now for seven days.

## View Reports

To view a report, do the following:

1. In the Live Agent interface, go to **Reports > Conversations**.
2. In the upper-left corner of the Report page, select a time period from the list.

The chart will be updated with the corresponding data related to the selected filter.

## Download Reports

To download a report as a CSV file, click **Download Report** in the upper-right corner of the report page.

## More information

* [Admin reports](/live-agent/reports/admin/overview)
* [Reports](/live-agent/reports/overview)


# Inboxes Report
Source: https://docs.cognigy.com/live-agent/reports/admin/inboxes

Optimize communication channels with the Live Agent Inboxes Report, showing conversation trends, response time, and resolution metrics per inbox.

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

This report shows detailed Inbox metrics within an account.

The report lets you easily track trends,
such as increases or decreases in the number of conversations, both incoming and outgoing messages,
the average response time from agents, the average resolution time for customer issues,
and the number of successfully resolved customer requests.

The Inboxes Report helps optimize communication channels, improve customer satisfaction, and enhance operational efficiency in your organization.

To learn more about metrics, refer to [Report Metrics](/live-agent/reports/admin/overview#report-metrics).

<Frame>
  <img alt="Live Agent Inbox report" />
</Frame>

## View Reports

To view a report, do the following:

1. Open the Live Agent interface.
2. Go to **Reports > Inboxes**.
3. In the upper-left corner of the report page, select a time period from the list.

The chart will be updated with the corresponding data related to the selected filter.

## Download Reports

To download a report as a CSV file, click **Download Report** in the upper-right corner of the report page.

## More information

* [Admin reports](/live-agent/reports/admin/overview)
* [Reports](/live-agent/reports/overview)


# Labels Report
Source: https://docs.cognigy.com/live-agent/reports/admin/labels

Monitor label usage in Live Agent with the Labels Report, showing conversation counts, response time, and resolution time for each conversation tag.

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

This report shows label-specific metrics, including the number of conversations, incoming and outgoing messages, average response time, average resolution time, and the count of successfully resolved user requests for each label.

With this report, you can monitor how many conversations or requests are linked to particular labels. This allows you to recognize the most frequent and critical issues or topics. For instance, you can check if there's a rise in the number of conversations marked as `billing issues` and then take necessary steps accordingly.

To learn more about metrics, refer to [Report Metrics](/live-agent/reports/admin/overview#report-metrics).

<Frame>
  <img alt="Live Agent Labels report" />
</Frame>

## View Reports

To view a report, do the following:

1. In the Live Agent interface, go to **Reports > Labels**.
2. In the upper-left corner of the report page, select a label and time period from the list.

The chart will be updated with the corresponding data related to the selected filter.

## Download Reports

To download a report as a CSV file, click **Download Report** in the upper-right corner of the report page.

## More information

* [Admin reports](/live-agent/reports/admin/overview)
* [Reports](/live-agent/reports/overview)


# Administrator Reports
Source: https://docs.cognigy.com/live-agent/reports/admin/overview

Access Live Agent administrator reports covering agents, teams, inboxes, labels, and conversations, with combined indicator and bar chart analytics.

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

Administrators can monitor the account's performance in terms of [agents](/live-agent/reports/admin/agents),
[labels](/live-agent/reports/admin/labels), [Inboxes](/live-agent/reports/admin/inboxes), [teams](/live-agent/reports/admin/teams),
and also track [conversation](/live-agent/reports/admin/conversations) metrics over time periods.

These metrics are presented in combined charts, including an indicator and a bar chart. This combined format allows for both a quick overview and a detailed, comprehensive understanding of the metrics. The indicator provides high-level information by displaying total and average values, while the detailed data is presented as bar charts.

## Combined Charts

The combined chart includes the following charts:

* **Indicator.** Represents a total or an average number. When there is a significant increase or decrease in conversation handling based on the last entry, it will be displayed next to the indicator value with a percentage figure related to the calculated average value of the selected time period. Increasing trends will be indicated with a plus and a green colored figure, and decreasing trends will be indicated with a minus in red color. For conversations, an increasing number means you have got more customer messages. For solving problems, increasing number, for example, `+25%`, means you got better at solving customer issues or got easier conversations.
* **Bar chart.** The horizontal axis typically represents dates or time periods, illustrating how the metrics change over time, while the vertical axis represents various performance metrics such as the number of messages or conversations.

When hovering the mouse cursor over the bar chart, details of metrics that happened can be seen.

You can display the values per day, week, month or year, depending on the selected filters.

For instance, in the image below, there is a combined chart for the QA team to analyze performance data.

<Frame>
  <img alt="Live Agent Team report" />
</Frame>

## Report Metrics

Each report includes charts with the following metrics:

| Metric              | Description                                                                                                                                                                                                          |
| ------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Conversations       | The number of conversations that occurred during the selected period.                                                                                                                                                |
| Incoming Messages   | The number of incoming messages sent by end users that occurred during the selected period.<br />This metric is not applicable to the Agents report.                                                                 |
| Outgoing Messages   | The number of outgoing messages sent by human agents that occurred during the selected period.                                                                                                                       |
| First Response Time | The time period for the first reply of a human agent to the end user across all conversations.                                                                                                                       |
| Resolution Time     | The amount of time it takes to resolve a conversation. The data for this metric is collected from all conversations, providing an overall picture of how long it takes to handle and complete customer interactions. |
| Resolution Count    | The number of conversations that are resolved by human agents in the selected period.                                                                                                                                |

## More Information

* [Reports](/live-agent/reports/overview)
* [Supervisor report](/live-agent/reports/supervisor-overview)


# Teams Report
Source: https://docs.cognigy.com/live-agent/reports/admin/teams

Evaluate team performance in Live Agent with the Teams Report, tracking conversations, average response time, and resolution time for each support team.

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

This report shows team-specific metrics, such as the number of conversations, outgoing messages,
average response time, average resolution time, and the number of successfully resolved user requests for each team.

The Teams report provides a comprehensive view of team performance and is essential for identifying areas that may need improvement. For instance, if the average response time is too high, it could mean that the team needs to improve its responsiveness. If the resolution time is long, it might indicate the need for more effective problem-solving strategies.

To learn more about metrics, refer to [Report Metrics](/live-agent/reports/admin/overview#report-metrics).

<Frame>
  <img alt="Live Agent Team report" />
</Frame>

## View Reports

To view a report, do the following:

1. In the Live Agent interface, go to **Reports > Teams**.
2. In the upper-left corner of the report page, select a team and time period from the list.

The chart will be updated with the corresponding data related to the selected filter.

## Download Reports

To download a report as a CSV file, click **Download Report** in the upper-right corner of the report page.

## More information

* [Admin reports](/live-agent/reports/admin/overview)
* [Reports](/live-agent/reports/overview)


# Overview
Source: https://docs.cognigy.com/live-agent/reports/overview

Live Agent analytics dashboards for supervisors and administrators. Monitor agent performance, team metrics, and real-time customer support data.

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

*Reports* are analytic dashboards for supervisors and managers in Live Agent. These reports offer real-time insights into the performance and activities of agents and teams, allowing for the monitoring and evaluation of interactions with customers, facilitating quick adjustments, and driving improvements in customer support processes.

Only users with `Administrator` or `Supervisor` roles have access to the reports.

Depending on your role, the following reports will be available:

* Supervisor:
  * [Overview](/live-agent/reports/supervisor-overview)
* Administrator:
  * [Conversations](/live-agent/reports/admin/conversations)
  * [Agents](/live-agent/reports/admin/agents)
  * [Labels](/live-agent/reports/admin/labels)
  * [Inboxes](/live-agent/reports/admin/inboxes)
  * [Teams](/live-agent/reports/admin/teams)

## View Reports

To view the reports, select **Reports** in the left-side menu of the Live Agent interface.

## More information

* [Insights: Live Agent report](/insights/reports/live-agent)
* [Live Agent Glossary](/live-agent/glossary)


# Supervisor Overview Report
Source: https://docs.cognigy.com/live-agent/reports/supervisor-overview

Monitor real-time conversation activity with the Live Agent Supervisor Overview Report, tracking agent status and key metrics for the past 24 hours.

<a href="/release-notes/2026.1"><Badge>Updated in 2026.1</Badge></a>

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

<Note>
  This report is only visible with the [Supervisor Role](/live-agent/roles#supervisor).
</Note>

This report offers an overview of all conversations related to a specific inbox,
including the number of conversations received within the last 24 hours.
All displayed data is relevant to this 24-hour time period.
Additionally, you will find an overview of Agent Status and conversation handling metrics.
At the bottom of the report, there is a list of conversations that indicate the agents' ongoing work.

The summary of key metrics is represented in the following charts:

* [Indicators](#indicators)
* [Line Charts](#line-charts)
* [Tables](#tables)

<Frame>
  <img alt="Live Agent Supervisor overview" />
</Frame>

## Indicators

### Open Conversations

Indicates the number of conversations that have the following statuses: `Open`, `Queued`, `Unassigned`, or `Escalated`.
The number of `Escalated` conversations changes only when the [Escalate Conversations to Supervisor](/live-agent/settings/inboxes#escalate-conversations-to-supervisors) feature is activated at the Inbox level and an agent clicks the **↑ Escalate to Supervisor** button in the [Conversation Actions](/live-agent/conversation/assign-conversations#assign-supervisors) section of the conversation chat.

### Today

Indicates the number of conversations that were handled on the current day.

### Agent Status

Indicates the number of agents who are currently `Online`, `Busy`, `Away`, or `Offline`.

### Conversation handling

Conversation handling indicates the following real-time metrics:

| Metrics                   | Description                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      |
| ------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| First Response Time (Avg) | The average time it takes for an agent to respond to a customer in a conversation.                                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
| Resolution Time (Avg)     | The average time required to resolve conversations with a customer.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              |
| Queue Time (Avg)          | The average time that an incoming conversation waits for assignment to a human agent.<br />This value is measurable in two scenarios:<br />- When [auto-assigning conversations](/live-agent/conversation/conversation-routing/automatic-mode#automatic-assignment) is turned off for the Inbox.<br />- When the [conversation limit per agent](/live-agent/settings/account-settings#conversation-limit-per-agent) is activated, and all agents in an Inbox with auto-assignment activated are at full capacity and unable to take on additional conversations. |

## Line Charts

### Conversation Metrics

Indicates conversation details in a 24-hour time grid, visualized with colored lines:

| Line graph                  | Description                                                     |
| --------------------------- | --------------------------------------------------------------- |
| Incoming Conversations      | Number of conversations that were created in the last 24 hours. |
| Conversation Duration (min) | Minutes it took to solve a conversation from start to finish.   |
| Queued Conversations        | Number of conversations that are not assigned to a human agent. |

## Tables

### Conversations by Agents

Provides information about the status and performance of agents:

| Title                   | Description                                                             |
| ----------------------- | ----------------------------------------------------------------------- |
| STATUS                  | Indicates the status of the Agent: `Away`, `Busy`, `Online`, `Offline`. |
| OPEN                    | Number of active conversations pending.                                 |
| QUEUED                  | Indicates the number of assigned conversations waiting.                 |
| AVG FIRST RESPONSE TIME | Indicates the average value when an Agent responded first.              |
| AVG RESOLUTION TIME     | Indicates the average value when an Agent resolved a conversation.      |
| RESOLVED                | Indicates how many conversations were resolved by Agents.               |

You can modify an agent's status when they are online by clicking an agent icon in the **Agent** column of the table.
You can set their status as `Busy` or `Away`.
However, if the agent is offline, any changes you attempt to make to their status will not be applied.

Additionally, you can filter the table data by agent status.

<Frame>
  <img alt="Live Agent Supervisor overview 2" />
</Frame>

## View Reports

To view a report, do the following:

1. In the Live Agent interface, go to **Reports > Overview**.
2. At the top of the report, from the **Inbox** list, select an Inbox you want to inspect.

The report will be updated with the corresponding data related to the selected Inbox.

## More Information

* [Reports](/live-agent/reports/overview)
* [Admin reports](/live-agent/reports/admin/overview)


# Roles
Source: https://docs.cognigy.com/live-agent/roles

Learn about Live Agent user roles, including their permissions, responsibilities, and access levels in the system.

In Live Agent, human agents can have three distinct user roles:

* [Agent](#agents)
* [Supervisor](#supervisor)
* [Administrator](#administrator)

These roles define the permissions and responsibilities of individuals using the Live Agent system.

## User Roles

### Agent

Agents in this role can access inboxes and conversations, assign conversations to other agents or themselves, and resolve conversations. They can access only the inboxes they are part of, participate in conversations within those inboxes, edit their user profile settings, and add personal-level canned responses. Conversation assignment is limited to inbox members.

### Supervisor

Supervisors have unique access to the Supervisor Dashboard to monitor conversation metrics and can also view the Inbox assigned to them.

### Administrator

Administrators have all permissions of a normal agent and access to all Cognigy Live Agent features enabled for your account, including settings. Administrators don't have permission to:

* View the supervisor dashboard. This dashboard is exclusive to the supervisor role.
* Create and delete users. You can only do this in the [Cognigy Admin Center](/ai/administer/access/admin-center/access-control).

## Permissions

The following tables show single role permissions of Agents, Supervisors, and Administrators.

### Agents

| Actions                                              | Agent | Supervisor | Administrator |
| ---------------------------------------------------- | ----- | ---------- | ------------- |
| Create an Agent<sup><a href="#footnote1">1</a></sup> | -     | -          | +             |
| Edit an Agent<sup><a href="#footnote1">1</a></sup>   | -     | -          | +             |
| Delete an Agent<sup><a href="#footnote1">1</a></sup> | -     | -          | +             |
| Assign an Agent                                      | +     | +          | +             |

### Teams

| Actions                             | Agent | Supervisor | Administrator |
| ----------------------------------- | ----- | ---------- | ------------- |
| Create a team                       | -     | -          | +             |
| Assign a team within a conversation | +     | +          | +             |
| Edit a team                         | -     | -          | +             |
| Delete a team                       | -     | -          | +             |
| Allow Auto Assign for a team        | -     | -          | +             |

### Inboxes

| Actions                                                                           | Agent | Supervisor | Administrator |
| --------------------------------------------------------------------------------- | ----- | ---------- | ------------- |
| Create an Inbox                                                                   | -     | -          | +             |
| Edit an Inbox                                                                     | -     | -          | +             |
| Delete an Inbox                                                                   | -     | -          | +             |
| Move Conversations from one Inbox to another<sup><a href="#footnote4">4</a></sup> | +     | +          | +             |
| Create Agent Copilot Bot                                                          | -     | -          | +             |
| Edit Agent Copilot Bot                                                            | -     | -          | +             |
| Delete Agent Copilot Bot                                                          | -     | -          | +             |
| Add Agents to the Inbox                                                           | -     | -          | +             |
| Remove Agents from the Inbox                                                      | -     | -          | +             |
| Configure Conversation Assignment Settings                                        | -     | -          | +             |
| View assigned Inboxes                                                             | +     | +          | +             |
| View all Inboxes                                                                  | -     | -          | +             |

### Conversations

| Actions                              | Agent | Supervisor | Administrator |
| ------------------------------------ | ----- | ---------- | ------------- |
| Resolve Conversations                | +     | +          | +             |
| Clear Conversation History           | -     | +          | +             |
| Mute Conversation                    | +     | +          | +             |
| Send Transcript                      | +     | +          | +             |
| Edit User Contact Profile            | -     | +          | +             |
| Assign an Agent                      | +     | +          | +             |
| Assign a Team                        | +     | +          | +             |
| Select Priority                      | +     | +          | +             |
| Filter Conversations                 | +     | +          | +             |
| Sort Conversations                   | +     | +          | +             |
| Search for messages in conversations | +     | +          | +             |

### Attachments

| Actions                                                                        | Agent | Supervisor | Administrator |
| ------------------------------------------------------------------------------ | ----- | ---------- | ------------- |
| Send Attachments within a conversation                                         | +     | +          | +             |
| Request Attachments within a conversation<sup><a href="#footnote2">2</a></sup> | +     | +          | +             |
| View Attachments                                                               | -     | -          | +             |
| Sort Attachments                                                               | -     | -          | +             |
| Filter Attachments                                                             | -     | -          | +             |
| Delete Attachments                                                             | -     | -          | +             |
| Download Attachments                                                           | -     | -          | +             |

### Audit Logs

| Actions         | Agent | Supervisor | Administrator |
| --------------- | ----- | ---------- | ------------- |
| Read Audit Logs | -     | -          | +             |

### Automation Rules

| Actions                     | Agent | Supervisor | Administrator |
| --------------------------- | ----- | ---------- | ------------- |
| Add Automation Rules        | -     | -          | +             |
| Edit Automation Rules       | -     | -          | +             |
| Clone Automation Rules      | -     | -          | +             |
| Delete Automation Rules     | -     | -          | +             |
| Activate Automation Rules   | -     | -          | +             |
| Deactivate Automation Rules | -     | -          | +             |

### Labels

| Actions                       | Agent | Supervisor | Administrator |
| ----------------------------- | ----- | ---------- | ------------- |
| Create a Label                | -     | +          | +             |
| Edit a Label                  | -     | +          | +             |
| Delete a Label                | -     | +          | +             |
| Add a Label to a conversation | +     | +          | +             |

### Skills

| Actions                | Agent | Supervisor | Administrator |
| ---------------------- | ----- | ---------- | ------------- |
| Add Skills             | -     | +          | +             |
| Edit Skills            | -     | +          | +             |
| Delete Skills          | -     | +          | +             |
| Add Skills to an Agent | -     | -          | +             |

### Canned Responses

| Actions                                                       | Agent | Supervisor | Administrator |
| ------------------------------------------------------------- | ----- | ---------- | ------------- |
| Create a Canned Response<sup><a href="#footnote3">3</a></sup> | +     | +          | +             |
| Use a Canned Response within a conversation                   | +     | +          | +             |
| Edit a Canned Response<sup><a href="#footnote3">3</a></sup>   | +     | +          | +             |
| Delete a Canned Response<sup><a href="#footnote3">3</a></sup> | +     | +          | +             |
| Import Canned Responses<sup><a href="#footnote3">3</a></sup>  | +     | +          | +             |

### Reports

| Actions                  | Agent | Supervisor | Administrator |
| ------------------------ | ----- | ---------- | ------------- |
| View the Overview Report | -     | +          | -             |
| View the Inbox Report    | -     | +          | +             |
| View the Agent Report    | -     | +          | +             |
| View the Label Report    | -     | +          | +             |
| View the Team Report     | -     | +          | +             |
| Download Reports         | -     | +          | +             |

### Assistants

| Actions                          | Agent | Supervisor | Administrator |
| -------------------------------- | ----- | ---------- | ------------- |
| **Agent Copilot**                |       |            |               |
| Activate Agent Copilot workspace | +     | +          | +             |
| **Agent Copilot Bot**            |       |            |               |
| Edit Agent Copilot Bot Settings  | +     | +          | +             |
| Reset Session                    | +     | +          | +             |

### Account and Profile Settings

| Actions                    | Agent | Supervisor | Administrator |
| -------------------------- | ----- | ---------- | ------------- |
| Configure Account Settings | -     | -          | +             |
| Configure Profile Settings | +     | +          | +             |

## Assign Roles

<a href="/release-notes/4.44"><Badge>Updated in 4.44</Badge></a>

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

You can assign roles to human agents.

To create an Agent with certain access rights, do the following steps:

1. Create an Agent from the [Cognigy Access Control](/ai/administer/access/admin-center/access-control).
2. Assign one of the following roles to the Agent:
   * `liveAgentAgent`
   * `liveAgentSupervisor`
   * `liveAgentAdmin` <br />If you have an Admin role in Cognigy, you do not need to add the Administrator role for your user because the Cognigy role already includes the right to administer Live Agent.
3. *(Optional)* Add any additional roles the Agent should have. For example, if you already have a `liveAgentAgent` role, you can add a `liveAgentSupervisor` role to allow access to the Overview report.
4. View the created user in Live Agent. Go to **Settings > Agents** in the Live Agent menu.

To see the roles applied to the Agent, the specific Agent must log in once to have the roles applied.

## Troubleshooting

### Assigned Roles are not Applied

If a human agent already has a specific role in Live Agent, assigned roles may not be applied.
For example, if you want to assign a user with `Administrator` rights the role of `Supervisor`,
the new role will only take effect once the user logs out of Cognigy and Live Agent.
After logging out and logging back in, the new roles should be applied successfully.

## More Information

* [Conversation Workflow](./conversation/overview)
* [Account Settings](./settings/account-settings)

***

<span>
  <sup>1</sup> This permission is granted on the Cognigy side.
</span>

<br />

<span>
  <sup>2</sup> By default, human agents cannot request files from users. To enable this feature, you need to install the [File Upload](https://github.com/Cognigy/WebchatPlugins/tree/master/plugins/file-upload) plugin in the [Webchat Endpoint](/webchat/v2/configuration).
</span>

<br />

<span>
  <sup>3</sup> Users with the Agent role can create canned responses for their own usage, edit, and delete canned responses that they created.
</span>

<br />

<span>
  <sup>4</sup> Users with the `Agent` role can assign conversations to the inboxes they have access to.
</span>

<br />


# Account Settings
Source: https://docs.cognigy.com/live-agent/settings/account-settings

Configure Live Agent account settings including general options, Agent Copilot, notifications, and conversation workflow. Only available to Administrators.

<a href="/release-notes/2026.17"><Badge>Updated in 2026.17</Badge></a>

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

*Account Settings* control the administration and optimization of Live Agent functionality within an organization.

Only users having the `Administrator` role have access to the **Account Settings**.

There are the following settings available within the account:

* [General Settings](#general-settings)
* [Agent Copilot](#agent-copilot)
* [Notifications](#notifications)

Additionally, you can copy an [Account ID](#account-id) and check a [current version](#version) of your Live Agent installation.

## General Settings

You can change the general settings for an account, including its name, language, and conversation workflow.

### Account Name

The name serves as the identifier and authentication for legal access to your organization's account. It will always be displayed to indicate the registered account.

### Site Language

The language represents the preferred site language for your organization's account.
You can select one of the available options in a dropdown list by clicking on the language field. The default language is English (en).
Setting a site language also translates system messages, push notifications, and email notifications.

### Conversation Limit per Agent

The conversation limit per human agent is the maximum number of conversations that a human agent can handle simultaneously. If a human agent reaches the limit in any Inbox, new conversations will remain unassigned, awaiting pickup by other human agents when their capacity utilization decreases. The conversation limit applies to the total number of conversations across all Inboxes.

For example, the human agent is assigned to Inboxes A and B, and the conversation limit is set to 2.
In this case, the human agent can handle two conversations simultaneously, regardless of the Inbox they belong to.

To set up a conversation limit per Agent, follow these steps:

1. Go to **Settings > Account Settings**.
2. On the **Account Settings** page, go to the **General Settings** section.
3. Activate the **Set conversation limit for agents** option.
4. In the **Max number of conversations per agent** field, specify the desired limit for agents.
5. In the upper-right corner, click **Update Settings**.

The limit will be applied to the account.

If you want to remove the limit, activate the **No conversation limit for agents** and click **Update Settings**.

### Auto Set Away

The **Auto Set Away** option is used to automatically set the human agent as away when they are not active after the amount of time in minutes specified.
This means the agent is not performing any action in the browser window while it is opened.
When the agent performs an action, it will be set back to online state automatically.

* Click **Disable Auto Set Away**, in case you do not want to use this option.
* Click **Enable Auto Set Away for agents availability** and enter the number of minutes in a separate field. After the specified time, the agent will then automatically switch to the `Away` status.

<Note>
  This setting does not work if the Live Agent interface is closed. In this case, the user is set to offline automatically when the browser window is closed.
</Note>

### TTL for Conversations

The **TTL for Conversations** option lets you automatically delete the conversations in the account after a set period.

To set a TTL for conversations, follow these steps:

1. Click **Enable TTL for conversations**.
2. Enter the TTL in the **Number of days after conversation is deleted** field.

By default, the **Delete only Resolved and Abandoned conversations** option is activated. Deselect this option to delete all conversations.

To deactivate the TTL for conversations, click **Disable TTL for conversations**.

### TTL for Attachments

The **TTL for Attachments** option lets you automatically delete the attachments in messages of all conversations in the account after a set period.

To set a TTL for attachments, follow these steps:

1. Click **Enable TTL for attachments**.
2. Enter the TTL in the **Number of days after attachment is deleted** field.

To deactivate the TTL for attachments, click **Disable TTL for attachments**.

### Auto Resolve Conversations

#### Auto Resolve Duration

The Auto Resolve Duration option automatically resolves the conversation after a certain number of minutes. After that time, you can assume that the customer's problem is solved, and the conversation can be closed.

* Click **Do not automatically resolve conversations**, in case you do not want to use this option.
* Click **Automatically resolve conversations if there is no activity** and enter the number of minutes in a separate field. After this period, the conversation should auto-resolve if there is no activity.

#### Auto Resolve Action

The **Auto Resolve Action** option resolves or abandons a conversation automatically after the set **Auto Resolve Duration** has expired.

The system starts a timer based on the last activity in the conversation.
This activity can be any interaction, such as adding or removing a label, changing the priority, assigning or unassigning a human agent.
If there's no activity within the set time, the system marks the conversation as ready to be closed.

You can select one of the following options:

* Select **Resolve** to close a conversation if the specified time has passed without any activity and the last message was sent by either a human agent or an end user.
* Select **Abandon** to close a conversation if the specified time has passed without any activity and the last message was sent by a human agent. If an end user replied last, the conversation won't be marked as abandoned.

### Automatically Delete Resolved Conversations

When conversations are resolved, you can delete them.
To automate this process, activate the **Automatically delete conversations when they are resolved** setting.
When you activate this setting, Live Agent automatically deletes conversations marked as resolved.
Note that this action will also delete all conversations that have already been resolved.

### Enable Agent Masking

Activating the **Enable Agent masking** setting masks all agent personal information,
such as name, email, and profile images.
For example, the agent name "Max Mustermann" would turn into "Agent x4gH5tZ".
Only the agent himself can see the clear name.
No one in your organization and of course,
no customer can connect the arbitrary masked name with a real human agent's name.

### Auto-Assign Conversations to Busy Agents

<a href="/release-notes/4.53"><Badge>Updated in 4.53</Badge></a>

By default, you can manually assign conversations to any human agents regardless of their status
(online, offline, away, busy).
When this setting is activated, the system can automatically assign incoming conversations to an agent,
even if the agent is busy and engaged in another conversation.
By default, only online agents are considered.

## Agent Copilot

Cognigy Agent Copilot offers the most comprehensive functions compared to the other assistants, enabling human agents to provide faster and more accurate customer support. The workspace is fully customizable with widgets that display the most relevant information for any chat and voice use case. When embedded into your contact center, it enhances the overall efficiency of your contact center.

In the **Account Settings**, administrators can activate Agent Copilot if [it has been installed previously](/agent-copilot/getting-started/chat).

If an administrator has already preset the **Agent Copilot Workspace** setting in the **Account Settings**,
this setting will take precedence over the user profile settings.

Before you can enable **Agent Copilot Workspace** for all human agents,
activate the **Enable overriding Agent Copilot Workspace preferences** setting.
The configuration set here applies as a preset to all human agents of your organization account
and overrides the [Profile Settings](/live-agent/profile-settings#notifications).

## Notifications

<a href="/release-notes/4.45"><Badge>Updated in 4.45</Badge></a>

You can configure email and push notifications as well as the notification sound of [audio notifications](/live-agent/notifications#configure-notifications-at-the-user-profile-level).

### Email and Push Notifications

<Note>
  * Push notifications must be enabled in Live Agent and in your browser. Otherwise, they are not available.
  * Push notifications are not supported by the Microsoft Edge browser.
</Note>

You can activate push and email notifications in the **Push Notifications** and **Email Notifications** sections for the following event options:

* A conversation is created.
* A conversation is assigned to you.
* You are mentioned in a conversation.
* You receive a message in an assigned conversation.

To unlock the event options, activate the **Enable overriding email notification preferences** setting. The event options you set apply to all human agents of your Live Agent organization's account and override individual agents' [Profile settings](/live-agent/profile-settings#notifications).

### Notification Sound

You can add a notification sound to the push notifications you configured. To configure the notification sound, follow these steps:

1. Go to **Settings > Account Settings**.
2. From the **Sound** list in the **Notification Sound** section, select one of the following options:
   * **Default** — a short sound is triggered when you send or receive a message.
   * **Custom** — the **Sound URL** field appears. Enter the URL for the sound file you want to set as audio notification. For example, `https://commondatastorage.googleapis.com/codeskulptor-assets/jump.ogg`. The notification sound is triggered when a push notification is triggered. Use a notification sound between 3 and 5 seconds, as longer notification sounds may overlap if you send or receive consecutive messages rapidly. When using a custom sound URL, make sure that the server hosting the sound file has the appropriate cross-origin resource sharing (CORS) headers configured.

For more information, read the [Notifications](/live-agent/notifications) documentation.

## Account ID

The Account ID, which is indicated here in the Account Settings, is the unique identifier of your account.
It is used in the Live Agent URLs to identify which account is accessed.
This ID is necessary when using the Live Agent API.

To find the Account ID alternatively in the Live Agent URL of your Live Agent **Inbox**, follow these steps:

1. Open your Live Agent Account and go to **Settings > Inboxes**.
2. In the Inboxes list, find the Inbox of your project, and click the **Settings** icon next to the Inbox.
3. The Inbox settings will be displayed, and you will see the address line of your browser as follows: `https://liveagent-<your-environment>/app/accounts/<account-id>/settings/inboxes/<inbox-id>`. For example, in the URL `https://liveagent-dev.cognigy.ai/app/accounts/2/settings/inboxes/32852`, the Account ID is 2.

<Frame>
  <img alt="Live Agent browser URL showing the Account ID and Inbox ID" />
</Frame>

For example, once you have found the account ID,
you can enter the Account ID in the [Handover Settings](/live-agent/getting-started/live-agent-setup/live-agent-setup-handover-endpoint#handover-to-manually-configured-live-agent) of your Endpoint in Cognigy,
to enable the handover of a customer conversation with your AI Agent to a human agent for support.

## Version

The indicated version is the current version of the Cognigy Live Agent interface. To check the features and fixes for each version, head over to [Release Notes](/release-notes/index).

## More information

* [User Roles](/live-agent/roles)
* [Profile Settings](/live-agent/profile-settings)
* [Agent Copilot](/live-agent/assistants/agent-copilot)
* [Notifications](/live-agent/notifications)


# Agents
Source: https://docs.cognigy.com/live-agent/settings/agents

Manage Live Agent support team members, configure agent statuses such as online, busy, and away, and understand administrator versus agent roles.

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

An Agent is a member of your customer support team. Agents will be able to view and reply to messages from your users.

On the Live Agent sidebar, click **Settings > Agents**.

Your name will be listed by default as Administrator.
An Administrator will have access to all Live Agent features enabled for your account,
including settings, along with all the normal Agents' privileges.

## Agent Statuses

Agent statuses help manage the availability and workload of human agents, ensure consistent and efficient customer service, and provide transparency regarding the agent's availability to handle chat requests.

In Live Agent, human agents can have the following statuses:

* **Online**. This status indicates that the agent is available and actively handling customer interactions.
* **Busy**. This status indicates that the agent is currently engaged in productive activities or actively assisting a customer. While in this state, the agent may not be immediately available to accept new chat requests. Agents in the `Busy` status temporarily defer new customer interactions until they return to the `Online` status. However, for enhanced workflow management, you can activate the **Assign conversation to a busy agent** setting in the [Account settings](/live-agent/settings/account-settings#auto-assign-conversations-to-busy-agents), empowering agents to efficiently handle new customer conversations even when they are occupied with ongoing tasks or interactions.
* **Away**. This status indicates that the agent is temporarily unavailable or not actively monitoring the chat system. They may take a break, attend to other responsibilities, or leave their workstation. Agents in this status cannot handle customer conversations until they are back in.
* **Offline**. This status indicates that the agent is not currently logged in to the Live Agent system. Agents in this status cannot handle customer interactions until they log back in.


# Attachments
Source: https://docs.cognigy.com/live-agent/settings/attachments

Manage file attachments in Live Agent conversations, including supported formats like PDF, DOCX, and images, and how human agents upload and access files.

<a href="/release-notes/4.70"><Badge>Updated in 4.70</Badge></a>

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

*Attachments* are files or documents that are sent or received during a conversation between a human agent and a user.

These attachments can include various file types; `jpeg`, `jpg`, `png`, `docx`, `wav`, `ogg`, `pdf`.

Attachments can provide additional information or media, for example, an invoice or a photo related to the issue a customer is facing.

## Upload Attachments as Agents

To upload a file in a conversation within Live Agent, follow these steps:

1. Go to Live Agent.
2. On the left side of the screen, select **Conversations**.
3. In the **Inboxes** column, select an Inbox.
4. Select your conversation.
5. At the bottom of the conversation chat, in the **Reply to User** section, click **Attach files**.
6. In the opened window, select the file you want to upload and click **Open**.
7. When the file is uploaded, click **Send**.

The file will be sent to the user and will appear in the chat. The user can download this attachment.

## Upload Attachments as End Users

End users can send attachments to human agents by using the **Add attachments** icon in the chat if you set up a [Storage File](/ai/agents/deploy/endpoints/file-storage) provider in the Endpoint Settings.

## Request Attachments with the Plugin

If you are using a [Webchat](/webchat/overview) Endpoint with Live Agent, in addition to the standard Attach file feature, human agents can request an attachment from a user. To enable this functionality, you need to add the [File Upload plugin](https://github.com/Cognigy/WebchatPlugins/tree/master/plugins/file-upload) to the Webchat Endpoint to upload files.

To request an attachment in a conversation within Live Agent, follow these steps:

1. Go to Live Agent.
2. On the left side of the screen, select **Conversations**.
3. In the **Inboxes** column, select an Inbox.
4. Select your conversation.
5. At the bottom of the conversation chat, in the **Reply to User** section, click **Request file**.
6. In the opened window, click **Send**. The file will be requested from the user, and the request will appear in the chat.
7. Wait for the attachment to be sent by the user. The file will appear in the chat.

When the file has been uploaded, you can view and download this file directly from the chat.

## View Attachments

To view all reviewed and sent attachments, follow these steps:

1. Go to Live Agent.
2. On the left side of the screen, select **Settings**.
3. Go to **Attachments**.

You will see the list of all attachments that were sent by end users and human agents.

## Download Attachments

To download an attachment within Live Agent, you can do this either in a conversation or on the **Attachments** page.

<Tabs>
  <Tab title="Attachment Page">
    1. Go to Live Agent.
    2. On the left side of the screen, select **Settings**.
    3. Go to **Attachments**.
    4. Go to the file you want to download.
    5. In the **ACTIONS** column, click **Download**.
  </Tab>

  <Tab title="Conversation Page">
    1. Go to Live Agent.
    2. On the left side of the screen, select **Conversations**.
    3. In the **Inboxes** column, select an Inbox.
    4. Select your conversation.
    5. Find the attachment in the conversation, click **Download**.
  </Tab>
</Tabs>

The file will be opened in a new tab of your browser. Click your file to save it to your local device.

## Delete Attachments

To delete an attachment within Live Agent, follow these steps:

1. Go to Live Agent.
2. On the left side of the screen, select **Settings**.
3. Go to **Attachments**.
4. Go to the file you want to download.
5. In the **ACTIONS** column, click **Delete**.
6. Confirm the deletion.

The file will be deleted from the Attachments page and will no longer be available in the conversation where it was uploaded.

## Filter Attachments

To filter attachments within Live Agent, follow these steps:

1. Go to Live Agent.
2. On the left side of the screen, select **Settings**.
3. Go to **Attachments**.
4. On the **Attachments** page, click **Filter Attachments**
5. In the **Filter Attachments** window, select the options you need to filter by:
   * **Cognigy Session ID**
   * **Cognigy User ID**
   * **Filename**
6. Enter the value you need to filter by and click **Apply filters**.

Optionally, you can add additional filters by clicking **+ Add Filter**.

The **Attachments** page will only display results that match the applied filter.

To drop the filter, click **Clear Filters**.

## Open the source of the Attachment

The source can be Inbox, Conversation, or Cognigy Session ID.

To investigate a conversation or Inbox where an attachment was uploaded, follow these steps:

1. Go to Live Agent.
2. On the left side of the screen, select **Settings**.
3. Go to **Attachments**.
4. Go to the row of the file you want to investigate.
5. On this row, click one of the following links:
   * **INBOX** — the option will navigate you to the Inbox of the conversation where the attachment was uploaded.
   * **CONVERSATIONS** — the option will navigate you to the conversation where the attachment was uploaded.
   * **COGNIGY SESSION ID** — the option will navigate you to the conversation where the attachment was uploaded.

## More Information

* [Conversation Workflow](/live-agent/conversation/overview)


# Audit Logs
Source: https://docs.cognigy.com/live-agent/settings/audit-logs

Track user activities and configuration changes in Live Agent with audit logs. Monitor create, update, and delete events for security and compliance.

<a href="/release-notes/4.57"><Badge>Updated in 4.57</Badge></a>

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

*Audit Logs* are a tool for maintaining transparency and accountability within your Live Agent resources.

These logs record important user activities, system events, and configuration changes, providing details for security and auditing purposes.

With audit logs, you can answer "Who did what, where, and when?" questions within your Live Agent environment.

The logs include events such as `create`, `update`, and `delete` for the following entities:

* [Agent Copilot Bot](/live-agent/assistants/agent-copilot-bot)
* [Automation Rules](/live-agent/settings/automation-rules)
* [Inboxes](/live-agent/settings/inboxes)
* [Inbox Collaborators](/live-agent/settings/inboxes)
* [Inbox Webhooks](/live-agent/settings/inboxes)
* [Teams](/live-agent/settings/teams)

Only users with the `Administrator` role can access these logs.

## Structure of Audit Log Entries

Every audit log entry provides a structured format that includes the user responsible for the activity, the specific action performed, and the timestamp of the event.

```txt theme={null}
| User             | Activity                          | Time                  |
| ---------------- | --------------------------------- | --------------------- |
| System           | Created a new inbox (#24725)      | Jun 22, 2023 03:06 PM |
| user@cognigy.com | Deleted an automation rule (#376) | Jun 22, 2023 12:48 PM |
```

In the first log entry, the system created a new inbox with the identifier `#24725`. The activity occurred on `Jun 22, 2023, at 03:06 PM`.

In the second log entry, the user with the `user@cognigy.com` login deleted an automation rule with the identifier `#376`. The activity occurred on `Jun 22, 2023, at 12:48 PM`.

## View Audit Logs

To view audit logs, go to **Settings > Audit Logs** in the Live Agent interface.

You will see a list of log entries containing user and system events.

### View Changes

For actions, you can view details on any changes that were made.

To do this, follow these steps:

1. In the Live Agent interface, go to **Settings > Audit Logs**.
2. Select a log from the list by clicking on it.

A window will open displaying the details of the changes.

For an `update` action, the **Before** section presents the original values or configuration, while the **After** section displays the new values after the update.

<Frame>
  <img alt="Audit log change details window showing Before and After sections for an update action" />
</Frame>


# Automation
Source: https://docs.cognigy.com/live-agent/settings/automation-rules

Set up Live Agent automation rules to auto-assign conversations, add labels, set priorities, and trigger actions based on custom workflow conditions.

<a href="/release-notes/4.61"><Badge>Updated in 4.61</Badge></a>

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

*Automation Rules* are a powerful tool to enhance efficiency and streamline processes by automating various tasks within your workflow. With these rules, you can easily assign conversations to specific agents or teams, add labels, adjust priorities, and more without manual intervention. By leveraging Automation Rules, you save time and resources, allowing your team to focus on their core competencies and minimize the time spent on routine tasks.

## Automation Rule Components

Each automation rule contains the following components:

* [Events](#events)
* [Conditions](#conditions)
* [Actions](#actions)

### Events

Automation Events serve as triggers for when you want your automation to be executed.

| Events               | Description                                                                                                          |
| -------------------- | -------------------------------------------------------------------------------------------------------------------- |
| Conversation created | Triggers when a new conversation is created in your system. This will include conversations created by all channels. |
| Conversation updated | Triggers when a conversation is updated.                                                                             |
| Message created      | Triggers when any new message inside a conversation is created.                                                      |
| Conversation opened  | Triggers when a conversation is reopened in your system.                                                             |

### Conditions

Conditions are the criteria that are verified before executing the action. The conditions are evaluated in the order they are defined.

| Condition Types | Description                             | Applicable Events                                                                |
| --------------- | --------------------------------------- | -------------------------------------------------------------------------------- |
| Assignee        | Based on conversation assignee.         | Conversation updated, Conversation opened                                        |
| Status          | Based on conversation status.           | Conversation created, Conversation updated                                       |
| Inbox           | Based on the Inbox of the conversation. | Conversation created, Conversation updated, Conversation opened, Message created |
| Priority        | Based on the conversation's priority.   | Conversation created, Conversation updated, Conversation opened                  |
| Team            | Based on the assigned team.             | Conversation updated, Conversation opened                                        |
| Message Type    | Based on the type of message.           | Message created                                                                  |
| Message Content | Based on the content of the message.    | Message created                                                                  |

### Actions

Actions are tasks or processes that are executed when the conditions are met.

| Actions               | Description                                      |
| --------------------- | ------------------------------------------------ |
| Assign an agent       | Assigns a specific agent to the conversation.    |
| Assign a team         | Assigns the conversation to a specific team.     |
| Add Label             | Adds a label to the conversation.                |
| Send Email to Team    | Sends an email notification to the team.         |
| Send Email Transcript | Sends an email with the conversation transcript. |
| Mute Conversation     | Mutes notifications for the conversation.        |
| Snooze Conversation   | Temporarily suppresses the conversation.         |
| Resolve Conversation  | Marks the conversation as resolved.              |
| Send Attachment       | Sends an attachment to the conversation.         |
| Send Message          | Sends a message to the conversation.             |
| Change Priority       | Adjusts the conversation's priority.             |

## Create Automation Rules

To create an automation rule, follow these steps:

1. In the Live Agent interface, go to **Settings > Automation**.
2. In the upper-right corner, click **+ Add Automation Rule**.
3. In the **Rule Name** field, enter a unique name for the rule.
4. In the **Description** field, enter a relevant description for the rule.
5. From the **Event** list, select the desired event.
6. In the **Conditions** section, set up a condition:
   * **Left operand** — select a [condition type](#conditions).
   * **Operator** — a comparison operator that checks whether the value of the left operand is `Equal to` or `Not Equal to` the value of the right operand.
   * **Right operand** — select values corresponding to condition types. For example, if you choose `Status` as a condition type, you can add multiple statuses, such as `Open`, `Pending`, `Resolved`.
7. *(Optional)* Add an additional condition by clicking **+ Add Condition**. Repeat the same steps as you did for the first condition and select one of the following operators:
   * **AND** — all conditions must be true for the overall condition to be true. For example, `Condition A AND Condition B` means that both `Condition A` and `Condition B` must be true to trigger the action.
   * **OR** — at least one of the conditions must be true for the overall condition to be true. For example, `Condition A OR Condition B` means that if either `Condition A` or `Condition B` is true, the action will be triggered.
8. In the **Actions** section, select an action. Note that you need to provide additional information depending on which type of action you selected. For example, if you select **Send an email to a Team**, you need to choose a team and write an email template.
9. *(Optional)* Add one more action.
10. Click **Create**.

The automation rule appears in the automation rule list and is automatically activated.

To deactivate the rule, move the toggle switch to the left position.

### Example

You can use multiple conditions with the `AND` or `OR` operator as in the example below.

When Agent Copilot is active, and a new conversation is created with the Open status, the conversation is assigned to a team.

* Name: Sales
* Description: This automation assigns new **Open** status conversations to the Sales team.
* Select the event as Conversation Created.
* Add two conditions and combine them using the `AND` operator:
  * Condition 1: Check if the **Conversation Status** is **Open**.
  * Condition 2: Verify that the **Inbox** is set to **AI Copilot bot-default** selected from the dropdown menu.
* Add an action: Assign the conversation to the Sales team by selecting it from the dropdown menu. Make sure the Sales team is created before executing this action.

<Frame>
  <img alt="Adding an automation rule with conditions and actions in Live Agent" />
</Frame>

## Edit Automation Rules

To edit an automation rule, click the **Edit** icon next to the desired automation rule in the list.

## Clone Automation Rules

To make a copy of an existing automation rule, click the clone icon from the Automation list you want to copy.
A replica of the automation rule with the same conditions and actions will be created.

## Delete Automation Rules

To delete an automation rule, click **Delete** next to the desired Automation in the list. Note that this action is permanent and cannot be undone.

## More Information

* [Settings](/live-agent/settings/overview)
* [Conversation Workflow](/live-agent/conversation/overview)
* [Conversation Routing](/live-agent/conversation/conversation-routing/overview)


# Canned Responses
Source: https://docs.cognigy.com/live-agent/settings/canned-responses

Create and manage canned response templates in Live Agent to help human agents reply faster with pre-saved, HTML-formatted quick replies to common questions.

*Canned responses* are pre-saved reply templates allowing human agents to respond to a conversation quickly.

Canned Response templates provide support for HTML formatting.

## Create Canned Responses

<a href="/release-notes/4.55"><Badge>Updated in 4.55</Badge></a>

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

To create a new canned response, follow these steps:

1. In the Live Agent interface, go to **Settings > Canned Responses**.
2. In the upper-right corner, click **Add Canned Response**.
3. In the **Add Canned Response** window, fill in the following fields:
   * **Short Code** — add a word that associates with the response.
   * **Content** — add a message that will be sent to the user. In this message, you can replace user contact profile data or human agent profile data with specific variables, such as <code>\{\{agent.name}}</code>, <code>\{\{contact.first\_name}}</code>, <code>\{\{contact.email}}</code>, and more. To include a variable, begin typing <code>\{\{</code> in the field. After that, a list of available variables will appear, allowing you to select the desired variable.
   * **Level** — select one of the following options from the list:
     * **Global** — all human agents can use a canned response.
     * **Personal** — only you can use a canned response.
     * **Team** — only human agents from the selected [team](/live-agent/settings/teams) can use a canned response. You can select only one team.
4. Click **Submit**.

The canned response will appear in the list on the Canned Responses page.

Alternatively, you can add a canned response or a list of canned responses by importing a CSV file.

## Import Canned Responses

<a href="/release-notes/4.52"><Badge>Updated in 4.52</Badge></a>

To import canned responses, follow these steps:

1. Prepare a CSV file with `short_code` and `content` as headers. For example:
   ```txt theme={null}
    | short_code | content                             |
    |------------|-------------------------------------|
    | delivery   | When do you want to place an order? |
   ```
2. Open the Live Agent interface.
3. Go to **Settings > Canned Responses**.
4. In the upper-right corner, click **Import Canned Response**.
5. In the **Import Canned Response** window, click **Select File**.
6. Upload the CSV file that you previously created.
7. Click **Import**.

The canned responses will appear in the list on the Canned Responses page.

## Use Canned Responses

To start using canned responses in the conversations, go to the [Send a Canned Response](/live-agent/conversation/send-reply#advanced-quick-reply).


# Inboxes
Source: https://docs.cognigy.com/live-agent/settings/inboxes

Create and manage Live Agent inboxes for routing conversations, configuring human agent assignments, message visibility, and integrating AI assistants.

<a href="/release-notes/2026.4"><Badge>Updated in 2026.4</Badge></a>

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

*Inboxes* are workspaces where human agents handle real-time chat interactions with customers and manage related tasks. Each Inbox ID represents the connection from Cognigy to Live Agent.

You can have an unlimited number of Inboxes in your Live Agent account, and you can view conversations specific to an Inbox by clicking on it.

Within the Inbox setting, you can edit the Inbox, add and remove agents, configure message visibility for them, set up conversation routing, and add assistant bots.

## Create Inboxes

To create an Inbox in Live Agent, do the following:

1. In the Live Agent interface, go to **Settings > Inboxes**.
2. In the upper-right corner, click **+ Add Inbox**. Live Agent will show you a screen to set up your Inbox.
3. On the **Inbox setup** page, fill in the following fields:
   * **Inbox Name** — create a unique Inbox name. For example, `Cognigy Handover`. The name can contain any characters, but can't start with `=`.
   * **Webhook URL** — enter the Cognigy Endpoint Config URL domain name, extended by `/handover/liveagent`. For example, `https://endpoint-trial.cognigy.ai/handover/liveagent`. You can find the Endpoint Config URL in the **Configuration Information** section of your Endpoint.
4. In the **Pick agents for the inbox** list, select agents by either clicking it or hovering your cursor over it and then pressing <kbd>Enter</kbd>. You can select multiple agents from the list. Agents can belong to different inboxes simultaneously.
5. Click the **Create Cognigy Inbox** button. Once you have successfully set up the Inbox, you are prompted to assign your agents to the new Inbox in the next step.
6. Select the required agents from the selection list, then click **Add Agents**.

A new Live Agent Inbox has been created. You can configure the Inbox settings by clicking **More settings**.

<Note>
  Users with the `Agent` or `Supervisor` role can view only the Inboxes to which they are assigned. Users with the `Administrator` role can view all Inboxes. For more information on user permissions, refer to [Roles](/live-agent/roles).
</Note>

## Edit Inboxes

If you want to change how an Inbox works or reuse an Inbox from a deleted Project, you need to edit the Inbox.

To edit an Inbox, follow these steps:

1. In the Live Agent interface, go to **Settings > Inboxes**.
2. On the **Inboxes** page, click <img alt="Inbox settings" /> next to the Inbox you want to edit.
3. In the **Inboxes Settings** section, update the following fields:
   * **Channel Avatar** — upload an image that represents the Inbox. The Channel Avatar helps you visually identify the Inbox.
   * **Inbox Name** — enter a unique Inbox name.
   * **Cognigy Project ID** — this field is pre-populated for [Inboxes created through Cognigy](/live-agent/getting-started/live-agent-setup/live-agent-setup-additional-inbox). Enter a valid Project ID. The Project ID allows Live Agent to retrieve [Contact Profile](/ai/agents/analyze/contact-profiles) data from Cognigy and binds the Inbox to the respective Project. You can [copy the Project ID](/ai/agents/develop/projects-and-flows/projects) from the Cognigy interface. After you assign the Inbox to a Project, [copy the Inbox ID](#copy-inbox-id) and enter it in the respective Cognigy Nodes of the Project, for example, the [Handover Agent](/ai/agents/develop/node-reference/service/handover-to-agent) and [Check Agent Availability](/ai/agents/develop/node-reference/service/check-agent-availability) Nodes. If you delete the Project to which the Project ID refers, Live Agent can still communicate with Cognigy Nodes, but the system cannot retrieve Contact Profile data through this Inbox.
   * **Webhook URL** — enter the Cognigy Endpoint Config URL domain name, extended by `/handover/liveagent`. For example, `https://endpoint-trial.cognigy.ai/handover/liveagent`. You can find the Endpoint Config URL in the **Configuration Information** section of your Endpoint.
4. Click **Update**.

## Copy Inbox ID

Each Inbox ID represents the connection between Cognigy and Live Agent and is used in API and UI URL paths.

To connect an Inbox to Cognigy,
copy the ID of the Inbox
and specify it in the **Live Agent Settings** section in the [Handover to Human Agent](/ai/agents/develop/node-reference/service/handover-to-agent) Node on the Cognigy side.

To copy an Inbox ID, follow these steps:

1. Go to Live Agent.
2. On the left side of the screen, select **Settings**.
3. In the settings column, select **Inboxes**.
4. On the **Configuration** tab, go to the **Inbox ID** section.
5. Click **Copy**.

## Delete Inboxes

Deleting an Inbox is irreversible and will cause a loss of synchronization with Cognigy.

To delete an Inbox within Live Agent, do the following:

1. In the Live Agent interface, go to **Settings > Inboxes**.
2. On the **Inboxes** page, to the right of the team you want to delete, click **Delete**.
3. You are prompted to enter the Inbox name you want to delete.
4. Click **Yes, Delete**.

Live Agent deletes the selected Inbox.

## Control Conversation Routing within Inboxes

To configure the conversation assignments within Live Agent, do the following:

1. In the Live Agent interface, go to **Settings > Inboxes**.
2. On the **Inboxes** page, to the right of the Inbox you want to change, click **Settings** and open the **Collaborators** tab.
3. In the **Conversation Assignment** sections, update the conversation assignment settings:
   * [Allow auto assignment](/live-agent/conversation/conversation-routing/automatic-mode#automatic-assignment)
   * [Allow conversations to be reassigned](/live-agent/conversation/conversation-routing/automatic-mode#automatic-reassignment)

The changes take effect.

## Escalate Conversations to Supervisors

To allow agents to escalate conversations to supervisors within an Inbox, do the following:

1. In the Live Agent interface, go to **Settings > Inboxes**.
2. On the **Inboxes** page, to the right of the Inbox you want to change, click **Settings** and open the **Collaborators** tab.
3. In the **Supervisors** section, select supervisors by either clicking it or hovering your cursor over it and then pressing <kbd>Enter</kbd>. You can select multiple supervisors from the list. Supervisors can belong to different inboxes simultaneously.
4. Once supervisors are added, configure escalation settings:
   * **Allow agents to escalate conversations to supervisors** — activate the option to let agents escalate conversations directly from the [Conversation Actions](/live-agent/conversation/assign-conversations) section.
   * **Send push notifications to supervisors** — activate the option to let supervisors receive real-time notifications when a conversation is escalated.
   * **Send email notifications to supervisors** — activate the option to let supervisors receive email notifications when a conversation is escalated.
5. Click **Update**. After applying these settings, the **↑ Escalate to Supervisor** button appears in the **Conversation Actions** section on the conversation page, allowing the agent to start an escalation. Refer to [Manual Assignment](/live-agent/conversation/conversation-routing/manual-mode) to learn more about supervisor assignment rules, and to [Conversation Queue](/live-agent/conversation/conversation-queue/overview) to see an example of conversation assignment.

## Add Agents to Inboxes

To add agents to an Inbox, do the following:

1. In the Live Agent interface, go to **Settings > Inboxes**.
2. On the **Inboxes** page, to the right of the Inbox you want to change, click **Settings**.
3. On the **Collaborators** tab, go to the **Agents** section.
4. In the agent list, select agents by either clicking it or hovering your cursor over it and then pressing <kbd>Enter</kbd>. You can select multiple agents from the list. Agents can belong to different inboxes simultaneously.
5. Click **Update**.

The changes take effect.

## Configure Agent Access to all Inbox Conversations

By default, your agents can see all messages within an Inbox.
To change this behavior for agents with the `Agent` role, do the following:

1. In the Live Agent interface, go to **Settings > Inboxes**.
2. On the **Inboxes** page, to the right of the Inbox you want to change, click **Settings** and open the **Collaborators** tab.
3. In the **Conversation Assignment** section, disable the **Allow Agents to see all conversations in this inbox** setting.

The changes take effect. Note that this setting does not affect agents with the `Administrator` or `Supervisor` roles.
They will still be able to see all messages within the Inbox.

## Remove Agents from Inboxes

To remove the Inbox appearance, do the following:

1. In the Live Agent interface, go to **Settings > Inboxes**.
2. On the **Inboxes** page, to the right of the Inbox you want to change, click **Settings**.
3. On the **Collaborators** tab, go to the **Agents** section.
4. In the agent list, remove an agent by clicking **x** on the right side of the agent's name.
5. Click **Update**.

## Add Agent Copilot Bots

Within an Inbox, you can add one or more Agent Copilot Bots to assist your human agents.
To create such a bot, you need to have a separate Flow and Endpoint on the Cognigy side.
For more details, refer to [Agent Copilot Bot](/live-agent/assistants/agent-copilot-bot).

## More Information

* [Agents](/live-agent/settings/agents)
* [Agent Copilot Bot](/live-agent/assistants/agent-copilot-bot)
* [Conversation Routing](/live-agent/conversation/conversation-routing/overview)


# Labels
Source: https://docs.cognigy.com/live-agent/settings/labels

Organize and prioritize conversations in Live Agent using labels. Learn how to create, edit, and delete conversation tags for custom workflows.

<a href="/release-notes/4.88"><Badge>Updated in 4.88</Badge></a>

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

*Labels* are tags that help you to categorize and prioritize conversations, providing additional context and information about the content or status of each conversation.

They can be used to create custom workflows, to make conversations easier to identify, filter, and organize.

Only Administrators and Supervisors can create, edit, and delete labels.

### Create Labels

To create a label within Live Agent, do the following:

1. Go to Live Agent.
2. On the left side of the screen, select **Settings**.
3. In the settings column, select **Labels**.
4. At the top right of the window, click **Add label**.
5. In the **Add label** window, fill in the following fields:
   * **Label Name** — a name for the label to identify its category. For example, `Premium_Customer`. The name can contain any characters but can't start with `=` and must be at least 2 characters long.
   * **Description** — a brief explanation of the label's meaning. For example, `This issue needs to be handled with priority`.
   * **Level** — a level of access to objects where the label is applicable. Select one of the following levels:
     * **Global** — the label is available for all Teams and Inboxes.
     * **Team** — the label is available only for a specific Team. Select a Team from the **Team** list. Note that you cannot select multiple Teams.
     * **Inbox** — the label is available only for a specific Inbox. Select an Inbox from the **Inbox** list. Note that you cannot select multiple Inboxes.
   * **Color** — a color for the label to make it visually identifiable. Select your desired color, click the small square again to confirm your selection.
   * **Show Label on Sidebar** — the option determines whether a label will be displayed on the sidebar. By using the label on the sidebar, you can filter conversations.
6. Click **Create**.

The label will be created, appear under the **Labels** section on the sidebar, and will be available to [add to Conversations](#add-labels-to-conversations).

<Tip>
  It is good practice to assign meaningful names to labels and use selected colors to visually identify the categories or topics to which a conversation belongs.
</Tip>

### Edit Labels

To edit a label within Live Agent, do the following:

1. Go to Live Agent.
2. On the left side of the screen, select **Settings**.
3. In the settings column, select **Labels**.
4. On the **Labels** page, to the right of the label you want to change, click **Edit**.
5. In the **Edit label** window, update the fields as needed.
6. Click **Edit**.

The label fields and settings will be updated.

### Delete Labels

To delete a label within Live Agent, do the following:

1. Go to Live Agent.
2. On the left side of the screen, select **Settings**.
3. In the settings column, select **Labels**.
4. On the **Labels** page, to the right of the label you want to delete, click **Delete**.
5. Confirm deletion by clicking **Yes, Delete**.

The label will be deleted and no longer appear under the **Labels** section on the sidebar, and will no longer be available to add to Conversations. Also, the label will be removed from the conversations where it was added earlier.

### Add Labels to Conversations

To start using Labels in Conversations, go to [Assign Labels](./../conversation/assign-conversations#assign-labels).


# Settings
Source: https://docs.cognigy.com/live-agent/settings/overview

Overview of Cognigy Live Agent settings, including team management, inbox configuration, automation rules, canned responses, and account settings.

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

Cognigy Live agent settings offer many features, such as team and inbox management, automation rules, canned responses, attachments, audit logs, and account settings. These capabilities can significantly improve customer interactions and enhance support efficiency.

Live Agent settings include the following sections:

* [Agents](/live-agent/settings/agents)
* [Teams](/live-agent/settings/teams)
* [Inboxes](/live-agent/settings/inboxes)
* [Labels](/live-agent/settings/labels)
* [Automation Rules](/live-agent/settings/automation-rules)
* [Canned Responses](/live-agent/settings/canned-responses)
* [Attachments](/live-agent/settings/attachments)
* [Audit Logs](/live-agent/settings/audit-logs)
* [Account Settings](/live-agent/settings/account-settings)


# Skills
Source: https://docs.cognigy.com/live-agent/settings/skills

Define agent skills in Live Agent for skill-based routing, assign language proficiency and domain expertise to improve conversation assignment.

<a href="/release-notes/4.60"><Badge>Updated in 4.60</Badge></a>

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

*Skills* are specific abilities that human agents possess, such as language proficiency and expertise in particular domains.

Only users with the Administrator and Supervisor roles can configure and assign skills.

## Create Skills

To create a skill, follow these steps:

1. In the Live Agent interface, go to **Settings > Skills**.
2. In the upper-right corner, click **Add Skill**.
3. In the **Add Skill** window, fill in the following fields:
   * **Type** — select one of the following types:
     * **Language** — a skill that refers to language expertise.
     * **Skill** — a skill that refers to specific knowledge.
   * **Name** — specify a unique name that reflects the skill. For example, `English`, `Spanish`, `German` for the **Language** type, or specific skills like `IT`, `HR`, `Customer Service` for the **Skills** type.
4. Click **Submit**.

The skill will appear in the list on the **Skills** page.
Afterward, you can edit or delete the added skill.

## Assign Skills

To assign a skill to a human agent, follow these steps:

1. In the Live Agent interface, go to **Settings > Agents**.
2. In the row with the agent, click **Skills**.
3. In the **Skills** window, click **Add skill**.
4. Select a skill type and its value. You can add multiple skills.
5. Click **Add Skills**.

The skills will appear in the **Skills** column of the row with the agent.

If you want to remove a skill from the row with the agent, open the **Skills** window and click <img alt="remove skill" /> on the right side of the selected skill.

## Apply Skills

To start using skills in conversation routing, you need to specify your created skills in the [Handover to Human Agent](/ai/agents/develop/node-reference/service/handover-to-agent) Node.

If you want to assign conversations only to agents relevant to their skills, in addition to the Handover to Human Agent Node, you can add the [Check Agent Availability](/ai/agents/develop/node-reference/service/check-agent-availability) Node.
In this case, if a relevant agent is not found, you can notify the user. This Node checks only online agents with the relevant skills and performs routing.


# Teams
Source: https://docs.cognigy.com/live-agent/settings/teams

Create and manage teams of human agents in Live Agent to handle specific customer inquiries, assign conversations, and collaborate efficiently.

<a href="/release-notes/4.88"><Badge>Updated in 4.88</Badge></a>

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

*Teams* are organized groups of human agents within your customer support team. These teams are formed based on their specific responsibilities or expertise in handling different types of customer inquiries or issues.

A human agent can be part of multiple teams.

Human agents can assign conversations manually to a team when they are working collaboratively.

## Create Teams

To create a team within Live Agent, do the following:

1. In the Live Agent interface, go to **Settings > Teams**.
2. In the upper-right corner, click **Create new team**.
3. In the **Create a new team** window, fill in the following fields:
   * **Team name** — enter a meaningful name of the team, that associates with its responsibilities. For example, `Team Customer Support`. The name can contain any characters but can't start with `=` and must be at least 2 characters long.
   * **Team Description** — enter a short description about this team.
4. Click **Create team**.
5. On the **Add agents to team** page, add available agents you want to your newly created team. This lets you collaborate as a team on conversations and get notified on new events in the same conversation. You can also select all available agents at once for the new team by activating the setting next to **AGENT**.
6. Click **Add agents** to confirm your selection.
7. Click **Finish**.

The new team will be added and displayed on the **Teams** page.

## Edit Teams

To edit a team within Live Agent, do the following:

1. In the Live Agent interface, go to **Settings > Teams**.
2. On the **Teams** page, to the right of the team you want to change, click **Edit**.
3. In the **Edit your team** window, update the fields as needed.
4. Click **Update team**.

The team fields and settings will be updated.

## Delete Teams

To delete a team within Live Agent, do the following:

1. In the Live Agent interface, go to **Settings > Teams**.
2. In the settings column, select **Teams**.
3. On the **Teams** page, to the right of the team you want to delete, click **Delete**.
4. You are prompted to enter the team name you want to delete, for example, `Team Customer Support`, then click **Delete Team Customer Support** to confirm.

Deleting the team will remove the team assignment from the conversations assigned to this team.

## Assign Teams to Conversations

To manually assign a team to a conversation within Live Agent, follow these steps:

1. In the Live Agent interface, go to **Conversations**.
2. Go to the relevant conversation.
3. In the **Details** menu, go to the **Conversation Actions** section.
4. From the **Assigned Team** list, select a team that can be assigned to this conversation. Note that the previously selected agent in the **Assigned Agent** list should be a part of this team.

The conversation will be assigned to the team.

## More Information

* [Agents](/live-agent/settings/agents)
* [Manual assignment](/live-agent/conversation/conversation-routing/manual-mode#manual-assignment)
* [Conversation assignment](/live-agent/conversation/assign-conversations)


# Live Agent OData Analytics Endpoint
Source: https://docs.cognigy.com/live-agent/tools/odata-endpoint

Access Live Agent analytics data via the OData endpoint. Query conversation and agent metrics for dashboards, BI tools, and data exports.

<a href="/release-notes/2025.25"><Badge>Updated in 2025.25</Badge></a>

<Warning>
  Live Agent has been put in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

Live Agent exposes an OData 4 analytics endpoint to retrieve analytics records. [OData](https://www.odata.org/), the best way to REST, is a powerful API framework. The OData endpoints allow you to retrieve the most relevant data models from Live Agent by using GET requests. It covers all your enterprise analytics needs to make fine-grained queries in your spreadsheets or build rich dashboards for your bots with your favorite data visualization tool.

The Live Agent OData endpoint supports only `GET` requests and doesn't support any other request types, such as `PATCH`, `DELETE`, or `POST`.

<blockquote>
  <span>▶️</span>

  <div>
    <div>
      <a href="https://support.cognigy.com/hc/en-us/articles/360019467199-Cognigy-Sessions-Analytics-OData">Technical video "Analytics & OData"</a>
    </div>

    <div>
      Watch this Episode of Cognigy Sessions for a technical deep dive
    </div>
  </div>
</blockquote>

## Supported Versions

| OData endpoint version | Live Agent version  | Status    | New in Version                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           |
| ---------------------- | ------------------- | --------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| 2.0                    | 2025.25.0 and later | Supported | The Team and AgentAvailabilityRecord collections were added. A new field, `priority`, was added to the Conversation model, and the changes introduced in 1.1 were rolled back.                                                                                                                                                                                                                                                                                                                           |
| 1.1                    | 2025.18.0 or later  | Supported | Rollback to pre-1.0 fields: - **Restored fields**: Conversation (`identifier`, `contact_last_seen_at`), Inbox (`greeting_enabled`, `greeting_message`, `email_address`), Message (`status`, `source_id`), User (`encrypted_password`, `reset_password_token`, `reset_password_sent_at`, `remember_created_at`, `current_sign_in_ip`, `last_sign_in_ip`, `confirmation_token`) - **No longer supported**: Conversation (`assigned_at`, `ended_at`), Label (`conversation_ids`), User (`last_sign_out_at`) |
| 1.0                    | 1.0.0 and later     | Supported | Initial release.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         |

## Usage and Authentication

<div />

To connect to the OData endpoint, get a Live Agent Access Token:

1. In the left-side menu of the Live Agent interface, go to **Profile Settings**.
2. In the **Access Token** section, copy the existing token.
3. Paste this token into your OData request.

An OData URL is composed of the service root, API version, service path, data collection, and Live Agent access token parameter as follows:

```txt theme={null}
GET /<api-version>/odata/<OData data model>?<odata query> HTTP/1.1
Host: <odata domain>
Authorization: Bearer <Live Agent access token>
```

The authentication can also be done using a query param such as follows:

```text theme={null}
/<api-version>/odata/<OData data model>?<odata query>&$top=1&$skip=0&apikey=<Live Agent access token>
```

<AccordionGroup>
  <Accordion title="OData Domain Name">
    The OData endpoint is available on a different domain from your Cognigy User Interface domain. For example, [https://odata-liveagent-trial.cognigy.ai/v1.0/odata](https://odata-dev-live-agent.cognigy.ai/v1.0/odata).

    For example, on our trial server, the OData endpoint URL for the Analytics Inputs Collection is `https://odata-liveagent-trial.cognigy.ai/v1.0/odata/Message?`. For on-premises installations, replace the `odata-trial.cognigy.ai` domain name with the domain name configured for your local installation.
  </Accordion>

  <Accordion title="Excel/Power BI">
    When using Power BI or Excel, you might be asked to authenticate. Simply choose `anonymous authentication` and pass the Live Agent access token as a query parameter `&apikey=<Live Agent access-token>`
  </Accordion>
</AccordionGroup>

## Endpoint Versions

<div />

<Tabs>
  <Tab title="Version 2.0">
    This endpoint version is available in Live Agent 2025.25.0 and later. In this version, the following OData collections are available:

    * [Message](#message) (`/Message`)
    * [Conversation](#conversation) (`/Conversation`)
    * [Inbox](#inbox) (`/Inbox`)
    * [Reporting Event](#reporting-event) (`/ReportingEvent`)
    * [Label](#label) (`/Label`)
    * [User](#user) (`/User`)
    * [Team](#team) (`/Team`)
    * [Agent Availability Record](#agentavailabilityrecord) (`/AgentAvailabilityRecord`)

    To see all the available OData models, use the following Endpoint:
    `https://<hostname>/v2.0/odata/`

    The URL for accessing the OData endpoint 2.0 is the following:
    `https://<hostname>/v2.0/odata/<collection>`
  </Tab>

  <Tab title="Version 1.1">
    This endpoint version is available in Live Agent 2025.18.0 or later. In this version, the following OData collections are available:

    * [Message](#message) (`/Message`)
    * [Conversation](#conversation) (`/Conversation`)
    * [Inbox](#inbox) (`/Inbox`)
    * [Reporting Event](#reporting-event) (`/ReportingEvent`)
    * [Label](#label) (`/Label`)
    * [User](#user) (`/User`)

    To see all the available OData models, use the following Endpoint:
    `https://<hostname>/v1.1/odata/`

    The URL for accessing the OData endpoint 1.1 is the following:
    `https://<hostname>/v1.1/odata/<collection>`
  </Tab>

  <Tab title="Version 1.0">
    This endpoint version is available in Live Agent 1.0.0 and later. In this version, the following OData collections are available:

    * [Message](#message) (`/Message`)
    * [Conversation](#conversation) (`/Conversation`)
    * [Inbox](#inbox) (`/Inbox`)
    * [Reporting Event](#reporting-event) (`/ReportingEvent`)
    * [Label](#label) (`/Label`)
    * [User](#user) (`/User`)

    To see all the available OData models, use the following Endpoint:
    `https://<hostname>/v1.0/odata/`

    The URL for accessing the OData endpoint 1.0 is the following:
    `https://<hostname>/v1.0/odata/<collection>`
  </Tab>
</Tabs>

## Querying

<div />

The Endpoint supports the following OData Query Language operators:

* `$skip`<sup>[1](#footnote1)</sup>
* `$top`<sup>[1](#footnote1)</sup>
* `$filter`
* `$select`
* `$orderby`
* `$expand`

### Example Queries

<Tabs>
  <Tab title="Version 2.0">
    `https://odata-liveagent-trial.cognigy.ai/v2.0/odata/Conversation(1)?$select=inbox_id,account_id&$top=1&$skip=0&apikey=<access-token>`
    Return the columns inbox\_id and account\_id for the Conversation with id=1.

    `https://odata-liveagent-trial.cognigy.ai/v2.0/odata/Conversation?$select=id,account_id&$top=1&$skip=0&apikey=<access-token>`
    Return the columns' ID and account\_id for all Conversations.

    `https://odata-liveagent-trial.cognigy.ai/v2.0/odata/Conversation?$select=id,account_id&$top=1&$skip=0&$filter=created_at le '2021-11-23T00:00:00'&$top=1&$skip=0&apikey=<access-token>`
    Return the columns' ID and account\_id for all the Conversations filtered by the created\_at column being lower or equal to '2021-11-23T00:00:00'
  </Tab>

  <Tab title="Version 1.1">
    `https://odata-liveagent-trial.cognigy.ai/v1.1/odata/Conversation(1)?$select=inbox_id,account_id&$top=1&$skip=0&apikey=<access-token>`
    Return the columns inbox\_id and account\_id for the Conversation with id=1.

    `https://odata-liveagent-trial.cognigy.ai/v1.1/odata/Conversation?$select=id,account_id&$top=1&$skip=0&apikey=<access-token>`
    Return the columns' ID and account\_id for all Conversations.

    `https://odata-liveagent-trial.cognigy.ai/v1.1/odata/Conversation?$select=id,account_id&$top=1&$skip=0&$filter=created_at le '2021-11-23T00:00:00'&$top=1&$skip=0&apikey=<access-token>`
    Return the columns' ID and account\_id for all the Conversations filtered by the created\_at column being lower or equal to '2021-11-23T00:00:00'
  </Tab>

  <Tab title="Version 1.0">
    `https://odata-liveagent-trial.cognigy.ai/v1.0/odata/Conversation(1)?$select=inbox_id,account_id&$top=1&$skip=0&apikey=<access-token>`
    Return the columns inbox\_id and account\_id for the Conversation with id=1.

    `https://odata-liveagent-trial.cognigy.ai/v1.0/odata/Conversation?$select=id,account_id&$top=1&$skip=0&apikey=<access-token>`
    Return the columns' ID and account\_id for all Conversations.

    `https://odata-liveagent-trial.cognigy.ai/v1.0/odata/Conversation?$select=id,account_id&$top=1&$skip=0&$filter=created_at le '2021-11-23T00:00:00'&$top=1&$skip=0&apikey=<access-token>`
    Return the columns' ID and account\_id for all the Conversations filtered by the created\_at column being lower or equal to '2021-11-23T00:00:00'
  </Tab>
</Tabs>

## Reference Documentation

For complete reference documentation, refer to the [extensive collection of resources at OData.org](https://www.odata.org/) and the [Oasis OData URL Convention Documentation](http://docs.oasis-open.org/odata/odata/v4.0/odata-v4.0-part2-url-conventions.html).

## Data Protection & Analytics

<div />

Only users with an admin role in Live Agent will be able to query OData Models.

## Live Agent OData Collections

<div />

This section details the data types that exist within the OData Collections that can be retrieved from the OData endpoint. The following Collections are available:

* [Message](#message)
* [Conversation](#conversation)
* [User](#user)
* [Reporting Event](#reporting-event)
* [Label](#label)
* [Inbox](#inbox)
* [Team](#team)
* [Agent Availability Record](#AgentAvailabilityRecord)

### Message

**Description:**

<Tabs>
  <Tab title="Version 2.0">
    `https://odata-liveagent-trial.cognigy.ai/v2.0/odata/Message?$top=1&$skip=0`
  </Tab>

  <Tab title="Version 1.1">
    `https://odata-liveagent-trial.cognigy.ai/v1.1/odata/Message?$top=1&$skip=0`
  </Tab>

  <Tab title="Version 1.0">
    `https://odata-liveagent-trial.cognigy.ai/v1.0/odata/Message?$top=1&$skip=0`
  </Tab>
</Tabs>

**Data Types:**

When retrieving this data model, the endpoint returns the following fields:

<Tabs>
  <Tab title="Version 2.0">
    | Field Name           | Description                                                                          | Type    | Example                                                                  |
    | -------------------- | ------------------------------------------------------------------------------------ | ------- | ------------------------------------------------------------------------ |
    | `id`                 | The unique message ID                                                                | Number  | `9411`                                                                   |
    | `content`            | The text content of the message                                                      | String  | `"hi"`                                                                   |
    | `account_id`         | The ID of the account where the conversation belongs                                 | Number  | `2`                                                                      |
    | `inbox_id`           | The ID of the inbox where the conversation belongs                                   | Number  | `427`                                                                    |
    | `conversation_id`    | The ID of the conversation to which this message belongs                             | Number  | `1424`                                                                   |
    | `message_type`       | The type of message: `{ incoming: 0, outgoing: 1, activity: 2, template: 3 }`        | Number  | `0`                                                                      |
    | `created_at`         | The timestamp when the message was created                                           | Date    | `"2023-01-20T13:01:24.141Z"`                                             |
    | `updated_at`         | The timestamp when the message was last updated                                      | Date    | `"2023-01-20T13:01:24.141Z"`                                             |
    | `private`            | Indicates if the message is private (visible only to agents)                         | Boolean | `false`                                                                  |
    | `content_type`       | The content type: `text: 0`, `input_text: 1`, `input_textarea: 2`, `input_email: 3`  | Number  | `0`                                                                      |
    | `content_attributes` | JSON object with additional content details, for example, items, external timestamps | JSON    | `{"items":[{"text":"hi","data":null}],"external_created_at":1674219684}` |
    | `sender_type`        | The type of sender: `"Contact"`, `"Agent"`, or `"cognigy"`                           | String  | `"Contact"`                                                              |
    | `sender_id`          | The unique ID of the sender                                                          | String  | `"1045"`                                                                 |
  </Tab>

  <Tab title="Version 1.1">
    | Field Name           | Description                                                                                                                                                                                                        | Type    | Example                                 |
    | -------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------- | --------------------------------------- |
    | `id`                 | Unique message ID                                                                                                                                                                                                  | Number  | 1                                       |
    | `content`            | The message sent                                                                                                                                                                                                   | String  | `Welcome to Live Agent`                 |
    | `account_id`         | Account ID                                                                                                                                                                                                         | Number  | 1                                       |
    | `inbox_id`           | Inbox ID                                                                                                                                                                                                           | Number  | 1                                       |
    | `conversation_id`    | Conversation ID                                                                                                                                                                                                    | Number  | 1                                       |
    | `message_type`       | Type of message `{ incoming: 0, outgoing: 1, activity: 2, template: 3 }`                                                                                                                                           | Number  | 3                                       |
    | `created_at`         | Message Creation Date                                                                                                                                                                                              | Date    | `2023-01-19T13:31:18.553Z`              |
    | `updated_at`         | Message Updated Date                                                                                                                                                                                               | Date    | `2023-01-19T13:31:18.553Z`              |
    | `private`            | Message Visibility                                                                                                                                                                                                 | Boolean | false                                   |
    | `status`             | Message Status `{ sent: 0, delivered: 1, read: 2, failed: 3 }`                                                                                                                                                     | Number  | 3                                       |
    | `source_id`          | Message source ID                                                                                                                                                                                                  | String  | default                                 |
    | `content_type`       | `text: 0`, `input_text: 1`, `input_textarea: 2`, `input_email: 3`, `input_select: 4`, `cards: 5`, `form: 6`, `article: 7`, `incoming_email: 8`, `input_csat: 9`, `typescognigy: 2016`, `request_file_upload: 2017` | Number  | 0                                       |
    | `content_attributes` | JSON Attribute/Value Definitions                                                                                                                                                                                   | JSON    | `{"items":[{"title":"Acme Hardware"}]}` |
    | `sender_type`        | Provides information on who originated the message                                                                                                                                                                 | String  | `cognigy`                               |
    | `sender_id`          | Sender unique ID                                                                                                                                                                                                   | Number  | 1                                       |
  </Tab>

  <Tab title="Version 1.0">
    | Field Name           | Description                                                                                                                                                                                                        | Type    | Example                                 |
    | -------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------- | --------------------------------------- |
    | `id`                 | Unique message ID                                                                                                                                                                                                  | Number  | 1                                       |
    | `content`            | The message sent                                                                                                                                                                                                   | String  | `Welcome to Live Agent`                 |
    | `account_id`         | Account ID                                                                                                                                                                                                         | Number  | 1                                       |
    | `inbox_id`           | Inbox ID                                                                                                                                                                                                           | Number  | 1                                       |
    | `conversation_id`    | Conversation ID                                                                                                                                                                                                    | Number  | 1                                       |
    | `message_type`       | Type of message `{ incoming: 0, outgoing: 1, activity: 2, template: 3 }`                                                                                                                                           | Number  | 3                                       |
    | `created_at`         | Message Creation Date                                                                                                                                                                                              | Date    | `2023-01-19T13:31:18.553Z`              |
    | `updated_at`         | Message Updated Date                                                                                                                                                                                               | Date    | `2023-01-19T13:31:18.553Z`              |
    | `private`            | Message Visibility                                                                                                                                                                                                 | Boolean | false                                   |
    | `content_type`       | `text: 0`, `input_text: 1`, `input_textarea: 2`, `input_email: 3`, `input_select: 4`, `cards: 5`, `form: 6`, `article: 7`, `incoming_email: 8`, `input_csat: 9`, `typescognigy: 2016`, `request_file_upload: 2017` | Number  | 0                                       |
    | `content_attributes` | JSON Attribute/Value Definitions                                                                                                                                                                                   | JSON    | `{"items":[{"title":"Acme Hardware"}]}` |
    | `sender_type`        | Provides information on who originated the message                                                                                                                                                                 | String  | `cognigy`                               |
    | `sender_id`          | Sender unique ID                                                                                                                                                                                                   | Number  | 1                                       |
  </Tab>
</Tabs>

### Conversation

**Description:**

A Conversation is the communication channel opened between an Agent and a client.

**Example Query:**

<Tabs>
  <Tab title="Version 2.0">
    `https://odata-liveagent-trial.cognigy.ai/v2.0/odata/Conversation?$top=1&$skip=0`
  </Tab>

  <Tab title="Version 1.1">
    `https://odata-liveagent-trial.cognigy.ai/v1.1/odata/Conversation?$top=1&$skip=0`
  </Tab>

  <Tab title="Version 1.0">
    `https://odata-liveagent-trial.cognigy.ai/v1.0/odata/Conversation?$top=1&$skip=0`
  </Tab>
</Tabs>

**Data Types:**

When retrieving this data model, the endpoint returns the following fields:

<Tabs>
  <Tab title="Version 2.0">
    | Field Name               | Description                                                                                     | Type   | Example                                                                                                                                                                                                                                                             |
    | ------------------------ | ----------------------------------------------------------------------------------------------- | ------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
    | `id`                     | The unique conversation ID                                                                      | Number | `74160`                                                                                                                                                                                                                                                             |
    | `account_id`             | The account ID where the conversation belongs                                                   | Number | `2`                                                                                                                                                                                                                                                                 |
    | `inbox_id`               | The inbox ID where the conversation belongs                                                     | Number | `34608 `                                                                                                                                                                                                                                                            |
    | `status`                 | The conversation status: open: 0, resolved: 1, pending: 2, snoozed: 3, abandoned: 4, deleted: 5 | Number | `2`                                                                                                                                                                                                                                                                 |
    | `assignee_id`            | The agent ID assigned to the conversation                                                       | Number | `null`                                                                                                                                                                                                                                                              |
    | `created_at`             | The conversation creation date                                                                  | Date   | `"2023-12-08T13:39:57.528Z"`                                                                                                                                                                                                                                        |
    | `updated_at`             | The date when the conversation was last updated                                                 | Date   | `"2023-12-08T14:58:04.695Z"`                                                                                                                                                                                                                                        |
    | `contact_id`             | The unique client ID                                                                            | Number | `"42992"`                                                                                                                                                                                                                                                           |
    | `display_id`             | The display ID for the conversation                                                             | Number | `6133`                                                                                                                                                                                                                                                              |
    | `agent_last_seen_at`     | The last seen date of the agent                                                                 | Date   | `"2023-12-19T06:51:27.875Z"`                                                                                                                                                                                                                                        |
    | `contact_inbox_id`       | The contact inbox ID                                                                            | Number | `"117788"`                                                                                                                                                                                                                                                          |
    | `last_activity_at`       | The date of the last activity in the conversation                                               | Date   | `"2023-12-08T14:58:04.693Z"`                                                                                                                                                                                                                                        |
    | `team_id`                | The team ID the conversation belongs to                                                         | Number | `null`                                                                                                                                                                                                                                                              |
    | `snoozed_until`          | The end date until which the conversation is snoozed                                            | Date   | `null`                                                                                                                                                                                                                                                              |
    | `assignee_last_seen_at`  | The last seen date of the assigned agent                                                        | Date   | `null`                                                                                                                                                                                                                                                              |
    | `first_reply_created_at` | The date when the first reply was created                                                       | Date   | `null`                                                                                                                                                                                                                                                              |
    | `assigned_at`            | The date when the conversation was assigned                                                     | Date   | `null`                                                                                                                                                                                                                                                              |
    | `ended_at`               | The date when the conversation ended                                                            | Date   | `null`                                                                                                                                                                                                                                                              |
    | `priority`               | The priority of the conversation: low: 0, medium: 1, high: 2, urgent: 3, escalation: 4.         | Number | `null`                                                                                                                                                                                                                                                              |
    | `additional_attributes`  | A JSON object containing any additional attributes                                              | JSON   | ""                                                                                                                                                                                                                                                                  |
    | `uuid`                   | The unique identifier                                                                           | String | `"e41a989a-81c8-49f7-af64-9a85482c67da"`                                                                                                                                                                                                                            |
    | `custom_attributes`      | A JSON object containing extra add-on attributes                                                | JSON   | `{"enableAgentInject":false,"cognigy_channel_type":"admin-webchat","cognigy_agent_assist_workspace_url":"https://agent-assist-trial.cognigy.ai/?sessionId=session-a369b88c-03fb-4497-a60b-495b94313210&userId=4cc1adcf-df86-401d-86ff-4effd5e3c9b1&URLToken=1234"}` |
    | `cognigy_session_id`     | The unique session ID                                                                           | String | `"session-a369b88c-03fb-4497-a60b-495b94313210"`                                                                                                                                                                                                                    |
  </Tab>

  <Tab title="Version 1.1">
    | Field Name              | Description                                                             | Type   | Example                                        |
    | ----------------------- | ----------------------------------------------------------------------- | ------ | ---------------------------------------------- |
    | `id`                    | Unique conversation ID                                                  | Number | 1                                              |
    | `account_id`            | Account ID (where the conversation belongs)                             | Number | 1                                              |
    | `inbox_id`              | Inbox ID (where the conversation belongs)                               | Number | 1                                              |
    | `status`                | Conversation status: `{ open: 0, resolved: 1, pending: 2, snoozed: 3 }` | Number | 1                                              |
    | `assignee_id`           | Agent ID (assigned to the conversation)                                 | Number | 1                                              |
    | `created_at`            | Conversation creation date                                              | Date   | `2023-01-19T13:31:18.553Z`                     |
    | `updated_at`            | Conversation updated date                                               | Date   | `2023-01-19T13:31:18.553Z`                     |
    | `contact_id`            | Client unique identification                                            | Number | 3                                              |
    | `display_id`            | Display ID                                                              | Number | 3                                              |
    | `contact_last_seen_at`  | Contact/Client date last seen                                           | Date   | `2023-01-19T13:31:18.553Z`                     |
    | `agent_last_seen_at`    | Agent date last seen                                                    | Date   | `2023-01-19T13:31:18.553Z`                     |
    | `additional_attributes` | JSON containing any additional attributes                               | JSON   | `{}`                                           |
    | `contact_inbox_id`      | Contact Inbox ID                                                        | Number | 1                                              |
    | `uuid`                  | Unique identifier                                                       | String | `89f20666-e17d-427c-9283-268ba7870283`         |
    | `identifier`            | Identifier                                                              | String | `89f20666-e17d-427c-9283-268ba7870283`         |
    | `last_activity_at`      | Last activity date                                                      | Date   | `2023-01-19T13:31:18.553Z`                     |
    | `team_id`               | Team ID (belonging to the conversation)                                 | Number | 1                                              |
    | `snoozed_until`         | Snoozed date end                                                        | Date   | `2023-01-19T13:31:18.553Z`                     |
    | `custom_attributes`     | Extra add-on attributes                                                 | JSON   | `{}`                                           |
    | `assignee_last_seen_at` | Date of last time the assignee has been seen                            | Date   | `2023-01-19T13:31:18.553Z`                     |
    | `cognigy_session_id`    | Unique session ID                                                       | String | "session-deedb62a-8062-490f-83ac-826de9ef94S3" |
  </Tab>

  <Tab title="Version 1.0">
    | Field Name              | Description                                                                                             | Type   | Example                                        |
    | ----------------------- | ------------------------------------------------------------------------------------------------------- | ------ | ---------------------------------------------- |
    | `id`                    | Unique conversation ID                                                                                  | Number | 1                                              |
    | `account_id`            | Account ID (where the conversation belongs)                                                             | Number | 1                                              |
    | `inbox_id`              | Inbox ID (where the conversation belongs)                                                               | Number | 1                                              |
    | `status`                | Conversation status: `{ open: 0, resolved: 1, pending: 2, snoozed: 3 }`                                 | Number | 1                                              |
    | `assignee_id`           | Agent ID (assigned to the conversation)                                                                 | Number | 1                                              |
    | `created_at`            | Conversation creation date                                                                              | Date   | `2023-01-19T13:31:18.553Z`                     |
    | `updated_at`            | Conversation updated date                                                                               | Date   | `2023-01-19T13:31:18.553Z`                     |
    | `contact_id`            | Client unique identification                                                                            | Number | 3                                              |
    | `display_id`            | Display ID                                                                                              | Number | 3                                              |
    | `agent_last_seen_at`    | Agent date last seen                                                                                    | Date   | `2023-01-19T13:31:18.553Z`                     |
    | `additional_attributes` | JSON containing any additional attributes                                                               | JSON   | `{}`                                           |
    | `contact_inbox_id`      | Contact Inbox ID                                                                                        | Number | 1                                              |
    | `uuid`                  | Unique identifier                                                                                       | String | "89f20666-e17d-427c-9283-268ba7870283"         |
    | `last_activity_at`      | Last activity date                                                                                      | Date   | `2023-01-19T13:31:18.553Z`                     |
    | `team_id`               | Team ID (belonging to the conversation)                                                                 | Number | 1                                              |
    | `snoozed_until`         | Snoozed date end                                                                                        | Date   | `2023-01-19T13:31:18.553Z`                     |
    | `custom_attributes`     | Extra add-on attributes                                                                                 | JSON   | `{}`                                           |
    | `assignee_last_seen_at` | Date of last time the assignee has been seen                                                            | Date   | `2023-01-19T13:31:18.553Z`                     |
    | `cognigy_session_id`    | Unique session ID                                                                                       | String | "session-deedb62a-8062-490f-83ac-826de9ef94S3" |
    | `assigned_at`           | The date and time when the conversation was assigned to a human agent                                   | Date   | `2023-01-19T13:31:18.553Z`                     |
    | `ended_at`              | The date and time when the human agent closed the conversation (marked it as `Resolved` or `Abandoned`) | Date   | `2023-01-19T13:31:18.553Z`                     |
  </Tab>
</Tabs>

### Inbox

**Description:**

The Inbox is where all Conversations from a specific set of Agents will be placed.

**Example Query:**

<Tabs>
  <Tab title="Version 2.0">
    `https://odata-liveagent-trial.cognigy.ai/v2.0/odata/Inbox?$top=1&$skip=0`
  </Tab>

  <Tab title="Version 1.1">
    `https://odata-liveagent-trial.cognigy.ai/v1.1/odata/Inbox?$top=1&$skip=0`
  </Tab>

  <Tab title="Version 1.0">
    `https://odata-liveagent-trial.cognigy.ai/v1.0/odata/Inbox?$top=1&$skip=0`
  </Tab>
</Tabs>

**Data Types:**

When retrieving this data model, the endpoint returns the following fields:

<Tabs>
  <Tab title="Version 2.0">
    | Field Name                 | Description                                                            | Type    | Example                                      |
    | -------------------------- | ---------------------------------------------------------------------- | ------- | -------------------------------------------- |
    | `id`                       | Unique Inbox record ID                                                 | Number  | `3`                                          |
    | `channel_id`               | The channel ID this Inbox is connected to                              | Number  | `3`                                          |
    | `name`                     | Inbox name                                                             | String  | `"cucumberTest_Project_by_template-default"` |
    | `account_id`               | Account ID that owns the Inbox                                         | Number  | `2`                                          |
    | `created_at`               | Timestamp when the Inbox was created                                   | String  | `"2023-01-10T17:36:00.052Z"`                 |
    | `updated_at`               | Timestamp when the Inbox was last updated                              | String  | `"2023-01-10T17:36:00.052Z"`                 |
    | `channel_type`             | Communication channel type, such as API, Webchat, Facebook (Meta)      | String  | `"Channel::Api"`                             |
    | `enable_auto_assignment`   | Whether conversations are automatically assigned to agents             | Boolean | `true`                                       |
    | `working_hours_enabled`    | Whether working hours scheduling is enabled                            | Boolean | `false`                                      |
    | `out_of_office_message`    | Custom message displayed when outside working hours                    | String  | `null`                                       |
    | `timezone`                 | Timezone of the Inbox                                                  | String  | `"UTC"`                                      |
    | `enable_email_collect`     | Whether email collection is enabled                                    | Boolean | `true`                                       |
    | `csat_survey_enabled`      | Whether CSAT survey is enabled                                         | Boolean | `false`                                      |
    | `cognigy_organization_id`  | Cognigy Organization ID                                                | String  | `null`                                       |
    | `cognigy_project_id`       | Cognigy Project ID                                                     | String  | `"63bda1ff4688585c8fa2ccae"`                 |
    | `enable_auto_reassignment` | Whether conversations can be automatically reassigned                  | Boolean | `true`                                       |
    | `conversations_show_all`   | Whether agents can see all conversations, even if not assigned to them | Boolean | `true`                                       |
  </Tab>

  <Tab title="Version 1.1">
    | Field Name                 | Description                                                                        | Type    | Example                    |
    | -------------------------- | ---------------------------------------------------------------------------------- | ------- | -------------------------- |
    | `id`                       | Unique record ID                                                                   | Number  | 1                          |
    | `channel_id`               | Channel ID                                                                         | Number  | 1                          |
    | `name`                     | Inbox name                                                                         | String  | `Welcome to Live Agent`    |
    | `email_address`            | Email address                                                                      | String  | `ai-agent@cognigy.com`     |
    | `account_id`               | Account ID                                                                         | Number  | 1                          |
    | `created_at`               | Inbox creation date                                                                | Date    | `2023-01-19T13:31:18.553Z` |
    | `updated_at`               | Inbox updated date                                                                 | Date    | `2023-01-19T13:31:18.553Z` |
    | `channel_type`             | Communication channel, for example, API, Webchat, Facebook (Meta)                  | String  | `Channel:Webchat`          |
    | `enable_auto_assignment`   | Feature that allows the agent to auto assign conversations                         | Boolean | `true`                     |
    | `greeting_enabled`         | Feature that allows an automatic greeting to be sent                               | Boolean | `false`                    |
    | `greeting_message`         | Automatic greeting message                                                         | String  | `Hi there!`                |
    | `working_hours_enabled`    | Feature that enables a working hours timetable                                     | Boolean | `true`                     |
    | `out_of_office_message`    | Feature that enables a working hours timetable                                     | Boolean | `true`                     |
    | `timezone`                 | Inbox timezone                                                                     | String  | "UTC"                      |
    | `enable_email_collect`     | Allows email collection                                                            | Boolean | `true`                     |
    | `csat_survey_enabled`      | Allows CSAT survey                                                                 | Boolean | `true`                     |
    | `cognigy_organization_id`  | Cognigy account ID                                                                 | String  | 1                          |
    | `cognigy_project_id`       | Cognigy Project ID                                                                 | String  | 1                          |
    | `enable_auto_reassignment` | Allows conversation reassignment                                                   | Boolean | `true`                     |
    | `conversations_show_all`   | Allows you to see all conversations in the Inbox, even if they don't belong to you | Boolean | `true`                     |
  </Tab>

  <Tab title="Version 1.0">
    | Field Name                 | Description                                                                        | Type    | Example                    |
    | -------------------------- | ---------------------------------------------------------------------------------- | ------- | -------------------------- |
    | `id`                       | Unique record ID                                                                   | Number  | 1                          |
    | `channel_id`               | Channel ID                                                                         | Number  | 1                          |
    | `name`                     | Inbox Name                                                                         | String  | `Welcome to Live Agent`    |
    | `account_id`               | Account ID                                                                         | Number  | 1                          |
    | `created_at`               | Inbox creation date                                                                | Date    | `2023-01-19T13:31:18.553Z` |
    | `updated_at`               | Inbox updated date                                                                 | Date    | `2023-01-19T13:31:18.553Z` |
    | `channel_type`             | Communication channel, for example, API, Webchat, Facebook (Meta)                  | String  | `Channel:Webchat`          |
    | `enable_auto_assignment`   | Feature that allows the agent to auto assign conversations                         | Boolean | `true`                     |
    | `working_hours_enabled`    | Feature that enables a working hours timetable                                     | Boolean | `true`                     |
    | `out_of_office_message`    | Feature that enables a working hours timetable                                     | Boolean | `true`                     |
    | `timezone`                 | Inbox Timezone                                                                     | String  | `UTC`                      |
    | `enable_email_collect`     | Allows email collection                                                            | Boolean | `true`                     |
    | `csat_survey_enabled`      | Allows CSAT survey                                                                 | Boolean | `true`                     |
    | `cognigy_organization_id`  | Cognigy Account ID                                                                 | String  | 1                          |
    | `cognigy_project_id`       | Cognigy Project ID                                                                 | String  | 1                          |
    | `enable_auto_reassignment` | Allows conversation reassignment                                                   | Boolean | `true`                     |
    | `conversation_show_all`    | Allows you to see all conversations in the Inbox, even if they don't belong to you | Boolean | `true`                     |
  </Tab>
</Tabs>

### Label

**Description:**

Labels are used to mark, identify, or group different Conversations.

**Example Query:**

<Tabs>
  <Tab title="Version 2.0">
    `https://odata-liveagent-trial.cognigy.ai/v2.0/odata/Label?$top=1&$skip=0`
  </Tab>

  <Tab title="Version 1.1">
    `https://odata-liveagent-trial.cognigy.ai/v1.1/odata/Label?$top=1&$skip=0`
  </Tab>

  <Tab title="Version 1.0">
    `https://odata-liveagent-trial.cognigy.ai/v1.0/odata/Label?$top=1&$skip=0`
  </Tab>
</Tabs>

**Data Types:**

When retrieving this data model, the endpoint returns the following fields:

<Tabs>
  <Tab title="Version 2.0">
    | Field Name         | Description                                           | Type             | Example                      |
    | ------------------ | ----------------------------------------------------- | ---------------- | ---------------------------- |
    | `id`               | A unique label ID                                     | String           | `"1"`                        |
    | `title`            | The label title                                       | String           | `"qa-60"`                    |
    | `description`      | The label description                                 | String           | `"sadc"`                     |
    | `color`            | The label color in HEX format                         | String           | `"#111113"`                  |
    | `show_on_sidebar`  | Indicates whether the label is shown on the sidebar   | Boolean          | `true`                       |
    | `account_id`       | The ID of the account to which the label belongs      | String           | `"2"`                        |
    | `created_at`       | The label creation timestamp                          | Date             | `"2023-01-19T13:31:18.553Z"` |
    | `updated_at`       | The label last updated timestamp                      | Date             | `"2023-09-18T12:34:19.579Z"` |
    | `conversation_ids` | A list of conversation IDs associated with this label | Array of Numbers | `[77455, 78729, 79906, …]`   |
  </Tab>

  <Tab title="Version 1.1">
    | Field Name        | Description                                 | Type    | Example                     |
    | ----------------- | ------------------------------------------- | ------- | --------------------------- |
    | `id`              | Unique label ID                             | Number  | 1                           |
    | `title`           | Label title                                 | String  | `label-complain`            |
    | `description`     | Label description                           | String  | `labels used for complains` |
    | `color`           | Label color                                 | String  | `red`                       |
    | `show_on_sidebar` | Allows the label to be shown on the sidebar | Boolean | `true`                      |
    | `account_id`      | Account to which the label will belong      | Number  | 3                           |
    | `created_at`      | Label creation date                         | Date    | `2023-01-19T13:31:18.553Z`  |
    | `updated_at`      | Label updated date                          | Date    | `2023-01-19T13:31:18.553Z`  |
  </Tab>

  <Tab title="Version 1.0">
    | Field Name         | Description                                            | Type             | Example                     |
    | ------------------ | ------------------------------------------------------ | ---------------- | --------------------------- |
    | `id`               | Unique label ID                                        | Number           | 1                           |
    | `title`            | Label title                                            | String           | `label-complain`            |
    | `description`      | Label description                                      | String           | `labels used for complains` |
    | `color`            | Label color                                            | String           | `red`                       |
    | `show_on_sidebar`  | Allows the label to be shown on the sidebar            | Boolean          | `true`                      |
    | `account_id`       | Account to which the label will belong                 | Number           | 3                           |
    | `created_at`       | Label creation date                                    | Date             | `2023-01-19T13:31:18.553Z`  |
    | `updated_at`       | Label updated date                                     | Date             | `2023-01-19T13:31:18.553Z`  |
    | `conversation_ids` | An array of conversation IDs associated with the label | Array of numbers | `[15628, 15964, 15965]`     |
  </Tab>
</Tabs>

### Reporting Event

**Description:**

The Reporting Event model represents actions or occurrences tracked in the system for reporting and analytics purposes.

**Example Query:**

<Tabs>
  <Tab title="Version 2.0">
    `https://odata-liveagent-trial.cognigy.ai/v2.0/odata/ReportingEvent?$top=10&$skip=0`.
  </Tab>

  <Tab title="Version 1.1">
    `https://odata-liveagent-trial.cognigy.ai/v1.1/odata/ReportingEvent?$top=10&$skip=0`.
  </Tab>

  <Tab title="Version 1.0">
    `https://odata-liveagent-trial.cognigy.ai/v1.0/odata/ReportingEvent?$top=10&$skip=0`.
  </Tab>
</Tabs>

**Data Types:**

When retrieving this data model, the endpoint returns the following fields:

| Field Name         | Description                                        | Type   | Example                                                                                                             |
| ------------------ | -------------------------------------------------- | ------ | ------------------------------------------------------------------------------------------------------------------- |
| `id`               | Unique Reporting Event ID                          | Number | 4                                                                                                                   |
| `event_start_time` | Start time of the event                            | Date   | `2023-01-08T18:34:43.162Z`                                                                                          |
| `event_end_time`   | End time of the event                              | Date   | `2023-01-08T18:34:43.162Z`                                                                                          |
| `name`             | Name of the event                                  | String | `message_created`, `conversation_resolved`, `first_reply_created`, `assignee_changed`, `conversation_inbox_changed` |
| `value`            | Value associated with the event, such as timestamp | Number | 1                                                                                                                   |
| `created_at`       | The date and time when the event was created       | Date   | `2023-01-08T18:34:43.213Z`                                                                                          |
| `updated_at`       | The date and time when the event was last updated  | Date   | `2023-01-08T18:34:43.213Z`                                                                                          |
| `account_id`       | ID of the account associated with the event        | Number | 2                                                                                                                   |
| `conversation_id`  | ID of the conversation associated with the event   | Number | 1                                                                                                                   |
| `inbox_id`         | ID of the inbox associated with the event          | Number | 1                                                                                                                   |
| `user_id`          | ID of the user associated with the event           | Number | 3                                                                                                                   |

<Note>
  To avoid potential query timeouts due to large data volumes, use the `$top` query option to limit the number of returned records.
</Note>

### User

**Description:**

This data model stores information about Live Agent users.

**Example Query:**

<Tabs>
  <Tab title="Version 2.0">
    `https://odata-liveagent-trial.cognigy.ai/v2.0/odata/User?$top=1&$skip=0`
  </Tab>

  <Tab title="Version 1.1">
    `https://odata-liveagent-trial.cognigy.ai/v1.1/odata/User?$top=1&$skip=0`
  </Tab>

  <Tab title="Version 1.0">
    `https://odata-liveagent-trial.cognigy.ai/v1.0/odata/User?$top=1&$skip=0`
  </Tab>
</Tabs>

**Data Types:**

When retrieving this data model, the endpoint returns the following fields:

<Tabs>
  <Tab title="Version 2.0">
    | Field Name             | Description                                                    | Type   | Example                                                                                                         |
    | ---------------------- | -------------------------------------------------------------- | ------ | --------------------------------------------------------------------------------------------------------------- |
    | `id`                   | A unique record ID                                             | Number | `1`                                                                                                             |
    | `provider`             | The provider, for example, email                               | String | `email`                                                                                                         |
    | `uid`                  | A provided unique identifier                                   | String | `some-email@cognigy.com`                                                                                        |
    | `confirmed_at`         | The date when the confirmation token was confirmed             | Date   | `2021-12-11 12:41:26.745`                                                                                       |
    | `confirmation_sent_at` | The date when the confirmation token was sent                  | Date   | `2021-12-11 12:41:26.745`                                                                                       |
    | `unconfirmed_email`    | An unconfirmed email                                           | String | `unconfirmed@cognigy.com`                                                                                       |
    | `name`                 | The user's name                                                | String | `Francisco`                                                                                                     |
    | `display_name`         | The user's display name                                        | String | `Paco`                                                                                                          |
    | `email`                | The user's email                                               | String | `fran@cognigy.com`                                                                                              |
    | `tokens`               | The user's tokens                                              | JSON   | `{"MN4iNP38tvwe5wer7pvpnAQ":{"token":"$2a$10$iA7Si0sBV9lMkO2mqvwiD.L90.uTCl27WQbEo/Kc0W","expiry":1644763988}}` |
    | `created_at`           | The date when the user was created                             | Date   | `2021-12-11 12:41:26.745`                                                                                       |
    | `updated_at`           | The date when the user was last updated                        | Date   | `2021-12-11 12:41:26.745`                                                                                       |
    | `pubsub_token`         | The user's publish/subscribe token                             | String | `MiVxY3FKry6oklambiuqL72nn`                                                                                     |
    | `availability`         | The user's availability                                        | Number | 1                                                                                                               |
    | `custom_attributes`    | The user's custom attributes                                   | JSON   | `{}`                                                                                                            |
    | `cognigy_user_id`      | The Cognigy user ID                                            | String | `610bEOk617uPiXy9L3HOq8nn6kdjJK2`                                                                               |
    | `sign_in_count`        | The total number of times the user has logged in to Live Agent | Number | 3                                                                                                               |
    | `current_sign_in_at`   | The date and time of the user's most recent login              | Date   | `2021-12-11 12:41:26.745`                                                                                       |
    | `last_sign_in_at`      | The date and time of the user's last login                     | Date   | `2021-12-11 12:41:26.745`                                                                                       |
    | `last_sign_out_at`     | The date and time of the user's last logout from Live Agent    | Date   | `2021-12-11 12:41:26.745`                                                                                       |
  </Tab>

  <Tab title="Version 1.1">
    | Field Name               | Description                                                       | Type   | Example                                                                                                         |
    | ------------------------ | ----------------------------------------------------------------- | ------ | --------------------------------------------------------------------------------------------------------------- |
    | `id`                     | Unique record ID                                                  | Number | 1                                                                                                               |
    | `provider`               | For example, email                                                | String | `email`                                                                                                         |
    | `uid`                    | Provided unique identifier                                        | String | `some-email@cognigy.com`                                                                                        |
    | `encrypted_password`     | Encrypted password                                                | String | `$2a$11$L.610bE0k617uÜiXy9L3H0q8nn6kdjJK2/X4p7ghlxPC38hjovasd`                                                  |
    | `reset_password_token`   | Token used to reset the password                                  | String | `$2a$11$L.610bE0k617uÜiXy9L3H0q8nn6kdjJK2/X4p7ghlxPC38hjsvaru`                                                  |
    | `reset_password_send_at` | Date when a password reset was sent                               | Date   | `2021-12-11 12:41:26.745`                                                                                       |
    | `remember_created_at`    | Date of remember creation                                         | Date   | `2021-12-11 12:41:26.745`                                                                                       |
    | `sign_in_count`          | Number of times the user has signed in                            | Number | 3                                                                                                               |
    | `current_sign_in_ip`     | Current sign in IP address                                        | String | `120.0.0.1`                                                                                                     |
    | `last_sign_in_ip`        | Last sign in IP address                                           | String | `120.0.0.1`                                                                                                     |
    | `confirmation_token`     | Confirmation token                                                | String | `$2a$11$L.610bE0k617uÜiXy9L3H0q8nn6kdjJK2/X4p7ghlxPC38hjovasd`                                                  |
    | `confirmed_at`           | Confirmation token confirmed at date                              | Date   | `2021-12-11 12:41:26.745`                                                                                       |
    | `confirmation_sent_at`   | Confirmation token sent at date                                   | Date   | `2021-12-11 12:41:26.745`                                                                                       |
    | `unconfirmed_email`      | Unconfirmed email                                                 | String | `unconfirmed@cognigy.com`                                                                                       |
    | `name`                   | User name                                                         | String | `Francisco`                                                                                                     |
    | `display_name`           | User display name                                                 | String | `Paco`                                                                                                          |
    | `email`                  | User email                                                        | String | `fran@cognigy.com`                                                                                              |
    | `tokens`                 | User tokens                                                       | JSON   | `{"MN4iNP38tvwe5wer7pvpnAQ":{"token":"$2a$10$iA7Si0sBV9lMkO2mqvwiD.L90.uTCl27WQbEo/Kc0W","expiry":1644763988}}` |
    | `created_at`             | User creation date                                                | Date   | `2021-12-11 12:41:26.745`                                                                                       |
    | `updated_at`             | User updated date                                                 | Date   | `2021-12-11 12:41:26.745`                                                                                       |
    | `pubsub_token`           | User publish/subscribe token                                      | String | `MiVxY3FKry6oklambiuqL72nn`                                                                                     |
    | `availability`           | This represents the user's availability                           | Number | 1                                                                                                               |
    | `custom_attributes`      | Custom attributes                                                 | JSON   | `{}`                                                                                                            |
    | `cognigy_user_id`        | Cognigy user ID                                                   | string | `610bEOk617uPiXy9L3HOq8nn6kdjJK2`                                                                               |
    | `sign_in_count`          | The total number of times the user has logged in to Live Agent    | Number | 3                                                                                                               |
    | `current_sign_in_at`     | The date and time of the user's most recent login                 | Date   | `2021-12-11 12:41:26.745`                                                                                       |
    | `last_sign_in_at`        | The date and time of the user's login previous to the current one | Date   | `2021-12-11 12:41:26.745`                                                                                       |
  </Tab>

  <Tab title="Version 1.0">
    | Field Name             | Description                                                       | Type   | Example                                                                                                         |
    | ---------------------- | ----------------------------------------------------------------- | ------ | --------------------------------------------------------------------------------------------------------------- |
    | `id`                   | Unique record ID                                                  | Number | 1                                                                                                               |
    | `provider`             | For example, email                                                | String | `email`                                                                                                         |
    | `uid`                  | Provided unique identifier                                        | String | `some-email@cognigy.com`                                                                                        |
    | `confirmed_at`         | Confirmation token confirmed at date                              | Date   | `2021-12-11 12:41:26.745`                                                                                       |
    | `confirmation_sent_at` | Confirmation token sent at date                                   | Date   | `2021-12-11 12:41:26.745`                                                                                       |
    | `unconfirmed_email`    | Unconfirmed email                                                 | String | `unconfirmed@cognigy.com`                                                                                       |
    | `name`                 | User name                                                         | String | `Francisco`                                                                                                     |
    | `display_name`         | User display name                                                 | String | `Paco`                                                                                                          |
    | `email`                | User email                                                        | String | `fran@cognigy.com`                                                                                              |
    | `tokens`               | User tokens                                                       | JSON   | `{"MN4iNP38tvwe5wer7pvpnAQ":{"token":"$2a$10$iA7Si0sBV9lMkO2mqvwiD.L90.uTCl27WQbEo/Kc0W","expiry":1644763988}}` |
    | `created_at`           | User creation date                                                | Date   | `2021-12-11 12:41:26.745`                                                                                       |
    | `updated_at`           | User updated date                                                 | Date   | `2021-12-11 12:41:26.745`                                                                                       |
    | `pubsub_token`         | User publish/subscribe token                                      | String | `MiVxY3FKry6oklambiuqL72nn`                                                                                     |
    | `availability`         | This represents the user's availability                           | Number | 1                                                                                                               |
    | `custom_attributes`    | Custom attributes                                                 | JSON   | `{}`                                                                                                            |
    | `cognigy_user_id`      | Cognigy user ID                                                   | string | `610bEOk617uPiXy9L3HOq8nn6kdjJK2`                                                                               |
    | `sign_in_count`        | The total number of times the user has logged in to Live Agent    | Number | 3                                                                                                               |
    | `current_sign_in_at`   | The date and time of the user's most recent login                 | Date   | `2021-12-11 12:41:26.745`                                                                                       |
    | `last_sign_in_at`      | The date and time of the user's login previous to the current one | Date   | `2021-12-11 12:41:26.745`                                                                                       |
    | `last_sign_out_at`     | The date and time of the user's last logout from Live Agent       | Date   | `2021-12-11 12:41:26.745`                                                                                       |
  </Tab>
</Tabs>

### Team

**Description:**

This data model stores information about Live Agent teams.

**Example Query:**

<Tabs>
  <Tab title="Version 2.0">
    `https://odata-liveagent-trial.cognigy.ai/v2.0/odata/Team?$top=1&$skip=0`
  </Tab>
</Tabs>

**Data Types:**

When retrieving this data model, the endpoint returns the following fields:

| Field Name                     | Description                                                     | Type    | Example                      |
| ------------------------------ | --------------------------------------------------------------- | ------- | ---------------------------- |
| `id`                           | The unique ID of the team                                       | String  | `"9"`                        |
| `name`                         | The name of the team                                            | String  | `"customer support team"`    |
| `description`                  | A description of the team                                       | String  | `"for urgent user requests"` |
| `allow_auto_assign`            | Whether conversations can be automatically assigned to the team | Boolean | `true`                       |
| `account_id`                   | The account ID the team belongs to                              | String  | `"2"`                        |
| `created_at`                   | The timestamp when the team was created                         | String  | `"2025-12-08 12:41:26.745"`  |
| `updated_at`                   | The timestamp when the team was last updated                    | String  | `"2025-12-08 12:41:26.745"`  |
| `members`                      | An array of user entities                                       | Array   | `[...]`                      |
| `members.id`                   | The member ID (internal Live Agent ID)                          | Number  | `57`                         |
| `members.provider`             | The authentication provider                                     | String  | `"cognigy"`                  |
| `members.uid`                  | The user UID from the provider                                  | String  | `"63ca7c200d733defd7512584"` |
| `members.confirmed_at`         | The timestamp when the user confirmed their account             | String  | `"2025-12-08 12:41:26.745"`  |
| `members.confirmation_sent_at` | The timestamp when the confirmation was sent                    | String  | `null`                       |
| `members.unconfirmed_email`    | The email pending confirmation                                  | String  | `null`                       |
| `members.name`                 | The full name of the user                                       | String  | `"John Doe"`                 |
| `members.display_name`         | The display name used in the UI                                 | String  | `"John"`                     |
| `members.email`                | The email address                                               | String  | `"john.doe@example.com"`     |
| `members.tokens`               | Authentication tokens                                           | String  | `null`                       |
| `members.created_at`           | The timestamp when the user was created                         | String  | `"2025-12-08 12:41:26.745"`  |
| `members.updated_at`           | The timestamp when the user was last updated                    | String  | `"2025-12-08 12:41:26.745"`  |
| `members.pubsub_token`         | The token for real-time updates                                 | String  | `"a5DXd5Q84djArRReTNfavEZk"` |
| `members.availability`         | The user availability status (`0 = offline`)                    | Number  | `0`                          |
| `members.custom_attributes`    | Custom attributes in JSON format                                | String  | `"{}"`                       |
| `members.cognigy_user_id`      | The user ID from Cognigy                                        | String  | `"63ca7c200d733eefd7512584"` |
| `members.sign_in_count`        | The number of times the user signed in                          | Number  | `17`                         |
| `members.current_sign_in_at`   | The timestamp of the current login session                      | String  | `"2025-12-08 12:41:26.745"`  |
| `members.last_sign_in_at`      | The timestamp of the last login                                 | String  | `"2025-12-08 12:41:26.745"`  |
| `members.last_sign_out_at`     | The timestamp of the last logout                                | String  | `"2025-12-08 12:41:26.745"`  |

### AgentAvailabilityRecord

**Description:**

This data model stores information about agent availability in Live Agent.

**Example Query:**

<Tabs>
  <Tab title="Version 2.0">
    `https://odata-liveagent-trial.cognigy.ai/v2.0/odata/AgentAvailabilityRecord?$top=1&$skip=0`
  </Tab>
</Tabs>

**Data Types:**

When retrieving this data model, the endpoint returns the following fields:

| Field Name     | Description                                                                           | Type            | Example                      |
| -------------- | ------------------------------------------------------------------------------------- | --------------- | ---------------------------- |
| `id`           | The unique ID of the availability record                                              | String          | `"1"`                        |
| `availability` | The agent's availability status (`0 = online`, `1 = offline`, `2 = busy`, `3 = away`) | Number          | `0`                          |
| `created_at`   | The timestamp when the record was created                                             | ISO 8601 String | `"2025-12-03T09:22:32.760Z"` |
| `updated_at`   | The timestamp when the record was last updated                                        | ISO 8601 String | `"2025-12-03T09:22:32.760Z"` |
| `account_id`   | The ID of the account the agent belongs to                                            | String          | `"2"`                        |
| `user_id`      | The ID of the user (agent)                                                            | String          | `"8953"`                     |

## Integrations

<div />

### Excel

When connecting from Microsoft Excel 2016, you must use the Power Query feature, which can be found under **Data > Get & Transform > New Query > From Other Sources > From OData Feed**. This will connect to our OData 4 feed.

### Power BI

Follow the [instructions in the Power BI documentation](https://docs.microsoft.com/en-us/power-bi/desktop-connect-odata).

### Tableau

<Danger>
  Cognigy supports OData 4.0, which implies that certain versions of Tableau aren't compatible.
</Danger>

Find instructions on how to connect an OData Feed in the [Tableau documentation](https://help.tableau.com/current/pro/desktop/en-gb/examples_odata.htm).

## OData Consumer Ecosystem

For a full list of available OData Consumer options, [follow the link to Consumers on OData.org](https://www.odata.org/ecosystem/#consumers).

## Client Libraries in .NET, Java, JavaScript, C++, and Other Platforms

For a full list of available OData Libraries, [see the latest directory of available libraries on OData.org](https://www.odata.org/libraries/).

***

<sup>1</sup>: This parameter is required to make a request to the Live Agent OData endpoint.


# Key Concepts
Source: https://docs.cognigy.com/ops-center/concepts

Learn about the key concepts in Ops Center, including components, subcomponents, metrics, alerts, and errors.

This article provides an overview of the key concepts in Ops Center.

<Frame>
  <img alt="Ops Center key concepts showing components, metrics, alerts, and errors" />
</Frame>

## Components

A *Component* is a Cognigy product, feature, resource, or service that can be part of the conversation lifecycle. Ops Center tracks components to provide information about the health of your AI Agents operation. Ops Center tracks the following components:

* [Voice Gateway](/voice-gateway/overview)
* [Endpoints](/ai/agents/deploy/endpoints/overview)
* [Flows](/ai/agents/develop/projects-and-flows/overview)
* [Handover Providers](/ai/escalate/handovers)

## Subcomponents

A *Subcomponent* is a feature or resource that composes or is used in a [component](#components). Ops Center also tracks subcomponents to provide information about the health of your AI Agents operation. The following table gives you an overview of the subcomponents that Ops Center tracks for each component:

<Accordion title="Voice Gateway">
  | Subcomponent                                                                              | Description                 |
  | ----------------------------------------------------------------------------------------- | --------------------------- |
  | [Speech-to-Text (STT)](/voice-gateway/webapp/speech-services#cloud-based-speech-services) | Messages processed with STT |
  | [Text-to-Speech (TTS)](/voice-gateway/webapp/speech-services#cloud-based-speech-services) | Messages processed with TTS |
</Accordion>

<Accordion title="Endpoints">
  | Subcomponent                                                                                      | Description                                                                |
  | ------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------- |
  | [Input transformer](/ai/for-developers/transformers/input-transformer)                            | Requests made to the input transformer.                                    |
  | [Output transformer](/ai/for-developers/transformers/output-transformer)                          | Requests made to the output transformer.                                   |
  | [Execution finished transformer](/ai/for-developers/transformers/execution-finished-transformer)  | Requests made to the execution finished transformer.                       |
  | [Inject transformer](/ai/for-developers/transformers/inject-transformer)                          | Requests made to the inject transformer.                                   |
  | [Notify transformer](/ai/for-developers/transformers/notify-transformer)                          | Requests made to the notify transformer.                                   |
  | [Real-time translation](/ai/platform-features/translation-and-localization/real-time-translation) | Requests made to a machine translation provider for real-time translation. |
</Accordion>

<Accordion title="Flows">
  | Subcomponent                                                               | Description                                                |
  | -------------------------------------------------------------------------- | ---------------------------------------------------------- |
  | Message Roundtrip                                                          | Volume of user input.                                      |
  | Flow execution                                                             | Flow execution time.                                       |
  | [Natural Language Understanding (NLU)](/ai/platform-features/nlu/overview) | NLU scoring time.                                          |
  | Outbound API calls                                                         | Requests made to external APIs and services from the Flow. |
  | [Extensions](/ai/for-developers/extensions)                                | Extension executions and processing time.                  |
  | [Large Language Models](/ai/agents/develop/gen-ai-and-llms/llms)           | LLM fallbacks and retries.                                 |
  | [Knowledge AI queries](/ai/agents/develop/knowledge-ai/overview) (KAI)     | Knowledge AI queries and KAI query latency.                |
</Accordion>

<Accordion title="Handover Providers">
  | Subcomponent                                                                                                      | Description                                                        |
  | ----------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------ |
  | [Live Agent](/live-agent/overview) and [third-party handover providers](/ai/escalate/handover-reference/overview) | Number of handover errors and API request durations for handovers. |
</Accordion>

## Metrics

*Metrics* are measurable and calculable values from [components](#components) and [subcomponents](#subcomponents) over time. You can use the metrics in Ops Center to monitor and evaluate the performance of your AI Agents operation. Ops Center updates them regularly on the dashboards. The following table gives you an overview of the metrics that Ops Center tracks for each component:

| Component          | Metrics                                                                                                                                              |
| ------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------- |
| Voice Gateway      | Voice input/min                                                                                                                                      |
| Endpoints          | Request/min, Endpoint traffic, transformer processing time, message processing time                                                                  |
| Flows              | 90th percentile processing time, Knowledge AI queries, Knowledge AI query latency, NLU scoring time, Extensions execution, Extension processing time |
| Handover Providers | Provider forwarding errors, API call request duration                                                                                                |

## Alerts

Ops Center shows alerts when the metrics of your AI Agents operation exceeds specific thresholds for a specific duration. You can [manage alerts](/ops-center/manage/alerts) to display them in the dashboard of the respective component tab and trigger email notifications to users about the cause of the alerts.

Alerts have the following statuses:

* **inactive** — the metric is below the threshold and no alert is displayed.
* **active** — the metric exceeds the threshold for a longer duration than the threshold duration. This leads to a red alert in Ops Center.

## Errors

Ops Center shows the errors that occur in components and subcomponents, and provides an overview of the errors on the respective component tab.


# Getting Started with AI Ops Center
Source: https://docs.cognigy.com/ops-center/getting-started

Learn how to create an AI Agent, monitor its performance metrics, track errors, and enable alerts in Ops Center.

In this tutorial, you'll create an AI Agent, explore its performance metrics, monitor errors, and enable alerts in Ops Center.

## Prerequisites

* By default, only the Overview page is available for viewing in Ops Center. To get a license and enable full access, contact your Customer Success representative.
* Assign the `admin` or `opsCenterUser` role to the user.

## Create a Flow

1. In the Cognigy interface, go to the Project and create a Flow in it.
2. In the Flow editor, add a Say Node.
3. In the Node editor, enter `You said {{input.text}}` in the **Text** field. Save changes.

## Create an Endpoint

1. In the left-side menu of the Project, go to **Deploy > Endpoints**.
2. On the **Endpoints** page, click **+ New Endpoint**.
3. In the **New Endpoint** section, do the following:
   1. Select the **Rest** Endpoint type and give your Endpoint a unique name.
   2. Select the previously created Flow from the list. Save changes.
4. In the Endpoint editor, copy the URL from the **Endpoint URL** field.

## Test the Flow

1. For testing purposes, use [Postman](https://web.postman.co/). In the Postman interface, create a `POST` request.

2. Paste the Endpoint URL you copied earlier into the **Request URL** field.

3. In the **Headers** section, enter the `Content-Type` key and its `application/json` value.

4. In the **Body** section, select the **raw** option and ensure the format is set to `JSON`.

5. Copy the following JSON and paste it into the **Body** field:

   ```json theme={null}
   {
     "userId": "uniqueUserId",
     "sessionId": "uniqueSessionId",
     "text": "Hello!",
     "channel": "webchat", 
     "data": {}
   }
   ```

6. Send the request. To send multiple requests in a loop, use the Postman Collection Runner.

7. In the request body, replace the `text` value with the `{{message}}`:

   ```json theme={null}
   {
     "userId": "testUser123",
     "sessionId": "testSession123",
     "text": "{{message}}",
     "channel": "webchat"
   }
   ```

   Save the request in a Collection.

8. Upload the [CSV](https://docs.cognigy.com/_assets/ops-center/test-dataset.csv) to the Collection Runner and set the iteration count.

9. Click **Run** to send multiple messages in a loop.

## View Metrics

1. In the Cognigy interface, navigate to the Admin Center using one of the following options:
   * Click **User Menu > Admin Center** on the **Main** page.
   * Select **Admin Center** from the User Menu in the bottom-left corner of the **Project** page.
2. In the left-side menu, click **Ops Center**. The **Overview** dashboard opens by default, where you can monitor key metrics.
3. Inspect the **Messages** chart. This chart provides insights into user interactions, showing the volume of user inputs sent to the system.
   Look at the peak in the chart, which indicates spikes in end-user activity. This information helps identify periods of high engagement, measured by the number of user inputs processed per minute. This metric is useful for tracking the overall system load.

   <Frame>
     <img alt="Messages chart showing user input volume per minute with activity spikes" />
   </Frame>

### Explore Endpoint Metrics

Next, go to the **Endpoints** tab to view detailed performance metrics for your Endpoints.

Focus on the following charts for insights into Endpoint performance:

<Accordion title="Endpoint Traffic">
  This chart tracks the volume of traffic (requests) handled by each Endpoint. A sudden increase or decrease can indicate changes in system demand or issues with Endpoint availability.

  Look at the peak in the chart, which indicates end-user activity in the Endpoints. This means that 38 user inputs were sent through the Endpoints per minute.

  <Frame>
    <img alt="Endpoint traffic chart showing request volume per Endpoint" />
  </Frame>
</Accordion>

<Accordion title="Message Processing Time">
  This chart tracks the time it takes to process each user input at the Endpoint. Monitoring this metric helps identify delays in message handling, potentially pointing to performance bottlenecks in your Flows or external services.

  The chart shows that at 9:03, there were 38 user inputs with processing times falling within the 100–500 millisecond time range.

  <Frame>
    <img alt="Message processing time chart showing latency distribution per Endpoint" />
  </Frame>
</Accordion>

### Explore Flow Performance Metrics

After examining the Endpoints, go to the **Flows** tab to analyze how your Flows are performing in terms of execution and natural language understanding.

Pay attention to the following charts:

<Accordion title="Flow Execution Time (90th Percentile)">
  This chart shows the distribution of Flow execution times, focusing on the 90th percentile. The 90th percentile indicates the time below which 90% of your Flows finish. In this case, the 90th percentile is 193 milliseconds (ms), meaning 90% of Flows completed in 193 ms or less, which suggests good performance.

  <Frame>
    <img alt="Flow execution time chart showing 90th percentile latency" />
  </Frame>
</Accordion>

<Accordion title="NLU Scoring Time">
  This chart tracks the time taken by the Natural Language Understanding (NLU) component to score user inputs. Longer scoring times may indicate that the NLU model is struggling to process certain inputs efficiently, potentially due to complex queries or system load.

  The chart shows that at 9:03, there were 28 user inputs with scoring times falling within the 50–100 millisecond time range.

  <Frame>
    <img alt="NLU scoring time chart showing intent recognition latency distribution" />
  </Frame>
</Accordion>

## Manage Alerts

1. In the Ops Center interface, go to the **Overview** dashboard to set up alerts.
2. Select the **Alerts** tab and click **Manage Alerts** to customize your alert preferences.
3. On the **Alerts** tab, activate the **HighMessageVolume** alert, which triggers when message volume exceeds a set threshold.
4. On the **Notification** tab, set up notifications by adding your email address, ensuring you are notified when the alert is triggered. Apply changes.
5. Go to the Postman Collection Runner and send multiple messages in a loop to increase the load on the server. If the system is receiving 50% more messages in the current 5-minute interval compared to the previous one, an alert will appear in the **Overview** dashboard. Also, a notification will be sent to the specified email.

## What's Next?

* Learn the [key concepts](/ops-center/concepts) of Ops Center.
* Familiarize yourself with the [specific metrics](/ops-center/monitor/metrics/overview) for each service.


# AI Ops Center Glossary
Source: https://docs.cognigy.com/ops-center/glossary

Glossary of terms used in Cognigy AI Operations Center (AI Ops Center).

AI Ops Center terms listed in alphabetical order. All terms in the glossary are provided for informational purposes.

## A

**Cognigy AI Operations Center**
A Cognigy product, also known as AI Ops Center or Ops Center, that provides a comprehensive view of the health of your AI Agents operations in dashboards, charts, and panels.

## C

**Chart**

A visual representation of data that displays values and trends over time. The Ops Center uses various chart types such as bars, lines, and heat maps to present data.

**Component**

A Cognigy product, feature, resource, or service that the Ops Center monitors, for example, Voice Gateway or Endpoints.

## D

**Dashboard**

A visual interface that provides a real-time, organized view of the performance of your AI Agents. Dashboards include charts and panels.

## I

**Indicator**

A chart type that shows a single number, representing a total or aggregate value.

## M

**Metric**

Measurable and calculable values from [components](#c) and [subcomponents](#s) over time, for example, voice inputs, requests, messages, and conversations per minute, which can be used for graphical representations. In Ops Center, metrics are updated regularly.

## P

**Panel**

A UI element that displays indicators and tiles in the Overview Dashboard.

## S

**Subcomponent**

A subcomponent is a feature or resource that composes or is used in a component, for example, speech-to-text (STT) and the Live Agent handover provider.

## T

**Threshold Duration**

The duration of time that a metric must exceed a threshold to trigger an alert.

**Tile**

A user interface (UI) element that displays charts and indicators.

**Trend**

A relative measure of the current metric value against the last metric value, expressed in percentage values, and displayed in panels and charts.


# Cognigy AI Operations Center
Source: https://docs.cognigy.com/ops-center/index



<div>
  <div>
    *Cognigy AI Operations Center (AI Ops Center or Ops Center)* is a monitoring platform for Cognigy AI Agents.

    <p>
      It allows you to track the performance of AI Agents by monitoring resources, errors, and conversations.
    </p>

    <p>
      You can use Ops Center to gain insights into how your AI Agents are performing and identify where issues might lie.
    </p>
  </div>

  <div>
    <div>
      <video aria-label="Cognigy AI Ops Center hero animation featuring a live chart with metrics that continuously increase.">
        <source type="video/mp4" />
      </video>

      <button type="button" aria-label="Pause video">
        ⏸
      </button>
    </div>
  </div>
</div>

<CardGroup>
  <Card title="About AI Ops Center" icon="info" href="/ops-center/overview">
    Explore Ops Center features, including how to get started and key concepts.
  </Card>

  <Card title="Monitor" icon="chart-line" href="/ops-center/monitor/metrics/overview">
    Monitor the health of Cognigy products with metrics, errors, and alerts.
  </Card>

  <Card title="Manage" icon="package" href="/ops-center/manage/alerts">
    Manage alerts, resolve errors, and assign user roles for effective operations.
  </Card>

  <Card title="Troubleshoot" icon="bug" href="/ops-center/troubleshooting">
    Troubleshoot common issues in your Ops Center to maintain smooth performance.
  </Card>
</CardGroup>


# Manage Alerts
Source: https://docs.cognigy.com/ops-center/manage/alerts

Learn how to activate alerts and set up notifications in Ops Center to monitor your AI Agents' operations.

<a href="/release-notes/2026.17"><Badge>Updated in 2026.17</Badge></a>

The alert system is deactivated by default.
To use this system, select the [alert](/ops-center/monitor/alerts) you want to activate.
You can activate alerts for all components, but component-specific activation is not available.

<Accordion title="1. Activate Alerts">
  1. In the Ops Center interface, go to the **Overview** tab.
  2. In the upper-left corner, select **Alerts**.
  3. On the **Alerts** page, click **Manage Alerts**.
  4. On the **Alerts** tab, select alerts that you want to activate and apply changes.

  Applying changes, whether you are activating or deactivating alerts, can take up to 15 minutes. Wait this time before checking the Alerts list to confirm your changes.
</Accordion>

<Accordion title="2. Send Notifications">
  You can send notifications either per email or via webhooks. On the **Notifications** tab, select one of the following options from the **Notification Method** list:

  <Tabs>
    <Tab title="Email">
      1. Enter the email addresses of the users who should receive notifications when an alert is triggered.
      2. Enter one email per field. Note that you can add only 5 email addresses. Apply the changes.
    </Tab>

    <Tab title="Webhook">
      1. **Webhook URL** — enter the URL of the service that must receive notifications when an alert is triggered. The URL must start with `http://` or `https://`. If the URL is invalid or uses an unsupported scheme, such as `ftp://`, an error is displayed and you can't save changes.
      2. **Webhook Connection** — select a webhook connection or create a new connection by clicking <img alt="add webhook connection" /> and configuring the following fields:
         * **Connection name** — enter a name for the webhook connection.
         * **Username** — enter a username to authenticate the webhook URL with the service.
         * **Password** — enter a password to authenticate the webhook URL with the service.
      3. Apply the changes.
    </Tab>
  </Tabs>
</Accordion>


# Resolve Errors
Source: https://docs.cognigy.com/ops-center/manage/errors

Learn how to resolve errors in Ops Center after investigating and fixing them.

Once you investigate the [error](/ops-center/monitor/errors) and fix it, you can remove it from the error list in your Ops Center.

Errors are retained for 7 days. If an error doesn't recur within this period, it is automatically marked as resolved.

To resolve an error manually before the 7-day deadline:

1. In the Ops Center interface, go to the tab where the error was found.
2. Click <img alt="menu" /> next to the error and select **Mark as Resolved**.


# Assign Roles
Source: https://docs.cognigy.com/ops-center/manage/roles

Learn how to assign roles to users so they can access Ops Center.

Only users with the `admin` and `opsCenterUser` roles can access Ops Center.
To assign this role to a user, go to **Access Control** in the [Admin Center](/ai/administer/access/admin-center/overview).
For more information about the global roles, read the [Access Control](/ai/administer/access/admin-center/access-control#before-you-begin) article.


# Alerts
Source: https://docs.cognigy.com/ops-center/monitor/alerts

Learn about the alerts available in Ops Center to monitor the performance and reliability of your Cognigy system.

You can enable alerts for all components.

<Frame>
  <img alt="Ops Center Alerts page showing available alert configurations" />
</Frame>

In Ops Center, the following alerts are available:

| **Alert**                                  | **Description**                                                                                                                                                                        | **Severity**                                                          | **Recommended Action**                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
| ------------------------------------------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Sudden message volume increase             | The system is receiving 50% more messages in the current 5-minute time interval compared to the previous 5-minute time interval.                                                       | Potential system strain during peak usage.                            | Open the [Messages](/ops-center/monitor/metrics/overview) chart on the **Overview** dashboard and the [Endpoints](/ops-center/monitor/metrics/endpoints) dashboard to view the Endpoint traffic. Monitor system load, Flow latency, and ensure performance remains stable.                                                                                                                                                                                                                                                                                                                                                                                 |
| Sudden message volume decrease             | The system is receiving 95% fewer messages in the current 5-minute interval compared to the previous 5-minute interval.                                                                | Possible downtime, channel issue, or low traffic.                     | Open the [Messages](/ops-center/monitor/metrics/overview) chart on the **Overview** dashboard and the [Endpoints](/ops-center/monitor/metrics/endpoints) dashboard to view the Endpoint traffic. Review the [logs](/ai/agents/test/logs), check Endpoint settings, integrations with third-party platforms, and traffic sources.                                                                                                                                                                                                                                                                                                                           |
| Increased transformer error rate           | The number of errors thrown by the Endpoint Transformers has exceeded 100 in the last 5 minutes.                                                                                       | Message transformations may fail or degrade in quality.               | Open the [Endpoints](/ops-center/monitor/metrics/endpoints) dashboard to view transformer metrics. Review the [logs](/ai/agents/test/logs) and verify the Transformer settings in Endpoints.                                                                                                                                                                                                                                                                                                                                                                                                                                                               |
| Increased outbound HTTP call error rate    | The number of errors thrown by API calls made through the HTTP Request Node has exceeded 50 in the last 5 minutes.                                                                     | Data loss, delayed Flows, or unavailable functionality for end users. | Open the [Flows](/ops-center/monitor/metrics/flows) dashboard to view outbound HTTP request metrics. Review the [logs](/ai/agents/test/logs), verify the HTTP Request Node configurations in your Flows and the availability of the external service's API.                                                                                                                                                                                                                                                                                                                                                                                                |
| Increased LLM error rate                   | The number of errors thrown during Large Language Model prompt execution has exceeded 100 in the last 5 minutes. Error count includes automatic retries (up to 3 attempts per prompt). | Failed executions may affect response quality.                        | Open the [Flows](/ops-center/monitor/metrics/flows) dashboard to view outbound LLM-related metrics. Review models configured for LLM-based Flows. Common issues include: <ul><li>**LLM configuration issues** – adjust the LLM settings if misconfiguration is evident in the logs or error messages.</li><li>**Rate limiting by the LLM provider** – contact the LLM provider to verify usage limits or quota issues.</li></ul>If the issue can't be resolved, contact Cognigy Product Support for further investigation.                                                                                                                                 |
| Increased Knowledge AI query error rate    | The number of failed knowledge search operations has exceeded 100 in the last 5 minutes.                                                                                               | Knowledge retrieval may fail.                                         | Open the [Flows](/ops-center/monitor/metrics/flows) dashboard to view Knowledge AI metrics. Check the embedding model configured for the Knowledge AI feature. Common issues include: <ul><li>**Rate limiting by the LLM provider** – check logs for details. If you are using an on-premises installation, inspect logs for `service-search-orchestrator`. For SaaS installations, contact Cognigy technical support.</li><li>**Invalid LLM credentials** – use the **Test** button in the LLM resource to verify credentials.</li><li>**Cognigy internal communication issues** – contact Cognigy technical support for further investigation.</li></ul> |
| Increased real-time translation error rate | The number of real-time translation errors has exceeded 100 in the last 5 minutes.                                                                                                     | Translations may fail.                                                | Check connections to translation providers:<ul><li>**At the project level** – verify credentials, use the **Test** button in **Manage > Settings** of your Projects.</li><li>**At the Endpoint level** – go to the Real-Time Translation settings in Endpoints. For some providers, such as DeepL Translate Pro, ensure you enter the correct values for fields like Formality or Glossary ID.</li></ul>                                                                                                                                                                                                                                                   |
| Increased NLU error rate                   | The number of failed NLU processing operations has exceeded 100 in the last 5 minutes.                                                                                                 | Intent recognition may fail.                                          | Open the [Flows](/ops-center/monitor/metrics/flows) dashboard to view NLU metrics. Review the [logs](/ai/agents/test/logs), verify the [Intent rules](/ai/platform-features/nlu/intents/rules) and [conditions](/ai/platform-features/nlu/intents/conditions), and check your NLU connector, if you have an external NLU connector. If the issue can't be resolved, contact [Cognigy Product Support](/help/get-help) for further investigation.                                                                                                                                                                                                           |
| Increased handover error rate              | The number of failed handover requests has exceeded 20 in the last 5 minutes.                                                                                                          | Handover to human agents may fail.                                    | Open the [Handover Providers](/ops-center/monitor/metrics/handover-providers) dashboard to view handover metrics. Review the [logs](/ai/agents/test/logs), verify the setup of your [handover provider](/ai/escalate/handovers), and check your handover provider's website for any issues with their services. If the issue can't be resolved, contact [Cognigy Product Support](/help/get-help) and provide them the [error log](/ai/agents/test/logs) for further investigation.                                                                                                                                                                        |

## More Information

* [All Metrics](/ops-center/monitor/metrics/overview)


# Errors
Source: https://docs.cognigy.com/ops-center/monitor/errors

Learn about the different error types in Ops Center and how to investigate and resolve them.

The error types are specific to each component. In Ops Center, you can view:

* All errors across services on the **Overview** tab.
* Errors for a specific component by opening its corresponding tab.

Errors from the past 7 days are displayed, including both run-time errors and those encountered in the Interaction Panel.

To investigate a specific error:

1. Open the **Errors** page. Click the error to see its details in the **Error Details** window.
2. Click **Go to `<resource name>`**, for example, **Go to Flow** or **Go to Endpoint**, to open the page where the error occurred. Find the error by its code and follow the recommended steps to fix it.

<Frame>
  <img alt="Ops Center Errors page with error details and navigation to resources" />
</Frame>

## List of Errors

| **Error Code**     | **Error Name**                                                                                                                                        | **Recommended Action**                                                                                                                                                                                 |
| ------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `ERR_ENDPOINT_001` | Endpoint Input Transformer Failed                                                                                                                     | Check the Endpoint's transformer settings, logs, and input mappings for errors.                                                                                                                        |
| `ERR_ENDPOINT_002` | Endpoint Output Transformer Failed                                                                                                                    | Check the Endpoint's transformer settings, logs, and output data format for errors.                                                                                                                    |
| `ERR_ENDPOINT_003` | Endpoint Execution Finished Transformer Failed                                                                                                        | Check the Endpoint's transformer settings. Validate the final execution transformer logic. Ensure it properly handles the final payload and any cleanup logic.                                         |
| `ERR_ENDPOINT_004` | Endpoint Inject Transformer Failed                                                                                                                    | Check the Endpoint's transformer settings. Check the injection logic that modifies request and response data. Ensure injected variables exist and are valid.                                           |
| `ERR_ENDPOINT_005` | Endpoint Notify Transformer Failed                                                                                                                    | Check the Endpoint's transformer settings. Confirm that the notification transformer logic is implemented correctly. Validate event payloads and downstream subscribers.                               |
| `ERR_FLOW_001`     | Flow Node Execution Failed                                                                                                                            | Go to the Node and review its parameters. Use logs and the Interaction Panel to debug the Node.                                                                                                        |
| `ERR_FLOW_002`     | Flow Outbound HTTP Call Failed                                                                                                                        | Check the outbound HTTP call settings: HTTP Request Node, endpoint URL, headers, auth, and payload. Test connectivity and response format.                                                             |
| `ERR_FLOW_003`     | Flow NLU Intent Misconfigured                                                                                                                         | Go to the Flow. Ensure the NLU Intent exists, is correctly configured, and trained. Retrain if needed.                                                                                                 |
| `ERR_FLOW_004`     | Flow NLU Embedding Model Misconfigured                                                                                                                | Go to the NLU embedding model in the Project settings and check that the model is correctly configured.                                                                                                |
| `ERR_FLOW_005`     | Flow Generative AI Model Misconfigured                                                                                                                | Go to the model in the LLM configuration. Check the model is correctly configured, test its connection, verify API keys.                                                                               |
| `ERR_FLOW_006`     | Flow Knowledge AI Embedding Model Misconfigured                                                                                                       | Go to the embedding model in the LLM configuration. Check the model is correctly configured, test its connection, verify API keys. Ensure the Knowledge AI embedding model is deployed and accessible. |
| `UNKNOWN`          | The error occurs when Ops Center can't determine the error type or identify which resource, such as LLM, Flow, NLU, or others, the issue occurred in. | Use the information from the **Message** section in the **Error Details** window, then go to the Project to troubleshoot the issue.                                                                    |

## More Information

* [Manage Errors](/ops-center/manage/errors)


# Endpoint Metrics
Source: https://docs.cognigy.com/ops-center/monitor/metrics/endpoints

Learn about the Endpoint metrics available in Ops Center to monitor the performance and reliability of your Endpoints.

The Endpoints dashboard shows metrics for your Endpoints through the following visualizations:

* [Line Charts](#line-charts)
* [Heatmaps](#heatmaps)

## Line Charts

### Endpoint Traffic

Shows the total number of requests received by Endpoints over time. This metric is useful for monitoring traffic volume and identifying usage patterns.

### Transformer Errors

Shows the total number of errors that occurred during the execution of Endpoint Transformers in Cognigy. Errors in Endpoint Transformers can disrupt the data flow between Cognigy and external services, potentially affecting the accuracy and reliability of integrations.

## Heatmaps

### Transformer Processing Time

Shows the time, in milliseconds, taken to process all active Endpoint transformers. This metric helps monitor the performance and responsiveness of active Endpoint transformers. High processing times may indicate bottlenecks, misconfigurations, or delays in communication with external services.

### Transformer Message Time

Shows the time, in milliseconds, taken to process a single message through Endpoint transformers. This metric helps identify problematic messages or anomalies in processing time.


# Flow Metrics
Source: https://docs.cognigy.com/ops-center/monitor/metrics/flows

Learn about the Flow metrics available in Ops Center to monitor the performance and reliability of your Flows.

The Flows dashboard shows metrics for your Flows through the following visualizations:

* [Line Charts](#line-charts)
* [Heatmaps](#heatmaps)

## Line Charts

### Flow Execution Time (90th percentile)

Shows the 90th percentile of Flow execution times, measured in milliseconds. This means that 90% of all executed Flows finish within this duration.

### Custom Flow Nodes (Extensions) Execution

Shows the number of Extensions that are executed per minute within your Flows.
This metric helps evaluate Extension usage and identify potential performance bottlenecks caused by custom code. The metric also provides insights into how often specific integrations or logic are triggered during Flow execution.

### Knowledge AI Queries

Shows the number of successful Knowledge AI queries executed against the vector database. This metric is useful for tracking query volume and system reliability.

### Outbound HTTP Requests

Shows the number of API calls made via the HTTP Request Node. This metric helps monitor the volume of external requests and track integration activity.

### LLM Fallbacks Triggered

Shows the number of times an LLM fallback was triggered because the main model failed.
This metric helps identify reliability issues with the main model.

### LLM Retries Triggered

Shows the number of API calls to LLM providers that were retried due to issues with a model.
This metric helps identify potential reliability or performance problems affecting model responses.

## Heatmaps

### Extension Processing Time

Shows the execution time, in milliseconds, for running Extension code. The metric helps assess the performance impact of the Extension code on Flow execution.
Monitoring this metric can identify potential delays or inefficiencies in the Extension code.

### NLU Scoring Time

Shows the duration, in milliseconds, that the NLU model takes to score a user input.
This metric helps evaluate the performance and responsiveness of the NLU model under load.

### Knowledge AI Query Latency

Shows the time, in milliseconds, to query Knowledge AI and receive a response.
This metric is useful for monitoring Knowledge AI performance.
High query latency can slow down the AI Agent. Slow responses make the AI Agent unresponsive, leading users to lose interest or abandon the conversation.

### Outbound HTTP Request Latency

Shows the time, in milliseconds, taken to complete outbound HTTP requests performed via the HTTP Request Node.
This metric helps identify potential latency issues in outbound HTTP requests.


# Handover Provider Metrics
Source: https://docs.cognigy.com/ops-center/monitor/metrics/handover-providers

Learn about the Handover Provider metrics available in Ops Center to monitor the performance and reliability of your handover providers.

<a href="/release-notes/2025.24"><Badge>Updated in 2025.24</Badge></a>

The Handover Providers dashboard shows metrics for handover providers created from the [handover provider list](/ai/escalate/handover-reference/overview) and handovers initiated via the [Handover to Human Agent Node](/ai/agents/develop/node-reference/service/handover-to-agent). These metrics are visualized through the following charts:

* [Line Charts](#line-charts)

## Line Charts

### Handover Requests

Shows the number of handover requests created at a specific time when the Cognigy Agent transfers conversations to contact centers, such as Live Agent, 8x8, Genesys, and others.
This metric reflects how many handover attempts were initiated and helps track time-based integration health to ensure smooth conversation transfers.

### Handover Messages

Shows the number of messages exchanged between an AI Agent and a human agent at a specific time during a handover process.
This metric reflects the actual message flow after a handover request has been accepted and is essential for monitoring the volume and progress of communication within the handover process.


# Overview
Source: https://docs.cognigy.com/ops-center/monitor/metrics/overview

View system-wide performance, error rates, and health status for all Cognigy components in the Ops Center Overview dashboard with key indicators.

The Overview dashboard provides a comprehensive view of system-wide performance, error rates, and health status across key components in the Cognigy platform.

<Frame>
  <img alt="Ops Center Overview dashboard showing system-wide performance and health status" />
</Frame>

View the following visualizations:

* [Indicators](#indicators)
* [Line Chart](#line-chart)
* [Panel](#panel)

## Indicators

### Active Alerts

The total number of alerts received across your organization.

### Errors

The total number of errors received across your organization.

## Line Chart

### Messages

The total number of user inputs processed per minute. This metric helps track the overall system load.

## Panel

### Voice Gateway

<Note>
  Voice Gateway metrics are under development and will be available in future releases.
</Note>

| **Subcomponent / Alert Status** | **🟢 Inactive** | **🔴 Active** |
| ------------------------------- | --------------- | ------------- |
| SIP Trunk Connectivity          | -               | -             |
| STT (Speech-to-Text)            | -               | -             |
| TTS (Text-to-Speech)            | -               | -             |

### Endpoints

| **Subcomponent / Alert Status** | **🟢 Inactive**                                                                                                                  | **🔴 Active**                                                                                                           |
| ------------------------------- | -------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------- |
| Transformer                     | The number of errors thrown by the Endpoint Transformers is less than 10% over 5 minutes, or the absolute count is less than 20. | The number of errors thrown by the Endpoint Transformers increases by more than 10% in 5 minutes and exceeds 20 errors. |
| Real-Time Translation           | The number of real-time translation errors is less than 10% over 5 minutes, or the absolute count is less than 20.               | The number of real-time translation errors increases by more than 10% in 5 minutes and exceeds 20 errors.               |

### Flows

| **Subcomponent / Alert Status** | **🟢 Inactive**                                                                                                                                        | **🔴 Active**                                                                                                                                 |
| ------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------ | --------------------------------------------------------------------------------------------------------------------------------------------- |
| Outbound HTTP Requests          | The number of errors from Outbound HTTP Request Node execution is less than 10% over 5 minutes, or the absolute count is less than 20.                 | The number of errors from Outbound HTTP Request Node execution increases by more than 10% in 5 minutes and exceeds 20 errors.                 |
| NLU                             | The number of NLU processing errors is less than 10% over 5 minutes, or the absolute count is less than 20.                                            | The number of NLU processing errors increases by more than 10% in 5 minutes and exceeds 20 errors.                                            |
| LLMs                            | The number of errors thrown during Large Language Model (LLM) prompt execution is less than 10% over 5 minutes, or the absolute count is less than 20. | The number of errors thrown during Large Language Model (LLM) prompt execution increases by more than 10% in 5 minutes and exceeds 20 errors. |
| Knowledge AI                    | The number of errors thrown by the knowledge search operations is less than 10% over 5 minutes, or the absolute count is less than 20.                 | The number of errors thrown by knowledge search operations increases by more than 10% in 5 minutes and exceeds 20 errors.                     |

### Handover Providers

| **Subcomponent / Alert Status** | **🟢 Inactive**                                                                                                      | **🔴 Active**                                                                                               |
| ------------------------------- | -------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------- |
| Live Agent                      | The number of requests to Cognigy Live Agent is less than 10% over 5 minutes, or the absolute count is less than 20. | The number of requests to Cognigy Live Agent increases by more than 10% in 5 minutes and exceeds 20 errors. |

## More Information

* [Alerts](/ops-center/monitor/alerts)
* [Errors](/ops-center/monitor/errors)


# Voice Gateway Metrics
Source: https://docs.cognigy.com/ops-center/monitor/metrics/voice-gateway

Learn about the Voice Gateway metrics available in Ops Center to monitor the performance and reliability of your Voice Gateway.

<Note>
  Voice Gateway metrics are under development and will be available in future releases.
</Note>

The Voice Gateway dashboard shows metrics for your Voice Gateway through the following visualization:

* [Line Chart](#line-chart)

## Line Chart

### Voice Gateway Incoming Messages

Shows the total number of user inputs that are transcribed and processed by Voice Gateway.


# What's AI Ops Center?
Source: https://docs.cognigy.com/ops-center/overview

Manage AI Agent deployments at scale with operational visibility, performance monitoring, and proactive alerts.

*AI Ops Center* is a Cognigy product developed to manage the deployment of AI Agents at scale. It serves as the solution for gaining operational visibility and control over AI Agents.

<div>
  <video title="AI Ops Center overview demo showing dashboards, alerts, and other monitoring features" aria-label="AI Ops Center overview demo showing dashboards, alerts, and other monitoring features">
    <source />
  </video>
</div>

## Who are AI Ops Center's Users?

Ops Center caters to a diverse range of roles across multiple industries. It is particularly beneficial if you are a system administrator managing platform performance, a conversation designer developing and maintaining AI Agents, or part of an operations team monitoring system health. Ops Center is also valuable for customer support teams tracking issues and optimizing customer experiences.

## Benefits of AI Ops Center

### Ensure Business Continuity

With Ops Center, you can minimize disruptions caused by platform or dependency issues.

### Provide Real-time Performance Insights

Ops Center offers clear visibility into how systems are performing moment-to-moment.

### Enable Proactive Issue Resolution

With Ops Center, you can notify users about potential issues before they escalate, enabling quicker fixes and mitigation.


# Common Issues
Source: https://docs.cognigy.com/ops-center/troubleshooting

Find solutions to common issues encountered while using Ops Center.

This page lists common issues you may encounter while using Ops Center and provides troubleshooting steps to resolve them.

<Accordion title="The Live Indicator Changed Color from Green to Gray">
  Refresh the page, clear the cache, and reopen the browser. If the issue persists, contact [Cognigy technical support](/help/get-help) for further assistance.
</Accordion>

<Accordion title="Service Loaded, but No Updates Visible in Ops Center">
  Data is updated every 30 seconds, so there might be a delay.
</Accordion>

<Accordion title="The Resource Column on the Errors Tab Shows a Dash in some Rows">
  The dash indicates that the resource, such as a Node, Flow, or LLM, associated with the error has been deleted.
</Accordion>


# 2025.10
Source: https://docs.cognigy.com/release-notes/2025.10



<Note>
  We are excited to announce a change in our software versioning system!
  This new system will make it easier for you to track releases by showing both their order and year, providing clearer insight into our updates and patches.
  Moving forward, we will use a new format: `YEAR.MINOR_RELEASE.PATCH_LEVEL`, where:

  * `YEAR` is the year of the release, for example, `2025`.
  * `MINOR_RELEASE` is the Nth minor release of the year, for example, the 10th minor release in 2025 is `10`.
  * `PATCH` is the Nth patch for a minor release, for example, the 1st patch for the 10th minor release is `1`. If no patches are released, the number is `0`.
</Note>

<Warning>
  As of April 2025, the AudioCodes Endpoint is deprecated. The full removal is planned for Q2 2026. Migrate to the [Voice Gateway Endpoint](/ai/agents/deploy/endpoint-reference/voice-gateway) before the removal date.

  Refer to the list of all upcoming removals on the [Deprecations and Removals](/release-notes/deprecations-and-removals) page.
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2025.10.2

### Cognigy

**Released** May 21, 2025

#### Bug Fixes

* Fixed the issue where the `baseUrl` validation failed for Salesforce MIAW, preventing the handover process from completing successfully

## 2025.10.1

### Cognigy

**Released** May 19, 2025

#### Improvements

* Increased the default timeout to prevent queue transfers from triggering the end of [Genesys Cloud Open Messaging handovers](/ai/escalate/handover-reference/genesys-cloud-open-messaging) from 2.5 to 10 seconds. Ending a chat before this timeout prevents the handover from ending. You can change the timeout with the newly introduced `GENESYS_OM_HANDOVER_ACW_AFTER_TRANSFER_TIMEOUT_IN_MS` environment variable in `values.yaml`. This timeout is useful when multiple after-call work events are sent through the event bridge and cause handovers to end during queue transfers

#### Bug Fixes

* Fixed the issue where [Salesforce handovers](/ai/escalate/handover-reference/salesforce) failed because the Salesforce API responses didn't return the affinity tokens required to create a new conversation

### Cognigy Voice Gateway

**Released** May 26, 2025

#### Bug Fixes

* Fixed the issue where users couldn't add phone numbers on the Phone numbers page in the Voice Gateway Self-Service Portal

### Cognigy Live Agent

**Released** June 2, 2025

#### Bug Fixes

* Fixed the issue where editing custom fields in the Contact Profile details caused an error
* Fixed the issue where editing the default fields in the Contact Profile details caused an error

## 2025.10.0

### Cognigy

**Released** May 13, 2025

#### Improvements

* Added the [Identity Token for Authenticated Users](/ai/escalate/handover-reference/salesforce-miaw) setting to include the customer identity token for Salesforce MIAW in the Handover to Agent Node. This setting enables user verification via the customer's identity provider (IdP) in Salesforce
* Improved the functionality of the Salesforce MIAW handover provider by preventing infinite loops when a handover expires without the human agent manually ending it. This improvement ensures that the system will no longer get stuck in a loop if the handover session ends due to timeout or other issues, providing a more reliable user experience
* Improved the behavior of the [Send Transcript as first message](/ai/escalate/handover-reference/salesforce-miaw) setting in the Salesforce MIAW handover provider by splitting the transcript into multiple messages when it exceeds the accepted message character limit
* Added email notifications to alert users whenever they change their password, ensuring they are promptly informed of security-related changes
* Removed the [100 MB Snapshot download limit](/ai/for-developers/developers/api-and-cli) in the Cognigy CLI

#### Bug Fixes

* Fixed the issue where a Go To Node showed unsaved changes after its target Node was deleted
* Fixed the issue where logs sorted by `_id` appeared out of chronological order
* Fixed the issue where Datepicker label translations in the Question Node didn't work correctly
* Fixed the issue where a handover wasn't canceled when triggered by an Intent
* Fixed the issue where no indication was displayed when the Flow failed due to a deleted Node

### Cognigy Voice Gateway

**Released** May 13, 2025

#### Improvements

* Added the capability to send a `gather` verb with nested text for a Question Node, reducing latency in recognition. This improvement enables faster processing, particularly when users provide quick responses. For example, when a `Yes/No` question is asked, Voice Gateway can now process a quick `yes` or `no` response more efficiently. Without this improvement, the delay in recognition could cause issues with fast replies. To enable this capability, set `FEATURE_ENDPOINT_VG_GATHER_NESTED: "true"` in the Cognigy `values.yaml` file

#### Bug Fixes

* Fixed the issue in `api-server` where emojis weren't supported in the application name in the Voice Gateway Self-Service Portal
* Fixed the issue in `feature-server` where the `bargein` parameter was disabled when a new `gather` verb was sent. Additionally, fixed the `bargein` behavior for the `dial` or `sip:refer` verbs to prevent call hang-ups and other unexpected issues
* Fixed the issue in the Voice Gateway Self-Service Portal where the wrong carrier was shown when the user opened the Edit Phone Number page
* Fixed the issue in `feature-server` where using non-streaming TTS with Microsoft resulted in an error. This issue occurred during the TTS process when Microsoft was selected as the provider, causing disruptions in the expected behavior

### Cognigy Webchat

**Released** May 13, 2025

#### Improvements

* Updated the Webchat v3 widget's [Close and Minimize](/webchat/v3/features#controlling-the-chat-window) buttons to trigger the `webchat/minimize` and `webchat/close` events
* Updated Webchat v3 to [v3.20.0](https://github.com/Cognigy/Webchat/releases/tag/v3.20.0)

### Cognigy Live Agent

**Released** May 13, 2025

#### Improvements

* Renamed Channel Name to Inbox Name at the [Create Inbox step](/live-agent/settings/inboxes) in Settings > Inboxes
* Removed deprecated Webhooks and Integrations from the Live Agent API documentation

#### Bug Fixes

* Fixed the issue where saved filters in the Cognigy Channel Type list, for example, WhatsApp, incorrectly showed conversations from other channels, for example, Webchat

### Cognigy Insights

**Released** May 13, 2025

#### Improvements

* Added the infinite loader to the Contains Steps, Starting Step, Endpoints, Snapshots, Tasks, and Goals filters
* Added the search functionality to filter steps by label in the Insights API

#### Bug Fixes

* Fixed the issue where the Contains Step filter in the Transcript Explorer didn't work properly

### Infrastructure

#### Cognigy Voice Gateway

* Introduced a new feature flag in Cognigy: `FEATURE_ENDPOINT_VG_GATHER_NESTED: "true"`. This flag determines whether a question is sent as a `gather` verb with a nested `say` verb, or as separate `say` and `gather` verbs. The default value is `false`

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and AI Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2025.11
Source: https://docs.cognigy.com/release-notes/2025.11



<Warning>
  * As of April 2025, the [Insights Beta API](https://api-trial.cognigy.ai/openapi#get-/insights/beta/messages/report/-queryHash-) is deprecated. The full removal is planned for Cognigy 2025.12.0. After the removal, use the [Cognigy OData endpoint](/ai/agents/analyze/odata) instead.
  * As of April 2025, the [Intent Sentence Generation, Lexicon Generation, Flow Generation, Adaptive Card Generation](/ai/agents/develop/gen-ai-and-llms/generative-ai#generate-cognigy-resources) lists in the Project settings are deprecated. The full removal is planned for Cognigy 2025.13.0. After the removal, select the LLM for all these features from the [Design-Time Generative LLM Features list](/ai/agents/develop/gen-ai-and-llms/generative-ai#design-time-generative-ai-features).

  Refer to the list of all upcoming removals on the [Deprecations and Removals](/release-notes/deprecations-and-removals) page.
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2025.11.2

### Cognigy

**Released** June 17, 2025

#### Bug Fixes

* Fixed the issue where applying changes to a Question Node failed without displaying errors

## 2025.11.1

### Cognigy Insights

**Released** June 5, 2025

#### Bug Fixes

* Fixed the issue where Dead Letter Exchange (DLX) queues had no consumers, causing failed or undeliverable messages to remain unprocessed. This issue led to delayed error handling and affected the performance of AI Agents

## 2025.11.0

### Cognigy

**Released** May 27, 2025

#### Improvements

* Removed the feature flag for the `amazon.titan-embed-text-v2:0` [embedding model](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature). Now, this model is available to all customers
* Added support for attachments with the [Salesforce MIAW handover provider](/ai/escalate//handover-reference/salesforce-miaw) from end users to human agents through Webchat v2 and Webchat v3
* Added the Log System Prompt and Log Tool Definitions options to the Debug Settings in the [AI Agent Node](/ai/agents/develop/node-reference/ai/ai-agent). When you activate these options, debug mode displays the AI Agent system prompt and the configured AI Agent tools information in the Interaction Panel
* Added support for streamable HTTP MCP servers to the MCP Tool Node, with support for existing SSE MCP servers as fallback servers

#### Bug Fixes

* Fixed the issue where different pages displayed a scrollable empty space at the bottom for MacBook Pro 14 users
* Fixed the issue where the Override Snapshot Connections option in the Endpoint settings was ignored
* Fixed the issue where undeprecated Nodes were considered deprecated and didn't show up in the Flow editor
* Fixed the issue where strict validation caused errors when:
  * using Extensions with time settings configured as optional
  * saving Question Nodes with the Question type set to Date and the Minimum Date or Maximum Date parameter set to `today`

### Cognigy Voice Gateway

**Released** May 27, 2025

#### Bug Fixes

* Fixed the issue where the user couldn't add phone numbers on the Phone numbers page in the Voice Gateway Self-Service Portal
* Fixed the issue where streaming TTS using ElevenLabs failed
* Fixed the issue where non-streaming TTS using Microsoft Azure Speech Services failed

### Cognigy AI Copilot

**Released** May 27, 2025

#### Improvements

* Added automatic display of the [AI Copilot workspace](/agent-copilot/overview). Now, human agents can view the AI Copilot workspace without waiting for user input

### Cognigy Live Agent

**Released** May 27, 2025

#### Improvements

* Added the `SKIP_GENESYS_EVENTBRIDGE_ORGANIZATION_ID_CHECK` feature flag to prevent the validation of the Cognigy organization ID during a handover. This approach is helpful when users have multiple organizations for one Genesys Cloud Open Messaging account

#### Bug Fixes

* Fixed the issue where the `$limit` and `$top` OData Query Language operators didn't work correctly for the user data model with the Live Agent OData endpoint. Additionally, removed unsupported operators from this endpoint. To gain overview of the supported operators, refer to the [OData Analytics endpoint documentation](/live-agent/tools/odata-endpoint)
* Fixed the issue where abandoned conversations were reassigned when a human agent's availability changed to online
* Fixed the issue where updating the Contact Profile details failed

### Cognigy Insights

**Released** May 27, 2025

#### Bug Fixes

* Fixed the issue where reports displayed data as strings instead of numbers
* Fixed the issue where deleting a Contact Profile deleted Contact Profile data in the Live Agent Escalations table
* Fixed the issue where the disabled Cognigy OData Endpoint didn't return 404 errors

### Infrastructure

#### Cognigy + Insights + AI Copilot

Cognigy 2025.11.0 and later versions are compatible with MongoDB 7.0:

* For SaaS installations, all environments hosted by Cognigy will be updated to MongoDB 7.0.
* For on-premises installations, update MongoDB to 7.0 after updating Cognigy to version 2025.11.0.

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and AI Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2025.12
Source: https://docs.cognigy.com/release-notes/2025.12



<Warning>
  The [Insights Beta API](https://api-trial.cognigy.ai/openapi#get-/insights/beta/messages/report/-queryHash-) has been removed. Use the [Cognigy OData endpoint](/ai/agents/analyze/odata) instead.

  Refer to the list of all upcoming removals on the [Deprecations and Removals](/release-notes/deprecations-and-removals) page.
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2025.12.2

### Cognigy

**Released** June 17, 2025

#### Bug Fixes

* Fixed the issue where applying changes to a Question Node failed without displaying errors

## 2025.12.1

### Cognigy

**Released** June 16, 2025

#### Bug Fixes

* Fixed the issue where custom attributes weren't included in the first message sent to Genesys Cloud Open Messaging. This message contained the conversation transcript between the end user and the AI Agent. As a result, the Genesys flow didn't capture the custom attributes as its first step
* Fixed the issue where the handover restarted during the connection to MongoDB Atlas
* Fixed the issue where Snapshots with a handover provider rendered the Handover to Agent Node's provider field empty, causing the handover to fail
* Fixed the issue where an update in the message processing queue caused missing user inputs

## 2025.12.0

### Cognigy

**Released** June 10, 2025

#### Features

##### Collaborative Editing

Introduced [Collaborative Editing](/ai/agents/develop/projects-and-flows/collab-editing), a new feature designed to boost team productivity and collaboration while designing Flows. From now on, you will see other active users in the Flow and within individual Nodes, including their mouse cursors and real-time update indicators showing the number of changes made. To reduce conflicts from parallel updates, we've also introduced Node locking, preventing others from editing a Node while someone else is working on it. Finally, if you have questions for other project members, you can now start a discussion directly on a Node and mention teammates. This feature enables your team to work faster and more efficiently when developing AI Agents together.

The feature is activated in a SaaS installation.

##### Native Integration with Zoom

Launched a [Zoom Contact Center Endpoint](/ai/agents/deploy/endpoint-reference/zoom) for chat with AI Agents, improving collaboration and automating workflows.

#### Improvements

* Set a maximum limit of 1 year for Analytics Data Expiration, Insights Session Data Expiration, and Steps Data Expiration in the Management UI
* Added support for a [custom authentication header name](/ai/agents/develop/gen-ai-and-llms/providers/openai-compatible) field to the OpenAI-compatible LLM provider settings
* Improved the stability of the `service-handover` database connection to MongoDB and added caching for handover data while sending messages to handover providers. This change enhances the speed and reliability of the handover process
* Updated the [Amazon EventBridge integration](/ai/escalate/handover-reference/genesys-cloud-open-messaging-with-amazon-eventbridge) with the Genesys Open Messaging handover provider to prevent conversations from ending during queue transfers. If you integrate with Genesys Open Messaging via EventBridge, update your configuration to use the new `v2.detail.events.conversation.{id}.customer.end` topic. Previously, you could end conversations using WebSocket events or Trigger-based methods. These methods have been removed and are no longer supported

#### Bug Fixes

* Fixed the issue where the `FEATURE_DISABLE_NPM_PACKAGES_IN_SECURE_CONTEXT` flag wasn't passed to the `service-function-execution` child process
* Fixed the issue preventing updates to handover providers imported via Snapshots
* Fixed the issue where the `disabled` status of Knowledge Sources imported from a Package was overwritten
* Fixed the issue where the HTTP Request Node encoded query parameters, resulting in altered query values
* Fixed the issue where Projects using Voice Gateway Endpoints with the Live Agent handover provider stopped delivering human agent messages to end users
* Fixed the issue where goals imported via a Package or Snapshot weren't referenced in the Track Goal Node

### Cognigy Voice Gateway

**Released** June 10, 2025

#### Improvements

* Added support for the Barge In feature to AI Agent Nodes. Now, when a Set Session Config Node precedes an AI Agent Node, the Barge In settings from the Set Session Config Node are applied to the AI Agent Node
* Improved `feature-server` by adding the entire `say` verb content to the summary on the Tracing tab in Recent Calls within the Voice Gateway Self-Service Portal

#### Bug Fixes

* Fixed the issue in `feature-server` where it stopped working when attempting to play a silent overlay
* Fixed the issue where the nested `gather` verb wasn't applied when the Repeat Question toggle was disabled under Reprompt Options in the Question or Optional Question Node
* Fixed the issue in `feature-server` where DTMF inputs were gathered after the maximum number of digits had been reached
* Fixed the issue of duplicate inputs due to concurrent gathers
* Fixed the issue where the `CALL_COMPLETED` event wasn't sent to the Endpoint

### Cognigy Webchat

**Released** June 10, 2025

#### Improvements

* Added support for the [`speak`](/webchat/v3/accessibility#adaptive-cards) parameter for adaptive cards, which allows you to include spoken text summary of the adaptive card for better accessibility and user engagement

#### Bug Fixes

* Fixed accessibility issues in Webchat v3

### Cognigy Live Agent

**Released** June 10, 2025

#### Bug Fixes

* Fixed the issue where skill-based routing and combined routing, which match human agents to conversations by language and other skills, didn't work as expected
* Fixed the issue where an `invalid team name` error was shown incorrectly during team creation or editing Teams in Settings > Teams
* Fixed the issue where the Contact Profile in conversations didn't display the Memories field after the Edit contact window was opened

### Cognigy Insights

**Released** June 10, 2025

#### Features

##### Business Dashboard

Introduced the [Business Dashboard](/insights/dashboards/business). It gives you a quick overview of your key session metrics through interactive charts and tables, and includes location-based activity, active sessions by AI or human agents, and business outcomes.

#### Improvements

* Removed the Insights Beta API

#### Bug Fixes

* Fixed the issue in the Step Explorer and Transcript Explorer where applying a Snapshot filter caused Step filters to show duplicate results. Now, Explorers accurately display only the selected steps, even when a Snapshot filter is active

### Infrastructure

#### Cognigy + Insights + AI Copilot, Voice Gateway

Traefik, the Kubernetes ingress controller, has been updated to enhance functionality and address known security vulnerabilities in the dependencies of Cognigy and Cognigy Insights. This update upgrades Traefik to version 3.4.0 and includes the latest Helm chart, reflecting a commitment to continuous improvement.

##### Traefik Update

Make updates for Cognigy and Voice Gateway separately, because they have different Helm charts and `values.yaml` files.

Before installing the Helm chart, follow these steps:

1. In the new Traefik Helm chart, the `traefik.ports.<port_name>.redirectTo` has been replaced with `traefik.ports.<port_name>.redirections`. If you've modified the Traefik port redirection in your local `values.yaml` file, update it accordingly.

   Before:

   ```yaml theme={null}
   ports:
     web:
       redirectTo:
         port: websecure
   ```

   Now:

   ```yaml theme={null}
   ports:
     web:
       redirections:
         entryPoint:
           to: websecure
           scheme: https
   ```

2. Deploy the latest CRDs of Traefik by executing the following command:

   ```bash theme={null}
   kubectl apply --server-side --force-conflicts -k https://github.com/traefik/traefik-helm-chart/traefik/crds/?ref=v35.4.0
   ```

For more detailed information, refer to the official [release notes](https://github.com/traefik/traefik-helm-chart/releases) provided by Traefik.

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and AI Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2025.13
Source: https://docs.cognigy.com/release-notes/2025.13



<Warning>
  From version 2025.13.0 onward, creating empty Knowledge Chunks via the [GUI or API is no longer allowed](/ai/agents/develop/knowledge-ai/knowledge-chunk/knowledge-chunk). Existing empty Knowledge Chunks remain temporarily visible in the Chunk editor and will soon be removed automatically.
</Warning>

<Warning>
  The Intent Sentence Generation, Flow Generation, Adaptive Card Generation, Lexicon Generation settings have been removed from [Project settings](/ai/agents/develop/gen-ai-and-llms/generative-ai#configure-llms-for-design-time-features). You can configure an LLM for these features altogether with the Design-Time LLM Features setting. If you had configured an LLM for one of the removed features and configure the Design-Time LLM Features setting as None, this setting automatically uses the previously configured LLM.

  Refer to the list of all upcoming removals on the [Deprecations and Removals](/release-notes/deprecations-and-removals) page.
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2025.13.1

### Cognigy Live Agent

**Released** June 25, 2025

#### Bug Fixes

* Fixed the issue where upgrading the Live Agent Helm chart from version 2025.13.0 to 2025.13.1 resulted in a migration error

## 2025.13.0

### Cognigy

**Released** June 24, 2025

#### Features

##### New LLM Prompt Node and Load AI Agent Node

Introduced two new Nodes to give you more flexibility when building AI Agents:

* [New LLM Prompt Node](/ai/agents/develop/node-reference/service/llm-prompt) — provides more flexibility and includes advanced AI functionalities, including tools, image support, and use of the conversation transcript that the AI Agent Node uses. The legacy LLM Prompt Node will still be available.
* [Load AI Agent Node](/ai/agents/develop/node-reference/ai/load-ai-agent) — lets users load the AI Agent configuration and persona to the Input or Context object, for example, for reusing the conversation transcript or the persona in the new LLM Prompt Node.

#### Improvements

* Added support for [OpenAI's](/ai/agents/develop/gen-ai-and-llms/providers/openai) and [Microsoft Azure OpenAI's](/ai/agents/develop/gen-ai-and-llms/providers/microsoft-azure-openai) `gpt-4.1`, `gpt-4.1-mini`, and `gpt-4.1-nano` models, [Anthropic's](/ai/agents/develop/gen-ai-and-llms/providers/anthropic) `claude-opus-4-0` and `claude-sonnet-4-0`, and [Mistral AI's](/ai/agents/develop/gen-ai-and-llms/providers/mistral) `mistral-medium` model. Additionally, changed the version mapping of [Anthropic's](/ai/agents/develop/gen-ai-and-llms/providers/anthropic) and Mistral AI's models to include the `*-latest` suffix. This suffix ensures that the latest version of the model is used when you select it. The versioning mapping applies to the following models:
  * Anthropic's `claude-3-5-sonnet` and `claude-3-7-sonnet` are now listed as `claude-3-5-sonnet-latest` and `claude-3-7-sonnet-latest` respectively.
  * Mistral AI's `mistral-large-2411`, `mistral-small-2501`, and `pixtral-large-2411` are now listed as `mistral-large-latest`, `mistral-small-latest`, and `pixtral-large-latest` respectively. `mistral-medium` is listed as `mistral-medium-latest`
* Added options to upload a Knowledge Source from a URL or file in the [AI Agent creation wizard](/ai/agents/develop/manage-ai-agents)
* Added capability to collapse debug messages in the [Interaction Panel](/ai/agents/test/interaction-panel/overview) and copy them to the clipboard
* Added the [Render Markdown option](/ai/agents/test/interaction-panel/overview#configure-the-interaction-panel) support to the Interaction Panel settings. Now, you can test how AI Agents output Markdown formatting
* Improved stability of the handover database connection and optimized message sending to handover providers by using caching

#### Bug Fixes

* Fixed the issue where the Transcript Steps setting in the Search Extract Output Node was fixed to 2 in the backend, even though users could change the setting in the Node editor. Now, changes to this setting apply accordingly
* Fixed the issue where Knowledge Chunks with empty text were included in a Knowledge Source
* Fixed the issue where deleting an entry from the Intent Trainer list caused an issue with the Add to Intent button
* Fixed the issue where a Voice Gateway Endpoint using Live Agent for handovers didn't receive messages from human agents
* Fixed the issue where empty Knowledge Chunks could be created in the Chunk Editor
* Fixed the issue where 2024 and 2025 were missing in the Conversation report year selector in Management UI
* Fixed the issue where the Question Node editor displayed the Google Model field twice in the STT section of the Voice Gateway channel
* Fixed the issue where nested `gather` verbs weren't applied when Repeat Question toggle was deactivated in the Reprompt Options section
* Fixed the issue where changing the Priority field in the Live Agent handover provider settings in the Handover to Agent Node didn't activate the Save Node button

### Cognigy Voice Gateway

**Released** June 24, 2025

#### Improvements

* Improved general stability of Voice Gateway

### Cognigy Webchat

**Released** June 24, 2025

#### Improvements

* Updated Webchat v3 to [v3.24.0](https://github.com/Cognigy/Webchat/releases/tag/v3.24.0)
* Added support for Markdown to the [Privacy Notice text field](/webchat/v3/configuration#privacy-notice). Now, you can customize your Privacy Notice with bold and italic formatting, and hyperlinks

### Cognigy Live Agent

**Released** June 24, 2025

#### Improvements

* Added new parameters to the following collections in the [Live Agent OData endpoint](/live-agent/tools/odata-endpoint): Label, User, and Conversation

#### Bug Fixes

* Fixed the issue of the `selectOne` API endpoint. Previously, requesting a specific label, for example, `/Label(100)`, that didn't exist returned a `200 OK` status with an empty array. Now, the request correctly returns a `404 Not Found` status for missing labels, aligning with standard REST API practices

### Infrastructure

#### Cognigy + Insights + AI Copilot

Cognigy 2025.13.0 and later versions are compatible with MongoDB 7.0.15:

* For SaaS installations, all environments hosted by Cognigy will be updated to MongoDB 7.0.15.
* For on-premises installations, update MongoDB to 7.0.15 after updating Cognigy to version 2025.13.0 following the [instructions in Cognigy's GitHub repository](https://github.com/Cognigy/cognigy-mongodb-helm-chart?tab=readme-ov-file#upgrading-helm-release)

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and AI Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2025.14
Source: https://docs.cognigy.com/release-notes/2025.14



<Warning>
  If you have the Knowledge AI feature enabled and Qdrant is deployed, upgrading Cognigy to version 2025.14.0 will automatically upgrade Qdrant to v1.13.4. If you need to roll back Cognigy 2025.14.0 to an earlier version, refer to the [Infrastructure section](#infrastructure).
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2025.14.2

### Cognigy

**Released** July 18, 2025

* The following headers added in Cognigy 2025.14.0 were removed because they broke some xApps by restricting iframe embedding and cross-origin resource loading:

  * Content-Security-Policy with frame-src
  * Cross-Origin-Resource-Policy: cross-origin
  * Cross-Origin-Embedder-Policy: require-corp

## 2025.14.1

### Cognigy

**Released** July 11, 2025

#### Bug Fixes

* Fixed the issue where Flows with Extensions didn't throw errors during execution

## 2025.14.0

### Cognigy

**Released** July 8, 2025

#### Improvements

* Set the [platform-provided LLM as the default](/ai/agents/develop/gen-ai-and-llms/providers/cognigy-platform-provided) for new Projects. Now, when creating a new Project, you can immediately use design-time features without needing to select any LLM manually
* Added the Log LLM Latency option to [AI Agent](/ai/agents/develop/node-reference/ai/ai-agent) and [LLM Prompt Nodes](/ai/agents/develop/node-reference/service/llm-prompt). This option allows you to log the latency of LLM calls, which can help you monitor and optimize performance
* Added the [Custom URL](/ai/agents/deploy/endpoints/file-storage) option to the Amazon S3 Bucket connection for the File Storage feature. This option allows you to specify URLs for S3-compatible storage services, such as Open Telekom Cloud, DigitalOcean Spaces, or Backblaze B2
* Added support for passing the user's language to Salesforce MIAW handover provider. The language is taken from the Flow's locale and mapped to the Messaging User Language field in the session details on the Salesforce MIAW side
* Added support for rendering automated responses when using the Salesforce MIAW handover provider. This ensures that any event-based messages configured in Salesforce (such as session start or inactivity notifications) are delivered during the conversation
* Added the `DISABLE_ABORT_HANDOVER_ON_SALESFORCE_MIAW_PARTICIPANT_CHANGE` environment variable to control whether a Salesforce MIAW conversation automatically closes when a human agent is removed. By default, this environment variable is set to `false`. Setting it to `true` keeps the conversation open and continues polling for new events even after the human agent leaves
* Added gravatar to the [Main page](/ai/overview/user-interface/main-page). Now, you can see your gravatar in the upper-right corner of the Cognigy UI, making it easier to identify your account
  Added support for Salesforce's automated messaging components when using the [Salesforce MIAW handover provider](/ai/escalate/handover-reference/salesforce-miaw#automated-messaging-components)

#### Bug Fixes

* Renamed the Log Request and Completion option to [Log System Prompt & Completion](/ai/agents/develop/node-reference/service/llm-prompt) in the Debug Settings section of the LLM Prompt Node to better reflect the option's functionality
* Fixed the issue where the cache wasn't invalidated when deleting connection fields
* Fixed the issue where inconsistent timeout and retry handling in Generative AI features resulted in longer-than-expected execution times for Nodes using LLMs
* Fixed the issue where an inactive conversation using Salesforce MIAW caused handovers to restart
* Fixed the issue where the version number and UI elements in the User Menu overlapped
* Fixed the issue where clicking the Create New Token button didn't open the creation form if the Edit Token window was already open
* Fixed the issue that prevented changes to the Question Node from being applied successfully
* Fixed the issue where importing a Knowledge Store via a Package caused incorrect formatting of the `sourceMetaData.sourceName` property within the `topK` object returned by the knowledge search
* Fixed the issue where triggering two consecutive handovers within the same session caused an error
* Fixed the issue where the Playbook's input text was incorrect if the Flow contained the Execute Flow or Go To Nodes

### Cognigy Voice Gateway

**Released** July 8, 2025

#### Improvements

* Improved general stability of Voice Gateway

### Cognigy Webchat

**Released** July 8, 2025

#### Features

* Added [Style Presets](/webchat/v3/configuration#style-presets) to the Webchat v3 Endpoint. Now, you can apply a ready-made style with one click instead of configuring each setting manually

#### Improvements

* Improved Webchat v3 accessibility. Now, Webchat v3 supports [relative font sizes](/webchat/v3/accessibility#text-resizing), so you can adjust the text size using your browser settings
* Updated Webchat v3 and Demo Webchat to [v3.25.2](https://github.com/Cognigy/Webchat/releases/tag/v3.25.2)

### xApps

**Released** July 8, 2025

#### Bug Fixes

* Fixed the issue where xApps weren't loaded in Demo Webchat

### Cognigy Live Agent

**Released** July 8, 2025

#### Improvements

* Renamed Team name to Team Name on the [Create a new Team page](/live-agent/settings/teams) to improve consistency in the Live Agent UI

### Infrastructure

#### Cognigy + Insights + AI Copilot

##### Qdrant upgrade from v1.8.3 to v1.13.4

For environments where the Knowledge AI feature is enabled and Qdrant is deployed, upgrading Cognigy to version 2025.14.0 will automatically upgrade Qdrant to v1.13.4.

If you need to roll back Cognigy 2025.14.0 to an earlier version, don't roll back Qdrant to v1.8.3. The Qdrant upgrade isn't reversible.

To prevent issues, before rolling back Cognigy 2025.14.0 to an earlier version, specify the v1.13.4 tag in the Qdrant configuration as follows:

```yaml theme={null}
qdrant:
  image:
    tag: "v1.13.4"
...
```

This configuration ensures that Qdrant will continue to run on v1.13.4 even if Cognigy 2025.14.0 and its Qdrant dependency Helm chart are rolled back.

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and AI Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2025.15
Source: https://docs.cognigy.com/release-notes/2025.15



<Warning>
  As of June 30, 2025, `claude-3-opus` is deprecated. The expected removal date for this model is January 2026. [Anthropic recommends migrating](https://docs.anthropic.com/en/docs/about-claude/model-deprecations) to `claude-opus-4-0` before January 2026. We strongly recommend carrying out this migration since `claude-3-opus` support will be removed from future Cognigy versions.

  Refer to the list of all upcoming removals on the [Deprecations and Removals](/release-notes/deprecations-and-removals) page.
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2025.15.2

### Cognigy Voice Gateway

**Released** August 14, 2025

#### Bug Fixes

* Fixed the issue where a duplicate SIP `BYE` message was sent when the call was terminated

## 2025.15.1

### Cognigy

**Released** July 24, 2025

#### Bug Fixes

* Fixed the issue where only the `baseUrl` parameter was encoded in HTTP Request Nodes
* Fixed the issue where errors appeared on the Logs page each time a handover to a human agent occurred, although the handover worked as expected

### Cognigy Voice Gateway

**Released** August 12, 2025

#### Bug Fixes

* Fixed the issue where updating an application via a `PUT` API call to `/v1/Applications/<app_sid>` led to the removal of `call_hook`, `call_status_hook`, and `messaging_hook`. This issue occurred when these values weren't included in the request body

## 2025.15.0

### Cognigy

**Released** July 22, 2025

#### Improvements

* Removed the model selection for the deprecated GPT Conversation Node from the Generative AI Settings section in Manage > Settings
* Added the Legacy label to the `claude-3-opus` [model](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) to reflect its deprecation by Anthropic
* Added the capability to display typing indicators to end users in [Webchat v3](/webchat/v3/configuration#webchat-behavior) when using [Genesys Cloud Open Messaging](/ai/escalate/handover-reference/genesys-cloud-open-messaging)

#### Bug Fixes

* Fixed the issue where `api.getContext()` and `actions.getContext()` returned `null` for falsy values
* Fixed the issue where the platform-provided LLM wasn't the default setting for design-time features in Projects created from a template
* Fixed the issue where users couldn't clear the Knowledge Store selection in the AI Agent wizard after selecting a Knowledge Store
* Fixed the issue where errors weren't properly handled during the Node execution
* Fixed the issue where the Interaction Panel didn't display specific debug messages when the Flow included a Go To Node with the Parse Intents option activated

### Cognigy Voice Gateway

**Released** July 22, 2025

#### Improvements

* Added automatic trimming of unnecessary silence at the end of generated TTS audio. This improvement enhances the recognition of short spoken replies, such as Yes or No, by ensuring they're detected more quickly and accurately after the AI Agent finishes speaking

### Cognigy Webchat

**Released** July 22, 2025

#### Improvements

* Updated Webchat v3 and Demo Webchat to [v3.26.0](https://github.com/Cognigy/Webchat/releases/tag/v3.26.0)

### Cognigy Live Agent

**Released** July 22, 2025

#### Bug Fixes

* Fixed the issue where, during high load or when many conversations were created simultaneously, system messages about labels added by automation rules appeared as if added by a human agent in the format: `<human-agent-name> added <label-name>`. Now, labels added by automation rules are consistently shown in the chat as messages from Automation System in the format: `Automation System added <label-name>`
* Fixed the issue where audio messages and attachments were sent after the conversation had been resolved

### Cognigy Insights

**Released** July 22, 2025

#### Bug Fixes

* Fixed the issue where retrieving data through `analyticsdata.intent` didn't work properly
* Fixed the issue where Assertions didn't validate SMS data after the execution of Send SMS Nodes
* Fixed the issue where a session wasn't created in Insights when the user didn't respond or continued the conversation in Webchat v3

### Infrastructure

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and AI Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2025.16
Source: https://docs.cognigy.com/release-notes/2025.16



<Warning>
  * As of September 24, 2025, `gemini-1.5-flash` and `gemini-1.5-pro` are deprecated. [Google recommends migrating](https://cloud.google.com/vertex-ai/generative-ai/docs/learn/model-versions#legacy-stable) `gemini-1.5-flash` to `gemini-2.0-flash` and `gemini-1.5-pro` to `gemini-2.0-flash`. We strongly recommend carrying out this migration since support for `gemini-1.5-flash` and `gemini-1.5-pro` will be removed from future Cognigy versions.

  * As of April 21, 2025, `gemini-1.0-pro` and `text-bison-001` are deprecated. Ensure you aren't using these models. [Google recommends migrating](https://cloud.google.com/vertex-ai/generative-ai/docs/learn/model-versions#legacy-stable) to Gemini 2.0 Flash Lite. Support for `gemini-1.0-pro` or `text-bison-001` has been removed.

  * On July 21, 2025, Anthropic retired `claude-3-sonnet`. Ensure you aren't using this model. [Anthropic recommends migrating](https://docs.anthropic.com/en/docs/about-claude/model-deprecations#2025-01-21%3A-claude-2%2C-claude-2-1%2C-and-claude-sonnet-3-models) to `claude-sonnet-4`. Support for `claude-3-sonnet` has been removed.

  Refer to the list of all upcoming removals on the [Deprecations and Removals](/release-notes/deprecations-and-removals) page.
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2025.16.1

### Cognigy Voice Gateway

**Released** August 14, 2025

#### Bug Fixes

* Fixed the issue where a duplicate SIP `BYE` message was sent when the call was terminated

## 2025.16.0

### Cognigy

**Released** August 5, 2025

#### Features

##### Visual Warnings for Deprecated and Removed LLM Models

Cognigy now highlights [outdated LLM models](/ai/agents/develop/gen-ai-and-llms/llms#deprecations-and-removals) in the interface to help you stay informed about deprecations and removals, ensuring your AI Agents run smoothly

#### Improvements

* Added the ability to override the Flow Locale with a custom language for the Salesforce MIAW handover provider. To use this feature, go to the Handover to Agent Node, select Salesforce MIAW from the Handover Provider list, and add the language in the Conversation Routing Attributes fields using the format: `"language": "es_MX"`
* Deprecated the `actions` object in the Code Node. To achieve the same functionality, use `api` instead
* Improved handling of attachments in [Salesforce MIAW handovers](/ai/escalate/handover-reference/salesforce-miaw) for Webchat v3. Webchat v3 now correctly displays attachments sent by human agents on the Salesforce side, ensuring a smoother and more consistent user experience during handovers
* Added the possibility to handle routing errors while making handovers to [Salesforce MIAW](/ai/escalate/handover-reference/salesforce-miaw) by selecting the Handover Status type in the Lookup Node
* Added the capability to display [typing indicators](/webchat/v3/configuration#webchat-behavior) to end users in Webchat when using [Salesforce MIAW](/ai/escalate/handover-reference/salesforce-miaw)
* Introduced the Analytics 3.0 API. Now, you can use the enhanced call and conversation counters via the [Analytics 3.0 endpoints](https://api-trial.cognigy.ai/openapi#get-/v3.0/callcounter)
* Added the capability to retrieve all job configurations and the associated tools of AI Agents via the Cognigy API endpoint [`GET /v2.0/aiagents/{aiAgentId}/jobs`](https://api-trial.cognigy.ai/openapi#get-/v2.0/aiagents/-aiAgentId-/jobs)
* Added a grace period to proactively terminate Salesforce MIAW chat sessions before the access token expires. This approach prevents unnecessary API calls to expired chat sessions

#### Bug Fixes

* Fixed the issue where whitespace wasn't trimmed from URLs entered when creating a Knowledge Source from web pages. Now, leading and trailing whitespaces are automatically removed during URL input validation to prevent this error
* Fixed the issue where the Add Transcript Step Node with Tool Call type and Assistant role didn't add entries to the conversation transcript as expected
* Fixed the issue where the example AI Agent template was incorrect in the Cognigy API documentation

### Cognigy Voice Gateway

**Released** August 5, 2025

#### Bug Fixes

* Fixed the issue where updating an application via a `PUT` API call to `/v1/Applications/<app_sid>` led to the removal of `call_hook`, `call_status_hook`, and `messaging_hook`. This issue occurred when these values weren't included in the request body

### Cognigy Webchat

**Released** August 5, 2025

#### Improvements

* Updated Webchat v3 and Demo Webchat to [v3.28.0](https://github.com/Cognigy/Webchat/releases/tag/v3.28.0)

### Cognigy Live Agent

**Released** August 5, 2025

#### Improvements

* Enhanced conversation routing. Now, Live Agent reassigns conversations with the status [Snoozed (until next reply)](/live-agent/conversation/conversation-routing/automatic-mode) and unassigns them if the human agent has reached their conversation limit

#### Bug Fixes

* Fixed the issue that prevented the profile image from displaying in the Contact Profile Details, even when a valid image URL was set

### Cognigy Insights

**Released** August 5, 2025

#### Bug Fixes

* Fixed the issue where users assigned to many Projects had a JWT that exceeded the HTTP header size limit, resulting in authentication failures with a 431 error

### Infrastructure

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2025.17
Source: https://docs.cognigy.com/release-notes/2025.17



<Warning>
  * As of September 24, 2025, `gemini-1.5-flash` and `gemini-1.5-pro` are deprecated. [Google recommends migrating](https://cloud.google.com/vertex-ai/generative-ai/docs/learn/model-versions#legacy-stable) `gemini-1.5-flash` to `gemini-2.0-flash` and `gemini-1.5-pro` to `gemini-2.0-flash`. We strongly recommend carrying out this migration since the support for `gemini-1.5-flash` and `gemini-1.5-pro` will be removed from future Cognigy versions.

  Refer to the list of all upcoming removals on the [Deprecations and Removals](/release-notes/deprecations-and-removals) page.
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2025.17.2

### Cognigy

**Released** September 17, 2025

#### Improvements

* Added the End Handover on Participant Change setting to the Handover to Agent Node for the Salesforce MIAW provider. Enabling this setting ends the handover when an agent is removed, relying on the participant change event because [Salesforce can't always guarantee receipt](https://help.salesforce.com/s/articleView?id=000397177\&type=1) of the conversation close event

#### Bug Fixes

* Fixed the issue where configuring OAuth credentials in a Genesys Endpoint failed with the error `Validation failed. Field oAuth2Connection was not allowed`
* Fixed the issue where user IDs with big integer values weren't processed correctly during handover

## 2025.17.0

### Cognigy

**Released** August 19, 2025

#### Improvements

* Added a legacy tag to ElevenLabs legacy voice names. ElevenLabs legacy voices aren't available when using the ElevenLabs EU endpoint
* Optimized how Contact Profile data is fetched for the RingCentral Engage handover provider while sending user messages and ensured the data is present when the handover starts
* Added support for [OpenAI-compatible](/ai/agents/develop/gen-ai-and-llms/providers/openai-compatible) embedding models
* Added the Enable User Connects Message and the Enable User Disconnects Message options to the [Salesforce MIAW](/ai/escalate/handover-reference/salesforce-miaw#configuration-on-the-cognigy-side) in the [Handover to Agent Node](/ai/agents/develop/node-reference/service/handover-to-agent#parameters). These options allow a message to be sent to the human agent in Salesforce MIAW when the user reconnects to or disconnects from the conversation
* Added logic to prevent the end of [Salesforce MIAW](/ai/escalate/handover-reference/salesforce-miaw) handovers when the human agent is removed from the conversation. With this logic, you no longer need to set the `DISABLE_ABORT_HANDOVER_ON_SALESFORCE_MIAW_PARTICIPANT_CHANGE` feature flag
* Added OAuth 2.0 support for the [Azure OpenAI LLM provider](/ai/agents/develop/gen-ai-and-llms/providers/microsoft-azure-openai#add-models-via-gui)
* Added the Include Rich Media Context option to the [AI Agent](/ai/agents/develop/node-reference/ai/ai-agent#parent-node), [LLM Prompt](/ai/agents/develop/node-reference/service/llm-prompt#parameters), and [Get Transcript](/ai/agents/develop/node-reference/service/get-transcript#parameters) Nodes. With this option, you can provide either structured content or textual description to AI Agents as more context for rich media like Text with Buttons and Quick Replies

#### Bug Fixes

* Fixed the issue where editing values of type Number in the Contact Profiles sent strings in the payload, leading to failed requests
* Fixed the issue where the Check Agent Availability Node settings from restored Snapshots or imported Packages didn't include previously configured handover providers
* Fixed the issue with handover providers from restored Snapshots or imported Packages. Previously, handover providers contained the user ID of the user who originally created the Snapshot or Package. Now, handover providers correctly use the user ID of the user who restored the Snapshot or imported the Package
* Fixed the issue where the search functionality of the Search Extract Output Node threw content filter errors for user inputs that didn't violate the content policy
* Fixed the issue where restoring Snapshots displayed an empty handover provider for Locales other than the default
* Fixed the issue where YouTube videos didn't load in Demo Webchat
* Fixed the issue where debug messages displayed only strings and JSON-formatted code
* Fixed the issue where the Get Token Node execution failed when used with the Microsoft Bot Framework with a multitenant bot

### Cognigy Voice Gateway

**Released** August 19, 2025

#### Features

##### ElevenLabs and Deepgram EU Endpoint Support

Introduced support for the endpoint for [ElevenLabs](/voice-gateway/webapp/speech-services) and [Deepgram](/voice-gateway/webapp/speech-services) services located in the EU. These endpoints host the data processed by ElevenLabs and Deepgram in the EU and can help meet the necessary requirements of your data residency policy. To use the EU endpoint, select this option in the speech services settings in the Voice Gateway Self-Service Portal.

#### Bug Fixes

* Fixed the issue in the Voice Gateway Self-Service Portal where the wrong carrier was assigned to a phone number even though the UI showed the correct carrier
* Fixed the issue where Continuous ASR remained active when sending a Gather if multiple Set Session Config Nodes were used

### Cognigy Webchat

**Released** August 19, 2025

#### Bug Fixes

* Fix the issue where user text messages were sanitized even when `disableTextInputSanitization` was set to `true`

### Cognigy Live Agent

**Released** August 19, 2025

#### Bug Fixes

* Fixed the issue where unassigned conversations weren't automatically assigned to available human agents

### Cognigy Insights

**Released** August 19, 2025

#### Bug Fixes

* Fixed the issue where the Transcript Explorer incorrectly displayed messages about Flow switching in debug mode, even when no Flow switch had occurred

### Infrastructure

#### Cognigy Live Agent

* Added the capability to use an existing Persistent Volume Claim (PVC) for storage. Now, users who have an on-premises installation and a volume claim in Microsoft Azure or AWS Elastic Kubernetes Service can set the PVC in `existingClaim` in `values.yaml`

#### Cognigy Voice Gateway

* Added the capability to set the database host and port in the `heplify-server` and `homer` configuration

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2025.18
Source: https://docs.cognigy.com/release-notes/2025.18



<Warning>
  * As of September 24, 2025, `gemini-1.5-flash` and `gemini-1.5-pro` are deprecated. [Google recommends migrating](https://cloud.google.com/vertex-ai/generative-ai/docs/learn/model-versions#legacy-stable) `gemini-1.5-flash` to `gemini-2.0-flash` and `gemini-1.5-pro` to `gemini-2.0-flash`. We strongly recommend carrying out this migration since the support for `gemini-1.5-flash` and `gemini-1.5-pro` will be removed from future Cognigy versions.

  Refer to the list of all upcoming removals on the [Deprecations and Removals](/release-notes/deprecations-and-removals) page.
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2025.18.2

### Cognigy

**Released** September 17, 2025

#### Improvements

* Added the End Handover on Participant Change setting to the Handover to Agent Node for the Salesforce MIAW provider. Enabling this setting ends the handover when an agent is removed, relying on the participant change event because [Salesforce can't always guarantee receipt](https://help.salesforce.com/s/articleView?id=000397177\&type=1) of the conversation close event

#### Bug Fixes

* Fixed the issue where configuring OAuth credentials in a Genesys Endpoint failed with the error `Validation failed. Field oAuth2Connection was not allowed`
* Fixed the issue where user IDs with big integer values weren't processed correctly during handover

## 2025.18.1

### Cognigy Voice Gateway

**Released** September 12, 2025

#### Bug Fixes

* Fixed the issue where it wasn't possible to enter new credentials when you selected Deepgram as a speech provider in the Voice Gateway Self-Service Portal for Cognigy SaaS installations

## 2025.18.0

### Cognigy

**Released** September 2, 2025

#### Improvements

* Removed the Beta flag from the new [Handover Providers](/ai/escalate/migration) interface

#### Bug Fixes

* Fixed the issue where the Absorb Context parameter in the Go To Node didn't merge the source and destination Flow contexts as expected
* Fixed the issue in the Cognigy API documentation. Now, GET APIs show the response for the `next` and `previous` cursors
* Fixed the spacing, sizing, and styling issues in the Lexicon and Any Slot input fields to align them with System Slot fields
* Fixed the issue where goals weren't created in Analytics when uploading a Snapshot, which caused missing data in the Goals Performance chart in Insights
* Fixed the issue where users with `projectAdmin` roles couldn't upload a file within the Project they administer. For example, when creating an LLM connection using the Upload JSON File option
* Fixed the issue where users couldn't save CognigyScript expressions in Node fields that support Manual Input, such as the Select Flow and Select Node fields in the Go To Node
* Fixed the issue where the AI Enhanced Output feature failed on Azure-hosted LLMs due to content policy violations
* Fixed the issue where the AI Enhanced Output feature didn't work with the following models: `gpt-4.1`, `gpt-4.1-nano`, `gpt-4.1-mini`, `claude-sonnet-4-0`, `claude-opus-4-0`
* Fixed the issue in the Endpoint settings where the Migrate to Handover Providers button for the Live Agent saved the pre-configured account instead of the manually entered configuration
* Fixed the issue in the Question Node where the Large Language Model list didn't display available models in the following scenarios:
  * When LLM Prompt was selected under Reprompt Options > Reprompt Method
  * When LLM-Extract Entity was selected from the Question Type list, the LLM Entity Extraction Options section showed an empty model list
* Fixed 500 errors and stack traces caused by malformed OData URLs
* Fixed the issue where additional category IDs weren't applied during handover to RingCentral Engage

### Cognigy Voice Gateway

**Released** September 2, 2025

#### Bug Fixes

* Fixed the `termination_reason` field in the Call Detail Record (CDR) for failed calls in `sbc-inbound`. Now, the `caller abandoned` reason correctly indicates that the caller initiated the call but canceled it before it was answered
* Fixed the issue where multiple `BYE` messages were sent when the voice AI Agent hung up a call in `sbc-inbound`. This fix ensures that each call termination generates only a single event, preventing duplicate call termination notifications

### Cognigy Webchat

**Released** September 2, 2025

#### Features

##### Webchat v3 npm Package

* With [version 3.29.1](https://github.com/Cognigy/Webchat/releases/tag/v3.29.1), Webchat v3 is now officially available as an [npm package](/webchat/v3/embedding/npm). This change makes it easier for developers to integrate, update, and manage Webchat v3 in modern web applications.

#### Bug Fixes

* Fixed the issue in Webchat v3 to correctly track conversations with or without a channel property, showing distinct icons and channel names in Transcript Explorer

### Cognigy Live Agent

**Released** September 2, 2025

#### Bug Fixes

* Fixed the issue where the end user's profile picture didn't update in the Contact Profile details and the Conversations list. Now, profile pictures display correctly throughout the Live Agent interface

### Cognigy Insights

**Released** September 2, 2025

#### Improvements

* Added support for additional country flags in the Top Sessions list on the Business dashboard

#### Bug Fixes

* Fixed the issue where deactivating a Contact Profile caused inconsistencies in Analytics data
* Fixed the issue where empty messages appeared in the Transcript Explorer
* Fixed the issue where WhatsApp messages appeared as Webchat messages in the Transcript Explorer
* Fixed the issue where deleting Analytics records caused inconsistencies
* Fixed the issue where the Analytics database failed due to a field length limitation
* Fixed the issue where searching for standard session IDs or contact IDs in the Transcript Explorer returned multiple unrelated results instead of an exact match

### Infrastructure

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2025.19
Source: https://docs.cognigy.com/release-notes/2025.19



<Warning>
  * As of September 24, 2025, `gemini-1.5-flash` and `gemini-1.5-pro` are deprecated. [Google recommends migrating](https://cloud.google.com/vertex-ai/generative-ai/docs/learn/model-versions#legacy-stable) `gemini-1.5-flash` to `gemini-2.0-flash` and `gemini-1.5-pro` to `gemini-2.0-flash`. We strongly recommend carrying out this migration since the support for `gemini-1.5-flash` and `gemini-1.5-pro` will be removed from future Cognigy versions.

  Refer to the list of all upcoming removals on the [Deprecations and Removals](/release-notes/deprecations-and-removals) page.
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2025.19.0

### Cognigy

**Released** September 16, 2025

#### Features

##### Enhanced AI Agent Job and Tool Management

Introduced a sleek, user-friendly interface for [managing AI Agents' jobs and tools](/ai/agents/develop/manage-ai-agents), with Flow Logic for AI Agents accessible via the Advanced Editor button, making workflows more intuitive and easier to configure.

#### Improvements

* Added the `message id` key to the [`transcript`](/ai/agents/develop/ai-agent-memory/input#nested-objects) Input object to uniquely identify each message within a conversation transcript
* Added a loading banner to show progress when creating Knowledge Sources
* Added the End Handover on Participant Change setting to the Handover to Agent Node for the Salesforce MIAW provider. Enabling this setting ends the handover when an agent is removed, relying on the participant change event because [Salesforce can't always guarantee receipt](https://help.salesforce.com/s/articleView?id=000397177\&type=1) of the conversation close event
* Enhanced Playbook text assertions with [advanced fuzzy search](/ai/agents/test/assertions) that handles case differences, extra spaces, tabs, newlines, and escaped characters

#### Bug Fixes

* Fixed the issue where pressing backspace in the JSON editor of an Agent Copilot Grid Node redirected the screen to the Flow editor
* Fixed the issue where configuring OAuth credentials in a Genesys Endpoint failed with the error `Validation failed. Field oAuth2Connection was not allowed`
* Fixed the issue in the Endpoint settings where password managers were automatically filling the Client Secret field in the Handover Settings section for Genesys Cloud Open Messaging
* Fixed the issue in the Endpoint settings where Macedonian wasn't correctly translated when auto-detection was enabled in the Real Time Translation Settings section
* Fixed the issue where design-time features didn't work with `gpt-4` family models, `claude-sonnet-4-0`, and `claude-opus-4-0` models
* Fixed the issue where tools from child AI Agent Nodes placed below a Tool Node were also returned as part of the parent Agent's tools, causing them to appear in both lists. The endpoint now returns only the jobs that belong to the specified parent `agentId`
* Fixed the issue where user IDs with big integer values weren't processed correctly during handover

### Cognigy Webchat

**Released** September 16, 2025

#### Improvements

* Updated Webchat v3 and Demo Webchat to [v3.30.1](https://github.com/Cognigy/Webchat/releases/tag/v3.30.1)

### Cognigy Live Agent

**Released** September 16, 2025

#### Improvements

* Added the [TTL for conversations](/live-agent/settings/account-settings#ttl-for-conversations) to the Account Settings, allowing automatic deletion of old conversations from the UI and the Live Agent OData Endpoint

#### Bug Fixes

* Fixed the issue where Japanese translations were incomplete in the UI. Now all strings are fully translated
* Fixed the issue where activity messages about conversation assignments weren't translated to the account locale
* Fixed the issue where auto-assigned conversations to an online agent were treated as self-assignments, causing email notifications to be skipped

### Cognigy Insights

**Released** September 16, 2025

#### Improvements

* Removed the charts related to the understood/misunderstood message concept from the Overview and NLU Performance reports to simplify the reports and focus on more actionable metrics
* Removed the Top Tasks chart from the Engagement report because the chart was already included in the Overview report, avoiding duplication and improving clarity

#### Bug Fixes

* Fixed the issue where Voice Gateway messages didn't show up in the Transcript Explorer
* Fixed the issue where the Playbook editor execution didn't store Analytics data
* Fixed the issue where Insights failed to load due to a missing role during validation

### Infrastructure

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2025.20
Source: https://docs.cognigy.com/release-notes/2025.20



<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2025.20.2

### Cognigy Voice Gateway

**Released** October 15, 2025

#### Improvements

* Removed the [`CALL_RECONNECTED`](/voice-gateway/references/events/overview) event, which was previously sent to the Flow after a WebSocket reconnection. This change prevents it from registering as user input and causing unintended behavior

## 2025.20.1

### Cognigy

**Released** October 7, 2025

#### Bug Fixes

* Fixed the issue where certain integrations broke due to reliance on an undocumented property in the `agentReply` message

### Infrastructure

**Released** October 1, 2025

### Cognigy Voice Gateway

* Updated the value in `{{ .Values.jobs.image }}` from `vg-alpine-curl:3.20` to `vg-alpine-k8s-tools:3.22`. This change was done because the image used in the `audit-logs-cleanup` CronJob and in the `audit-cleanup-admin-extractor` Job were no longer accessible in the public [docker.io](https://www.docker.com/) registry

## 2025.20.0

### Cognigy

**Released** September 30, 2025

#### Features

##### Webhook Logging for AI Agent and LLM Prompt Nodes

Introduced the [Send Request Logs to Webhook](/ai/agents/develop/node-reference/ai/ai-agent) option to the Debug Settings in AI Agent and LLM Prompt Nodes. Enabling this option allows developers to capture and forward request logs to a specified webhook endpoint for debugging and monitoring purposes, helping to troubleshoot issues more effectively.

#### Improvements

* Added the [Custom Headers](/ai/agents/develop/node-reference/ai/ai-agent) parameter to the Advanced section in the MCP Tool Node. This parameter allows you to specify any key-value pairs to be added to your MCP connection headers
* Extended the [Textual Description](/ai/agents/develop/node-reference/basic/say) parameter capabilities for the output types: Video, Audio, Image, Adaptive Card, Gallery, and List. For these types, you can now include text from the Textual Description parameter in the prompt or transcript
* Updated the JSON example in the Source Data parameter of the [Fuzzy Search Node](/ai/agents/develop/node-reference/ai/fuzzy-search). This update improves clarity for users when configuring the Node and ensures the example aligns with common practical scenarios
* Added new audit events. Now, Cognigy OData requests are logged as `odataRequest` in the Admin Center for better transparency and monitoring
* Removed the `FEATURE_ENABLE_OAUTH2_AZURE_CONNECTION_WHITELIST` feature flag for the OAuth2 authentication method in the [Azure OpenAI LLM configuration](/ai/agents/develop/gen-ai-and-llms/providers/microsoft-azure-openai)

#### Bug Fixes

* Fixed the issue where JSON responses weren't parsed correctly when using non-streamed requests to Azure OpenAI
* Fixed the issue where a deleted Project couldn't be removed from global resources
* Fixed the issue where uncollapsed Knowledge Chunks in the Chunk editor were sometimes overlapping
* Fixed the issue where long Knowledge Source titles and descriptions were displayed incorrectly in the Chunk editor
* Fixed the issue when updating an email using the Update Profile Node. Now, the system correctly saves and displays the email in the Contact Profile and the Cognigy OData Endpoint instead of showing it as `undefined`

### Cognigy Voice Gateway

**Released** September 30, 2025

#### Features

##### Advanced RTP Traffic Routing

The Anchor Media parameter is now [Media Path](/ai/agents/develop/node-reference/voice/voice-gateway/transfer), giving enhanced control over RTP routing for monitoring, transcoding, security, and NAT handling. The Full Media and Partial Media options remain compatible with previous settings (On/Off), while No Media allows direct routing without a media platform.

### Cognigy Live Agent

**Released** September 30, 2025

#### Improvements

* Introduced [version 1.1](/live-agent/tools/odata-endpoint) for the Live Agent OData endpoint

### Cognigy Insights

**Released** September 30, 2025

#### Improvements

* Fixed the issue where Snapshots were listed twice when the Snapshots filter was activated
* Fixed the issue where empty messages containing only timestamps were shown in the Transcript Explorer

### Infrastructure

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2025.21
Source: https://docs.cognigy.com/release-notes/2025.21



<Note>
  Qdrant has been updated from v1.13.4 to v1.15.4. For more information, refer to the [Infrastructure](#qdrant-upgrade-from-v1134-to-v1154) section.
</Note>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2025.21.3

### Cognigy Live Agent

**Released** October 24, 2025

#### Bug Fixes

* Fixed the issue where data couldn't be saved or read properly when a custom namespace for a Redis connection pool is used

## 2025.21.2

### Cognigy Live Agent

**Released** October 23, 2025

#### Bug Fixes

* Fixed the issue where the database migration job failed during the installation of a newly set-up Live Agent

## 2025.21.1

### Cognigy

**Released** October 15, 2025

#### Bug Fixes

* Fixed the release rollout process by deactivating a background job

### Cognigy Live Agent

**Released** October 15, 2025

#### Bug Fixes

* Fixed the connection issue that prevented Live Agent from starting

## 2025.21.0

### Cognigy

**Released** October 14, 2025

#### Features

#### Knowledge Connectors

* Introduced [Knowledge Connectors](/ai/agents/develop/knowledge-ai/knowledge-connectors) in Knowledge AI. This feature allows you to ingest knowledge from external sources, such as databases, APIs, or document libraries, and use it directly in your Flows. You can use prebuilt Knowledge Connectors from the Marketplace, such as Confluence and Diffbot, or create custom ones using the Extension framework.

#### Improvements

* Allowed arbitrary custom URLs for [Azure OpenAI LLMs](/ai/agents/develop/gen-ai-and-llms/providers/microsoft-azure-openai), so the URL is no longer restricted to Azure OpenAI's domain

#### Bug Fixes

* Fixed the issue where `console.log` statements in the NLU connector code logged to the system log instead of the app log
* Fixed `items` in the `GET /v2.0/conversations/{sessionId}` response to be an array instead of an object
* Fixed the issue preventing the export of manual Knowledge Sources from the UI
* Fixed the issue where Knowledge Source file names weren't properly decoded
* Fixed the issue where some Knowledge Chunks were created without metadata when processing files with Azure Document Intelligence
* Fixed the issue where the time zone wasn't considered when calculating the UTC offset in date Slots
* Fixed the issue where the Azure Bot Services endpoint displayed empty messages in the Microsoft Teams chat
* Fixed the issue where the `getConversationCounter` v3 API returned incorrect counts for sessions over 50 conversations. Data will be available the day after the v2025.21.0 upgrade, once records are corrected

### Cognigy Voice Gateway

**Released** October 14, 2025

#### Improvements

* Removed the [`CALL_RECONNECTED`](/voice-gateway/references/events/overview) event, which was previously sent to the Flow after a WebSocket reconnection. This change prevents it from registering as user input and causing unintended behavior

### Cognigy Live Agent

**Released** October 14, 2025

#### Bug Fixes

* Fixed the issue where the deletion process failed for conversations older than two years

### Infrastructure

#### Cognigy + Insights + Agent Copilot

##### Qdrant Upgrade from v1.13.4 to v1.15.4

For Cognigy SaaS where the Knowledge AI feature is enabled and Qdrant is deployed, upgrading Cognigy to version 2025.21.0 will automatically upgrade Qdrant to v1.15.4.

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2025.22
Source: https://docs.cognigy.com/release-notes/2025.22



<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2025.22.3

### Cognigy

**Released** November 20, 2025

#### Bug Fixes

* Fixed the issue where Microsoft Azure OpenAI models using the API Key connection type with custom URLs failed

## 2025.22.2

### Cognigy

**Released** November 5, 2025

#### Bug Fixes

* Fixed the issue where Snapshot deployments used the incorrect Project ID during Flow executions

### Cognigy Webchat

**Released** November 5, 2025

#### Bug Fixes

* Fixed the issue where Webchat displayed the human agent's name and logo from Genesys Open Messaging instead of the AI Agent's name and the Webchat logo

## 2025.22.1

### Cognigy

**Released** October 30, 2025

#### Bug Fixes

* Fixed the issue in the Helm chart that prevented the execution of Knowledge Connectors in the trial environment. For on-premises installations, set `persistentVolumeClaim.runtime.enabled` to `true` so that Knowledge Connectors work properly

## 2025.22.0

### Cognigy

**Released** October 28, 2025

#### Features

##### Cognigy AI Operations Center (AI Ops Center)

With [AI Ops Center](/ops-center/index), you can monitor and manage your AI Agents using real-time dashboards, performance insights, and proactive issue alerts. Full features are unlocked with a license.

#### Improvements

* Added [new LLMs](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature):
  * Azure OpenAI and OpenAI: `gpt-5`, `gpt-5-nano`, `gpt-5-mini`, `gpt-5-chat-latest`
  * Google Gemini: `gemini 2.5`, `gemini-2.5-flash`, `gemini-2.5-flash-lite`
* Added [a new event](/ai/escalate/handover-reference/genesys-cloud-open-messaging-with-amazon-eventbridge) to display the human agent's name and image in the conversation for the Genesys Open Messaging Handover provider with the Amazon EventBridge integration
* Added authorization and authentication actions to [Audit Events](/ai/administer/access/admin-center/audit-events) for successful and failed user login attempts
* Improved the debug message description for the [MCP Tool Node](/ai/agents/develop/node-reference/service/llm-prompt#mcp-tool). The debug message now indicates whether the tools were fetched from the cache or directly from the MCP Server
* Upgraded the Basic Parser to a more stable document processing solution to improve reliability, reduce maintenance overhead, and enhance security

#### Bug Fixes

* Fixed the Track Goal Name not displaying properly on the Node
* Resolved Anthropic responses in LLM prompt nodes occasionally displaying a JSON object instead of text output
* Updated the deprecation deadline for handover settings to January 2026

### Cognigy Voice Gateway

**Released** October 28, 2025

#### Improvements

* Added filtering of recent calls by application
* Improved reconnection behavior. Voice Gateway now reconnects faster and more reliably when it temporarily loses its connection to Cognigy

#### Bug Fixes

* Fixed SRTP Base64 padding to ensure compatibility with Zoom's crypto suite
* Fixed the issue where the Reason set on the Voice Gateway Endpoint for the Call Failover feature wasn't correctly added as an `X-Reason` header to SIP messages

### Cognigy Webchat

**Released** October 28, 2025

#### Bug Fixes

* Fixed the issue with streaming outputs in Demo Webchat

### Cognigy Live Agent

**Released** October 28, 2025

#### Bug Fixes

* Fixed the issue where the database migration job failed when installing a newly set-up Live Agent
* Fixed the issue where data couldn't be saved or read properly when a custom namespace for a Redis connection pool is used

### Cognigy Insights

**Released** October 28, 2025

#### Bug Fixes

* Fixed the Snapshots list loading issue in the Snapshot filter. The list now appears immediately and reliably when the filter is applied

### Infrastructure

#### Cognigy + Insights + Agent Copilot + AI Ops Center

##### Activate AI Ops Center

The activation process depends on the installation type:

* **SaaS.** All clusters include the Cognigy monitoring stack by default. Contact your Customer Success Manager to activate AI Ops Center.
* **On-Premises.** AI Ops Center is unavailable for on-premises installations. To conserve cluster resources, add the following to the localized `values.yaml` file:

  ```yaml theme={null}
  dbConnectionString:
    services:
      serviceSentinel:
        enabled: false

  serviceSentinel:
    enabled: false
  ```

#### Voice Gateway

* Enable Redis keep-alive ping. Add a new environment variable to `{{ .Values.cognigyEnv }}` named `REDIS_KEEP_ALIVE_INTERVAL_IN_SECONDS` with a default value of 10 seconds. All services with access to Redis will restart after applying this value.

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2025.23
Source: https://docs.cognigy.com/release-notes/2025.23



<Warning>
  The full removal of AudioCodes Nodes is planned for Q1 2026. After the removal, AI Agents using these Nodes will stop working properly. We strongly recommend migrating to [Voice Gateway Nodes](/ai/agents/develop/node-reference/voice/voice-gateway/overview) before the removal date.
</Warning>

<Warning>
  In Cognigy 2025.25.0, transformers will start using the `axios` library for handling HTTP requests internally instead of `request-promise`. This change occurs to improve security and maintainability. The following HTTP request options will still be supported:

  * **URL/Endpoint Configuration** — `uri`, `url`
  * **Request Method** — `method`
  * **Request Headers & Configuration** — `headers`
  * **Request Body/Data** — `form`, `json`
  * **Query Parameters** — `qs`, `qsStringifyOptions`
  * **Authentication** — `auth`
  * **Connection & Network** — `proxy`, `maxRedirects`, `gzip`, `encoding`

  If you use other HTTP request options, review your transformer code to ensure it is compatible with the new `axios` library before the release of Cognigy 2025.25.0.
</Warning>

<Warning>
  As of Cognigy 4.91.0, the Understood Messages concept is deprecated. Use the [Goals and Tasks](https://docs.cognigy.com/ai/analyze/goals-and-tasks/overview/) concepts to evaluate whether your AI Agent understood a particular phrase. In Cognigy 2025.19, the charts related to this concept were removed from the Insights reports.

  As of Cognigy 4.91.0, the `understood` field in the Cognigy OData endpoint is deprecated. The full removal of the `understood` field from the OData endpoint is planned for Cognigy 2026.1. Use the [Goals collection](https://docs.cognigy.com/ai/analyze/odata#goals) in the Cognigy OData endpoint.

  Refer to the list of all upcoming removals on the [Deprecations and Removals](https://docs.cognigy.com/release-notes/deprecations-and-removals/) page.
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2025.23.2

### Cognigy

**Released** December 3, 2025

#### Bug Fixes

* Fixed the issue where Cognigy OData audit events contained sensitive information

## 2025.23.1

### Cognigy

**Released** November 19, 2025

#### Bug Fixes

* Fixed the issue where Microsoft Azure OpenAI models using the API Key connection type with custom URLs failed
* Fixed the issue where a Tool Node couldn't be added to the LLM Prompt Node

### Cognigy Voice Gateway

**Released** December 3, 2025

#### Bug Fixes

* Fixed the issue in `feature-server` where the initial connection to Cognigy required more attempts than expected

## 2025.23.0

### Cognigy

**Released** November 11, 2025

#### Features

##### AI Agent Knowledge Tool

Introduced the knowledge tool for AI Agents. This feature allows you to easily add a knowledge tool to your AI Agent to access Knowledge Sources in the [AI Agent editor](/ai/agents/develop/manage-ai-agents), [AI Agent Node](/ai/agents/develop/node-reference/ai/ai-agent), or [LLM Prompt Node](/ai/agents/develop/node-reference/service/llm-prompt)

#### Improvements

* Added the option to define a custom header to authenticate through [OAuth 2.0 in the Azure OpenAI LLMs](/ai/agents/develop/gen-ai-and-llms/providers/microsoft-azure-openai)
* Added the [GET Latest Log Entries](https://docs.cognigy.com/api-reference-logs/logs/get-the-latest-log-entries) API endpoint to retrieve Cognigy logs with improved performance. By default, this endpoint has a time window of 2 hours. You can change the time window by setting the `FEATURE_ACTIVE_LOGGING_WITH_TAIL_CUT_OFF_HOURS` environment variable in the Helm chart. However, larger time windows may cause lower performance. If you need to retrieve logs over a very long period of time, for example, all your logs, you can use the [GET Log Entries](https://docs.cognigy.com/api-reference-logs/logs/get-log-entries) API endpoint. For better performance, pass the `startDate` and `endDate` parameters to the endpoint

#### Bug Fixes

* Fixed the issue where the Question Node didn't recognize email addresses containing the `old` substring
* Fixed the issue where missing or incorrect Intents were displayed in debug messages
* Fixed the issue where handover cancellation messages appeared inconsistently for human agents using Genesys Cloud Open Messaging
* Fixed the issue where user events messages were included in conversation transcripts
* Fixed the issue where Flows implemented from Snapshots checked the incorrect Project ID
* Fixed the issue where GET methods displayed `previous` and `next` parameters, and the `extension` parameter of the `/v2.0/flows/{flowId}/chart/nodes/{nodeId}` endpoint was listed as `enum` type. Now, the GET methods parameters were changed to `previousCursor` and `nextCursor`, and the `extension` parameter type was changed to `string`
* Fixed the issue where translated text didn't appear after the Locale change
* Fixed the issue where incorrect Knowledge Connectors were displayed for marketplace extensions
* Fixed the issue where the user interface and error messages incorrectly referred to Projects as Agents
* Fixed the issue where custom embedding models weren't listed correctly in the Project settings
* Fixed the issue where text assertions failed in Playbook runs that included messages with a button
* Fixed the issue where the loader position in the Knowledge Source list caused the scroll bar on the main list to stop working

### Cognigy Voice Gateway

**Released** November 11, 2025

#### Bug Fixes

* Fixed an issue where atmosphere sounds continued after executing a dial. Now, sounds stop immediately for a clean audio experience
* Fixed an issue where fallback transfers triggered by speech failover lacked the defined headers. Now, transfers include the correct headers

### Cognigy Webchat

**Released** November 11, 2025

#### Features

##### Custom Webchat Launcher Icon

Added the capability to [customize the Webchat v3 launcher icon](/webchat/v3/configuration). Now, you can select one of the 3 default icons or upload your own to better match your brand

#### Bug Fixes

* Fixed issue where Webchat v3 displayed the human agent's name and logo from Genesys Open Messaging instead of the AI Agent's name and the Webchat logo

### Cognigy Live Agent

**Released** November 11, 2025

#### Improvements

* Added translation logic for API availability messages. System messages such as `User left the conversation` and `User joined the conversation` are now displayed in the correct language

#### Bug Fixes

* Fixed the issue in the Supervisor Overview report where the conversation list could spin indefinitely when opened with an assignee\_id URL parameter before the agent list had loaded

### Cognigy AI Ops Center

**Released** November 11, 2025

#### Improvements

* Limited the options in the subcomponent filter to only show component-relevant options

#### Bug Fixes

* Fixed the issue where graphs displayed fractions on the vertical axis

### Cognigy Insights

**Released** November 11, 2025

#### Improvements

* Added automatic saving of Projects selected in the Business Dashboard when switching between dashboards

#### Bug Fixes

* Fixed the issue where Slot matches for the value `0` weren't rendered in the Transcript Explorer
* Fixed the issue where input data containing keys with dot notation, for example, `{"user.name": "value"}`, wasn't saved to the database and wasn't displayed in Insights

### Infrastructure

#### Voice Gateway

* Parameterize the Test Call Manager cron job schedule via `{{ .Values.testCallManager.cronjobSchedule }}`.

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2025.24
Source: https://docs.cognigy.com/release-notes/2025.24



<Warning>
  In Cognigy 2025.25.0, transformers will start using the `axios` library for handling HTTP requests internally instead of `request-promise`. For more details, read [Transformers Overview](/ai/for-developers/transformers/overview).
</Warning>

<Warning>
  The following Anthropic models have been removed from Cognigy because they are no longer supported by Anthropic: `claude-3-5-sonnet-20241022`, `claude-3-5-sonnet-latest`, `claude-instant-v1`, and `claude-v1-100k`.

  If you use these models in your Flows, replace them with available models from the [Model Support by Feature](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) list to ensure your AI Agents remain stable.
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2025.24.2

### Cognigy

**Released** December 5, 2025

#### Bug Fixes

* Fixed SSO and inconsistent login issues in Cognigy

### Cognigy Voice Gateway

**Released** December 3, 2025

#### Bug Fixes

* Fixed the issue in `api-server` and `helm chart` where `api-server` logged some errors due to an optional environment variable
* Fixed the issue in `feature-server` where the initial connection to Cognigy required more attempts than expected

## 2025.24.1

### Cognigy

**Released** December 3, 2025

#### Bug Fixes

* Fixed the issue where Cognigy OData audit events contained sensitive information

### Cognigy Voice Gateway

**Released** November 26, 2025

#### Bug Fixes

* Fixed the issue where Microsoft Azure's speech-to-text functions didn't work properly by rolling back the FreeSWITCH image from `cognigy.azurecr.io/vg-freeswitch-mrf:2024.24.0` to `cognigy.azurecr.io/vg-freeswitch-mrf:0.9.3`

## 2025.24.0

### Cognigy

**Released** November 25, 2025

#### Features

##### Simulator

Introduced the [Simulator](/ai/agents/test/simulator/overview) for AI Agents. The Simulator is an LLM-powered tool in Cognigy that allows users to model and test full conversational scenarios for LLM-based AI Agents. Use the Simulator to mimic real-world user interactions and evaluate the performance, behavior, and reliability of AI Agents in controlled environments. You can try the Simulator in the [trial environment](https://trial.cognigy.ai/).

##### New Handover Provider for storm from Content Guru

Introduced a new [handover provider for storm](/ai/escalate/handover-reference/storm), the contact center platform from Content Guru. With the storm handover connector, you can integrate Cognigy with storm, enabling end users to connect with human agents working in a contact center that uses storm.

### Improvements

* Retired [Anthropic's](/ai/agents/develop/gen-ai-and-llms/providers/anthropic) `claude-3-5-sonnet-latest`, `claude-instant-v1`, and `claude-v1-100k` models
* Added `base64Encode` and `base64Decode` [API functions](/ai/for-developers/code/api-functions) to Code Nodes
* Improved the visibility of hyperlinks in the [Interaction Panel](/ai/agents/test/interaction-panel/overview)
* Added an activity filter on the [Contact Profiles](/ai/agents/analyze/contact-profiles) page
* Introduced a [fallback value](/ai/agents/develop/nodes/overview#fallback-value-in-node-fields) indicator in the Node editor fields. This indicator helps users understand the fallback value that is used when no value is configured for the field

#### Bug Fixes

* Fixed the issue where resizing the Marketplace page prevented users from scrolling and displaying all Extensions
* Fixed the issue where icons indicating deprecation and retirement status for LLMs were hidden when the LLM name was longer than the Model Type column on the Large Language Models page
* Fixed the issue where logging in to Cognigy in a second environment with the same domain caused the user to be logged out of the first environment due to a lack of cookie isolation
* Fixed the issue where users couldn't add Tool Nodes to LLM Prompt Nodes
* Fixed the issue where a translation for the Generative Conversation Summary Node was missing
* Fixed the issue where dependent Flows didn't appear in the dependency list while packaging a Voice Gateway Endpoint
* Fixed the issue where the Global Connection page crashed when users with slow connections switched between pages
* Fixed the issue where the collaborator icon on a Node was hidden if there were no conversations on this Node
* Fixed the issue where real-time translations weren't working as expected when Live Agent sent messages to users
* Fixed the issue where the positioning of the loader in the Knowledge Sources page prevented the scroll bar of the Knowledge Source list from working
* Fixed the issue where retired models were displayed in Knowledge AI settings
* Fixed the issue where human agent names were shown for Genesys Cloud Open Messaging without customer consent
* Fixed the issue where Azure OpenAI models using API key connections with custom URLs didn't work properly

### Cognigy Live Agent

**Released** November 25, 2025

#### Improvements

* Added the [Profile Language option](/live-agent/profile-settings#profile-language) in the Profile settings to allow the human agent to set their preferred UI language. This option overrides the account language and translates email and push notifications into the selected language

#### Bug Fixes

* Fixed the issue where the right expanding arrow button of the chat panel wasn't visible when the Enable overriding AI Copilot preferences and Enable AI Copilot Workspace options were activated
* Fixed the issue where email delivery failed due to names including commas in the From field

### Cognigy AI Ops Center

**Released** November 25, 2025

#### Improvements

* Updated the [Handover Providers](/ops-center/monitor/metrics/handover-providers) dashboard: the Provider Forwarding Errors Count and API Call Request Duration charts were removed, and the Handover Requests and Handover Messages charts were added to track the number of handover attempts and end-user messages, ensuring smooth conversation transfers

#### Bug Fixes

* Fixed the issue where the General indicator was highlighted red when other subcomponents' indicators were highlighted red

### Infrastructure

#### Voice Gateway

* Improved Websocket stability for keep-alive pings, retries, and message orders

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2025.25
Source: https://docs.cognigy.com/release-notes/2025.25



<Warning>
  With this release, transformers start using the `axios` library for handling HTTP requests internally instead of `request-promise`. For more details, read [Transformers Overview](/ai/for-developers/transformers/overview).
</Warning>

<Warning>
  The following models have been removed from Cognigy because OpenAI and Microsoft Azure OpenAI no longer support them: `gpt-3.5-turbo`, `gpt-3.5-turbo-instruct`, and `gpt-4`.

  If you use these models in your Flows, replace them with available models from the [Model Support by Feature](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) list to ensure your AI Agents remain stable.
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2025.25.1

### Cognigy Live Agent

**Released** December 16, 2025

#### Bug Fixes

* Fixed the issue with the Cognigy OData endpoint where account-level filtering was not correctly applied to the `ReportingEvent` collection

### Cognigy Voice Gateway

**Released** December 10, 2025

#### Bug Fixes

* Fixed the issue where Microsoft Azure's speech services functions didn't work properly for on-premises installations by rolling back the FreeSWITCH image to `cognigy.azurecr.io/vg-freeswitch-mrf:0.9.3`

## 2025.25.0

### Cognigy

**Released** December 9, 2025

#### Features

##### Data Redaction

Introduced [data redaction](/ai/platform-features/data-redaction) to automatically redact any information, including Personally Identifiable Information (PII), from conversation logs and analytics.
This feature prevents exposure or misuse of predefined data while maintaining log readability and analytics accuracy.

##### New Tools for AI Agents

Introduced the [Send Email and Handover to AI Agent tools](/ai/agents/develop/manage-ai-agents), enabling AI Agents to send emails directly and transfer conversations to other AI Agents within workflows

##### New NiCE CXone AAH Endpoint

Provided full access to the [NiCE CXone AAH Endpoint](/agent-copilot/deploy/nice-aah) for all users, enabling human agents to integrate Agent Copilot into their NiCE contact center platform

#### Improvements

* Enabled [Simulator availability](/ai/agents/test/simulator/overview) across all SaaS installations
* Removed the beta flag for the [LLM Fallback](/ai/agents/develop/gen-ai-and-llms/fallback) feature
* Added the Tenant ID field and moved the Billing Time Zone to the Change Organization Information section in the [Management UI](/ai/administer/access/management-ui#manage-organizations)
* Removed the `gpt-3.5-turbo`, `gpt-3.5-turbo-instruct`, and `gpt-4` models
* Replaced the `request-promise` library with Axios for HTTP requests sent from the transformers
* Added support for custom hostnames when deploying the platform behind a CDN or firewall
* Updated logic for generating session IDs for Endpoints. Now, each Endpoint and third‑party account combination, for example, a WhatsApp Endpoint and a WhatsApp account, gets a unique session ID with no duplicates across Endpoints and Projects

#### Bug Fixes

* Fixed the issue where deprecated models were displayed without the Deprecated icon
* Fixed the issue where organization deletion didn't remove the NFS directories
* Fixed the issue where additional attributes in Genesys Open Messaging handovers were corrupted due to invalid data types
* Fixed the issue where the Storm handover provider wasn't working due to request verification issues
* Fixed the issue where the LLM Prompt tool Nodes couldn't be added
* Fixed the issue where custom models and request options for Google Gemini weren't forwarded to the provider
* Fixed the issue where Azure OpenAI LLMs using API key connections with custom URLs failed
* Fixed the issue where the custom attributes in Genesys prevented handovers from being created when they contained nested JSON attributes

### Cognigy Voice Gateway

**Released** December 9, 2025

#### Improvements

* Improved connection handling when multiple WebSocket connection attempts are made, reducing dropped and failed connections
* Improved handling of DTMF signals and Barge-in scenarios, making voice interactions more responsive and accurate

#### Bug Fixes

* Fixed the issue with Silence Overlay and Atmosphere sounds, improving call reliability and user experience
* Fixed the HTTP session failures, such as those occurring when starting a SIPREC session, ensuring reliable call recording and signaling

### Cognigy Webchat

**Released** December 9, 2025

#### Improvements

* Updated Demo Webchat to 3.34.0

### Cognigy Live Agent

**Released** December 9, 2025

#### Features

##### Escalation to Supervisors

Introduced the [escalation to supervisors](/live-agent/settings/inboxes#escalate-conversations-to-supervisors) feature. This feature allows human agents to escalate conversations from the Live Agent interface to a designated supervisor. Supervisors can then take over the conversation, ensuring faster resolution of complex issues and better support management.

##### Live Agent OData Endpoint v2.0

Introduced [Live Agent OData Endpoint version 2.0](/live-agent/tools/odata-endpoint), featuring enhanced functionality and a new teams collection for managing team data.

#### Bug Fixes

* Fixed the issue where memories in Contact Profiles weren't in a human-readable format
* Fixed the issue causing evaluation errors in Excel OData Power Query Editor

### Cognigy Insights

**Released** December 9, 2025

#### Improvements

* Added Markdown rendering support to the Transcript Explorer
* Updated the session's Locale information when switching Locales in the Flow

#### Bug Fixes

* Fixed the issue where a red error message appeared in the Transcript Explorer when quick reply texts were empty
* Fixed the issue where multiselect filters in the Transcript Explorer didn't work as expected

### Infrastructure

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2026.1
Source: https://docs.cognigy.com/release-notes/2026.1



<Warning>
  The `understood` field has been removed from the Cognigy OData endpoint. Use the [Goals collection](/ai/agents/analyze/odata#goals) in the Cognigy OData endpoint.
</Warning>

<Warning>
  As of Cognigy 4.83.0, Webchat v2 is deprecated. The full removal of Webchat v2 is planned for February 2026. Make sure you migrate to [Webchat v3](/webchat/migration) before the removal date.
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2026.1.5

### Cognigy

**Released** January 26, 2026

#### Bug Fixes

* Fixed the issue where users couldn't view the Simulation dashboard due to the `Unable to fetch Simulation Batches` error

## 2026.1.4

### Cognigy

**Released** January 23, 2026

#### Bug Fixes

* Fixed the issue where importing Knowledge Stores via Packages failed

## 2026.1.3

### Cognigy

**Released** January 22, 2026

#### Bug Fixes

* Fixed the issue where Set Session Config Nodes generated incorrect error logs

## 2026.1.2

### Cognigy

**Released** January 20, 2026

#### Bug Fixes

* Fixed the issue where Deepgram STT models weren't applied correctly

## 2026.1.1

### Cognigy

**Released** January 14, 2026

#### Improvements

* Improved service stability and performance

### Cognigy Voice Gateway

**Released** January 20, 2026

#### Bug Fixes

* Rolled back `vg-drachtio-server` due to an issue with handling umlauts and emojis in SIP messages

## 2026.1.0

### Cognigy

**Released** January 13, 2026

#### Features

##### Mocking

Introduced the [Mocking feature](/ai/agents/test/mocking). Mocking lets you change the behavior of a Node by replacing its logic with mock code execution without changing the original Node settings. This approach allows you to test specific Flow conditions in the Simulator, the Interaction Panel, and Endpoints without reconfiguring Nodes.

##### New Handover to Human Agent Tool

Introduced the [Handover to Human Agent tool](/ai/agents/develop/manage-ai-agents#handover-to-human-agent), allowing AI Agents to transfer conversations to human agents via handover providers if they can't address the user's request.

#### Improvements

* Added the Email Sender Name field to the [Email Notification Node](/ai/agents/develop/node-reference/data/email-notification)
* Published the [Confluence Extension 5.0.2](https://github.com/Cognigy/Extensions/releases/tag/confluence502) and [Diffbot Extension 1.0.2](https://github.com/Cognigy/Extensions/releases/tag/diffbot102) on the Cognigy Marketplace to address critical vulnerabilities of previous versions

#### Bug Fixes

* Fixed the issue where the max token limit in Cognigy didn't allow custom LLM models to use the max token values set by the user
* Fixed the issue where Amazon Bedrock custom options didn't work properly
* Fixed the issue where MCP tools didn't support custom headers containing CognigyScript
* Fixed the issue where validation failed when saving AI Agent Nodes
* Fixed the issue where Markdown formatting wasn't rendered in messages in the Simulator
* Fixed the issue where the `.ack()` function was called in a different `amqp` channel than the one that received a message
* Fixed the issue where the goals list on the Create Scenario page didn't display the existing goals
* Fixed the issue where handover providers' icons weren't displayed in the Handover Provider list in the Check Agent Availability Node
* Fixed the issue where Packages including a period in the name missed the `.zip` extension in the file name
* Fixed the issue where HTTP requests in Endpoint transformers broke when the payload included `json: true`
* Fixed the issue where the Contact Profiles page in a Project displayed Contact Profiles from other Projects

### Cognigy Voice Gateway

**Released** January 13, 2026

#### Bug Fixes

* Fixed the issue where `CALL_COMPLETED` events were missing for transferred calls when the carrier ended the call first
* Fixed the issue where TTS streaming failed when using ElevenLabs with non-global regions by limiting the TTS streaming to global regions

### Cognigy Live Agent

**Released** January 13, 2026

#### Improvements

* Added escalated conversations as a metric to the Open Conversations indicator in the [Supervisor Overview report](/live-agent/reports/supervisor-overview#open-conversations)

### Cognigy Insights

**Released** January 13, 2026

#### Bug Fixes

* Fixed the issue where the [GET all Conversations API method](/api-reference/conversations/get-conversations) didn't include the `endpointName`, `ratings`, and `ratingComments` properties in the response
* Fixed the issue where goals and Snapshots filters couldn't be applied at the same time

### Infrastructure

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2026.10
Source: https://docs.cognigy.com/release-notes/2026.10



<Warning>
  Removals and deprecations:

  * Support for Aleph Alpha's `luminous` and `luminous-embedding-128` has been deprecated.
  * Support for Anthropic's `claude-3-haiku` model has been removed.
</Warning>

<Warning>
  As of Cognigy 2026.8.3, the `ingress.ingresController` feature flag controls whether to create Traefik middleware resources. For on-premises installations, refer to the [Infrastructure section in the Cognigy 2026.8.0 release notes](/release-notes/2026.8#infrastructure).
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2026.10.2

### Cognigy Voice Gateway

**Released** May 14, 2026

#### Bug Fixes

* Fixed the issue with barge-in where a caller's voice input was ignored if it contained fewer words than the minimum barge-in word count, causing the AI Agent to wait for the caller to speak again

## 2026.10.1

### Cognigy

**Released** May 18, 2026

#### Improvements

* Updated Redis container images from PhotonOS-based to Debian-based versions to ensure compliance with contractual requirements for the use of Redis

### Cognigy Voice Gateway

**Released** May 13, 2026

#### Bug Fixes

* Fixed the issue where the FreeSWITCH media server stopped responding. Now, it automatically restarts when it becomes unresponsive

## 2026.10.0

### Cognigy

**Released** May 12, 2026

#### Features

##### New AWS Bedrock LLMs

Introduced support for [AWS Bedrock's](/ai/agents/develop/gen-ai-and-llms/providers/amazon-bedrock) `amazon.nova-premier-v1:0` and `amazon.nova-2-lite-v1:0` models. This improves performance by allowing requests to be routed within a region or worldwide.

##### New OpenAI and Microsoft Azure OpenAI LLM

Introduced support for the `gpt-5.5` model from [OpenAI](/ai/agents/develop/gen-ai-and-llms/providers/openai) and [Microsoft Azure OpenAI](/ai/agents/develop/gen-ai-and-llms/providers/microsoft-azure-openai). This model allows for higher-performance AI capabilities across different use cases.

#### Improvements

* Increased the maximum number of success criteria in the [Simulator](/ai/agents/test/simulator/overview) from 5 to 10
* Moved the [Enable Mocking toggle](/ai/agents/test/mocking#activate-mocking) outside of the Advanced section of the Start Simulation panel in the Simulator for more visibility
* Added the [`GET /management/v2.0/system/license`](/api-reference/management-ui/get-the-license-via-management-ui) and [`POST /management/v2.0/system/license`](/api-reference/management-ui/set-the-license-via-management-ui) API methods to install and view licenses in the Management UI. These methods let you manage licenses in installations with single sign-on (SSO) without requiring individual user SSO credentials
* Added support for the Response API for [OpenAI](/ai/agents/develop/gen-ai-and-llms/providers/openai), [Microsoft Azure OpenAI](/ai/agents/develop/gen-ai-and-llms/providers/microsoft-azure-openai), and [OpenAI-compatible](/ai/agents/develop/gen-ai-and-llms/providers/openai-compatible) models
* Added [audit events](/ai/administer/access/admin-center/audit-events) for user, tenant, and role sync between NiCE CXone and Cognigy
* Renamed Username to Access Key ID and Password to Access Key Secret in the NiCE CXone Authentication Outbound settings in the [NiCE CXone (Non-Integrated)](/ai/escalate/handover-reference/nice-cxone/nice-cxone-non-integrated) handover provider

#### Bug Fixes

* Fixed the issue where accessing the Simulator from the left-side menu and creating scenarios didn't work properly for installations with their own Traefik implementation
* Fixed the issue where logging out without selecting a Project threw the `Unable to load Agents – No RefreshToken provided` error
* Fixed the issue where Knowledge Stores didn't work with custom embedding models from OpenAI and Azure OpenAI
* Fixed the issue where users could save the storm handover provider without selecting a connection
* Fixed the issue where the `gpt-5.2` and `gpt-5.4` models weren't working with OpenAI-compatible providers
* Fixed the issue where the Question Node crashed when an unknown country prefix was detected with Pattern: IBAN as the question type
* Fixed the issue where the WhatsApp Endpoint crashed with `TypeError: formattedMessages is not iterable`. This issue occurred when a Live Agent handover provider sent an event with empty text and no message content to Cognigy, causing subsequent messages in the output buffer to be dropped

### Cognigy Voice Gateway

**Released** May 12, 2026

#### Bug Fixes

* Fixed the issue where updating Set Session Config Node settings removed the speech-to-text (STT) and text-to-speech (TTS) settings during Flow execution and changed the STT language to English
* Fixed the issue where STT failed mid-call due to the incorrect resolution of the credential label for speech providers after switching STT vendors during Flow execution
* Fixed the issue where inbound calls were rejected with `603 Decline` when Cognigy sent a redirect within the first few milliseconds of call setup, before the media endpoint was ready
* Fixed the issue where handover statistics under-counted successful call transfers when the carrier sent SIP `BYE` before the REFER `NOTIFY` confirmation arrived
* Fixed the issue where Voice Gateway dropped the call without signaling the transfer outcome to Cognigy when a SIP carrier took longer than 15 seconds to return `NOTIFY` status for a REFER
* Fixed the issue in the Voice Gateway Self-Service Portal where the connection to the Microsoft Azure speech provider with a custom STT endpoint and disabled TTS was displayed as failed even though the STT connection test succeeded

### Cognigy Webchat

**Released** May 12, 2026

#### Improvements

* Added React 19 support for [Webchat v3 embedded via npm](/webchat/v3/embedding/npm)

### Cognigy Live Agent

**Released** May 12, 2026

#### Bug Fixes

* Fixed the issue where conversations were incorrectly auto-resolved after a short period of no activity, even if a human agent had just updated the conversation, for example, by assigning it, changing the team, adding labels, or updating the priority
* Fixed the issue where conversations were incorrectly auto-resolved when a new message was received due to a timing issue between checking for inactivity in the conversation and updating its status

### Cognigy Insights

**Released** May 12, 2026

#### Bug Fixes

* Fixed the issue in the Transcript Explorer where using the search field while filtering results by User Messages incorrectly returned sessions where the search result appeared only in AI Agent messages
* Fixed the issue where sessions were displayed as duplicates in the Transcript Explorer and were registered twice for session counts in OData

### Infrastructure

#### Cognigy + Insights + Agent Copilot

* The Redis, Redis Persistent, Redis Sentinel, and Redis Exporter images were updated as follows:
  * `redis:7.4.0-debian-12-r4` to `redis:7.4.8-photon-5-r11`
  * `redis-sentinel:7.4.0-debian-12-r4` to `redis-sentinel:7.4.8-photon-5-r11`
  * `redis-exporter:1.63.0-debian-12-r0` to `redis-exporter:1.83.0-photon-5-r0`

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2026.11
Source: https://docs.cognigy.com/release-notes/2026.11



<Warning>
  Removals:

  * The LLM Prompt Node (legacy) has been removed. Flows that contain this Node will continue to work. However, we strongly recommend migrating to the new [LLM Prompt](/ai/agents/develop/node-reference/service/llm-prompt) Node. This Node offers greater flexibility and advanced AI features, including tool use, image support, and conversation transcript access.
  * Support for Aleph Alpha's `luminous-embedding-128` has been removed. Switch to the newer models to ensure your AI Agents remain stable.
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2026.11.1

### Cognigy Voice Gateway

**Released** May 29, 2026

#### Bug Fixes

* Fixed the issue where calls were silenced and terminated after being transferred to enterprise carriers due to incompatible SRTP encryption suites

## 2026.11.0

### Cognigy

**Released** May 26, 2026

#### Features

##### Custom Models for Google GenAI

Added support for custom models to the [Google GenAI LLM](/ai/agents/develop/gen-ai-and-llms/providers/google-genai) provider, allowing users to configure and use fine-tuned models based on the Gemini model family instead of being limited to predefined options.

#### Improvements

* Added the Agent Change Notification setting to Salesforce MIAW in the [Handover to Human Agent Node](/ai/agents/develop/node-reference/service/handover-to-agent#salesforce-miaw-settings) and in the [Handover to Human Agent](/ai/agents/develop/manage-ai-agents#salesforce-chat-settings) tool. The setting shows a message when a conversation is transferred from one human agent to another
* Added support for changing the model type for standard models in the [LLM provider editor](/ai/agents/develop/gen-ai-and-llms/providers/all-providers), allowing you to switch to a different model without recreating the configuration
* Added error messages when updating the NiCE CXone provider fails, helping you quickly identify and fix configuration issues
* Removed the AgentAssistV1 action from the NiCE CXone handover provider's studio script to allow users without an Agent Assist license to create the provider

#### Bug Fixes

* Fixed the issue where the [Set Translation](/ai/agents/develop/node-reference/other-nodes/set-translation) Node didn't correctly keep its settings during session and handover, which incorrectly turned translation back on
* Fixed the issue in the UI where copying text removed parts of the content due to incorrect HTML tag handling
* Fixed the issue where AI Agent avatars weren't displayed in some customer browsers and instead showed a placeholder image
* Fixed the issue where calling `createApiKeyManagement` from `@cognigy/rest-api-client` failed to create an admin API key and incorrectly returned `401/404` errors
* Fixed the issue where the names of the AI Judge success criteria weren't validated for uniqueness, allowing duplicates that blocked saving or cause confusion in scenarios and simulations. Also fixed charts and tooltips to distinguish criteria with the same label
* Fixed the issue where scenario generation failed when using Anthropic LLMs as the design-time LLM because the request didn't include required user input messages
* Fixed the issue where the Create button remained inactive after selecting a valid goal for a Goal Completed success criterion on the New Scenario page

### Cognigy Voice Gateway

**Released** May 26, 2026

#### Features

##### Speechmatics in a Voice Preview

Speechmatics is now available in [Voice Preview](/ai/agents/test/voice-preview). Configure it in the Speech Services settings to test real-time speech recognition in your voice Flows.

#### Improvements

* Increased the default value from 1 to 5 in the User No Input Retries parameter of the [Set Session Config](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details#user-input-timeout) Node. This change applies to newly added Nodes only; existing Nodes aren't affected
* Removed the SIP Realm setting from the [Accounts](/voice-gateway/webapp/accounts#settings) in the Voice Gateway Self-Service Portal
* Added a confirmation dialog requiring user approval before deleting a speech credential

#### Bug Fixes

* Fixed the issue where failed call transfers (`410 GONE`) were incorrectly reported as [`TRANSFER_REFER_SUCCESS`](/voice-gateway/references/events/TRANSFER_REFER_SUCCESS) instead of [`TRANSFER_REFER_ERROR`](/voice-gateway/references/events/TRANSFER_REFER_ERROR), preventing fallback and error flows from running
* Fixed the issue where the Hang Up Node's custom `X-Reason SIP` header value was overwritten by the default call termination reason, causing incorrect termination data to be propagated downstream

### Cognigy Insights

**Released** May 26, 2026

#### Bug Fixes

* Fixed the issue where the Simulation Overview dashboard displayed multiple success criteria with the same display name as a single row. Each criterion is now shown separately based on its position in the results, ensuring correct counts and percentages

### Infrastructure

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2026.12
Source: https://docs.cognigy.com/release-notes/2026.12



<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

<Warning>
  * The following Gemini models have been deprecated:

    * `gemini-2.0-flash`. Instead, use `gemini-3.5-flash`
    * `gemini-2.0-flash-lite`. Instead, use `gemini-3.1-flash-lite`
    * `gemini-3.1-flash-lite-preview`. Instead, use `gemini-3.1-flash-lite`

  * The State tab in the Interaction Panel, the `POST /v2.0/sessions/{sessionId}/state/inject` and `POST /v2.0/sessions/{sessionId}/state/reset` API methods, and all `/v2.0/flow/{flowId}/states` API routes have been removed.
</Warning>

## 2026.12.2

### Cognigy

**Released** June 12, 2026

#### Bug Fixes

* Fixed the issue where selecting `Deepgram (provided by NiCE)` or `Microsoft (provided by NiCE)` as a speech provider in the Set Session Config Node changed the selection to `Deepgram` or `Microsoft` after saving the Node

## 2026.12.1

### Cognigy

**Released** June 11, 2026

#### Bug Fixes

* Fixed the issue where the [`GET /v2.1/projects/{projectId}/logs`](/api-reference-logs/logs-v21/get-log-entries) API method was unavailable

### Cognigy Voice Gateway

**Released** June 17, 2026

#### Bug Fixes

* Fixed the issue where the STT language set via the Set Session Config Node wasn't applied to the speech service, leading to inaccurate transcriptions

## 2026.12.0

### Cognigy

**Released** June 9, 2026

#### Features

##### New Gemini Models

Added support for [Gemini's](/ai/agents/develop/gen-ai-and-llms/providers/google-genai) `gemini-3.5-flash` and `gemini-3.1-flash-lite` models. These models allow for higher-performance AI capabilities across different use cases.

##### OAuth 2.0 Support for MCP Server Endpoint

Added OAuth 2.0 support for the [MCP server Endpoint](/ai/agents/deploy/endpoint-reference/mcp-server). With this feature, you can make sure only authenticated requests reach the MCP server.

#### Improvements

* Added the Custom Attributes toggle to the NiCE CXone settings section in the [Question Node](/ai/agents/develop/node-reference/basic/question), [Handover to Human Agent Node](/ai/agents/develop/node-reference/service/handover-to-agent), and [Handover to Human Agent tool](/ai/agents/develop/node-reference/ai/ai-agent#handover-to-ai-agent-tool)
* Removed the Set Up Cognigy Live Agent section from the [Project creation wizard](/ai/agents/develop/projects-and-flows/projects)
* Extended the [Node search](/ai/agents/develop/nodes/node-search) to include expressions inside If-condition fields, so variables and literals used in If conditions can now be discovered via the search box

#### Bug Fixes

* Fixed the issue where the Simulator section was incorrectly displayed in the Management UI
* Fixed the issue where running a simulation from a Snapshot in the Interaction Panel showed an infinite loading spinner and never rendered the simulation output
* Fixed the issue where the Model field in the LLM editor wasn't deactivated for embedding models
* Fixed the issue where Voice Preview and Click To Call failed with `No text-to-speech service credentials for speechmatics` when Speechmatics was selected as the speech provider
* Fixed the issue where the AI Agent incorrectly switched from the configured Locale to the default one
* Fixed the issue in the AI Agent editor where editing the Instructions and Context field of an AI Agent reactivated previously disabled tool Nodes in AI Agent Nodes configured with the edited AI Agent
* Fixed the issue where the response of `GET /management/v2.0/system/license` continued to expose `quotaMaxKnowledgeChunks` after the Knowledge AI feature was deactivated in the installed license
* Fixed the issue where the MCP Endpoint didn't expose tools due to internal paginated API responses
* Fixed the issue where creating a Knowledge Store from a web page failed when you entered text in the Description field
* Fixed the issue in the Question Node where changing the Question Type field caused parsing errors
* Fixed the issue where the inject and notify transformers didn't log errors when an error occurred
* Fixed the issue where a cached invalid NiCE CXone handover provider configuration prevented updates from being saved

### Cognigy Voice Gateway

**Released** June 9, 2026

#### Features

##### Speechmatics for Click To Call

[Speechmatics](/voice-gateway/webapp/speech-services#speechmatics) is now supported in [Click To Call](/click-to-call/overview). [Configure Speechmatics in your Click To Call application](/click-to-call/configure#voice-configuration) in the Voice Gateway Self-Service Platform to apply this speech provider to your voice widget.

#### Bug Fixes

* Fixed the issue in the Mute Speech Input Node where activating both Mute Speech Input and Mute DTMF Input toggles caused a `TypeError: Cannot read properties of undefined (reading 'length')` to be logged as `could not send final ping command` during the final ping in the Voice Gateway Self-Service Platform. The error didn't affect call execution or user experience
* Fixed the issue where Cognigy didn't receive AMD events when a custom HTTP URL was used as the `amd.actionHook` value
* Fixed the issue where executing `executeFlow()` after a SIP REFER transfer delayed call-completion cleanup by the 3-minute safety timeout instead of running immediately after the final ping
* Fixed the issue where Additional Session Parameters in the Advanced section of the Set Settings Config Node didn't override other settings in this Node

### Cognigy Live Agent

**Released** June 9, 2026

#### Bug Fixes

* Fixed the issue where conversations were auto-resolved while a human agent was actively viewing them. Viewing a conversation now resets the inactivity timer

### Cognigy Insights

**Released** June 9, 2026

#### Features

##### Conversation Analyzer

Introduced [Conversation Analyzer](/insights/conversation-analyzer/overview), a quality evaluation tool in Cognigy Insights that automatically scores your AI Agent's conversations based on predefined and custom criteria. With this feature, you can define an evaluation topic and a set of metrics, and the Conversation Analyzer processes your conversations and provides results across dashboards. The Conversation Analyzer is in beta, and we encourage you to try it and provide us with feedback.

#### Improvements

* Renamed the Transcript explorer to [Conversations explorer](/insights/explorers/conversations) to better represent the user experience in the scope of the Conversation Analyzer.

### Infrastructure

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2026.13
Source: https://docs.cognigy.com/release-notes/2026.13



<Danger>
  The following models have been removed:

  * Anthropic: `claude-opus-4-0 and claude-sonnet-4-0`
  * Google Gemini: `gemini-2.0-flash-lite`
  * Google GenAI: `gemini-3.1-flash-lite-preview`
  * Mistral AI: `pixtral-12b-2409`

  These models are no longer available. Migrate to the latest versions of these models to continue using them in your Cognigy Agents.
</Danger>

<Warning>
  The following models have been deprecated:

  * Azure OpenAI and OpenAI: `gpt-5-chat-latest`, `gpt-4.1-nano`
  * Google Gemini: `gemini-2.5-flash-lite`
  * Amazon Bedrock: `amazon.nova-premier-v1:0`
  * Mistral AI: `mistral-medium-latest`

  These models will be removed as soon as LLM providers retire them. Before the removal, migrate to the latest versions to continue using them in your Cognigy Agents.
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2026.13.2

### Cognigy Voice Gateway

**Released** July 29, 2026

#### Bug Fixes

* Adjusted the default configuration for Deepgram text-to-speech and speech-to-text

## 2026.13.1

### Cognigy Voice Gateway

**Released** July 27, 2026

#### Bug Fixes

* Fixed the issue where Deepgram speech-to-text and text-to-speech could use an incorrect configuration

### Cognigy

**Released** June 29, 2026

#### Improvements

* Added authenticated access for runtime file downloads, ensuring that files stored in third-party storage are accessible only through authenticated requests. For configuration steps, refer to [File Storage](/ai/agents/deploy/endpoints/file-storage)

#### Bug Fixes

* Fixed the issue in the Set Session Config Node where the Disable Punctuation and Dynamic Hints fields weren't visible when the STT vendor was set to Deepgram Flux

## 2026.13.0

### Cognigy

**Released** June 23, 2026

#### Features

##### Real-Time Queue Notifications in Salesforce MIAW

[Real-time queue notifications](/ai/escalate/handover-features/real-time-queue-notifications) are now supported by the Salesforce MIAW handover provider to allow end users to receive real-time updates on their queue position and estimated wait time during handovers to human agents.

##### New Embedding Models in Google GenAI LLM Provider

Added support for the [`gemini-embedding-2`](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) model and custom embedding models for the [Google GenAI LLM provider](/ai/agents/develop/gen-ai-and-llms/providers/google-genai), giving you more flexibility in how you configure semantic search and knowledge retrieval in your Cognigy Agents.

##### API Key Authentication in Webhook Endpoints

Added [API key authentication to Webhook Endpoints](/ai/agents/deploy/endpoint-reference/webhook#api-key-authentication), ensuring that only authorized sources can trigger your Cognigy Agents.

##### Attachment Support for NiCE CXone and Webchat v3

Added support for sending and receiving attachments when using the NiCE CXone handover provider with a [Webchat v3 Endpoint](/webchat/v3/configuration). End users can now upload files from Webchat v3 and share them with NiCE CXone agents, and can also view file attachments sent by NiCE CXone agents directly in Webchat v3. For more information, refer to [NiCE CXone (Integrated)](/ai/escalate/handover-reference/nice-cxone/nice-cxone-integrated#configure-attachments) and [NiCE CXone (Non-Integrated)](/ai/escalate/handover-reference/nice-cxone/nice-cxone-non-integrated#configure-attachments).

#### Improvements

* Improved the `httpRequest` helper in [Transformer functions](/ai/for-developers/transformers/overview#limitations) to respect the configured execution timeout. When a request exceeds the timeout, an error is thrown and logged. The default timeout is 5 seconds and is configurable via `TRANSFORMER_FUNCTION_EXECUTION_TIMEOUT_MS`
* Added support for sending LLM requests to the webhook when an error occurs for the AI Agent and LLM Prompt Nodes, such as invalid credentials, for easier LLM call debugging

#### Bug Fixes

* Fixed the issue where the Flow Transition debug message wasn't shown in the Interaction Panel when a Go To Node switched Flows
* Fixed the issue where AI Agent screens became cramped or unusable when the interaction panel or pinned left navigation reduced the available width
* Fixed the issue where leading or trailing whitespace in an AI Agent's name was saved and caused knowledge store name validation to fail on document upload
* Fixed the issue where the Interaction Panel agent selector didn't update after creating, renaming, or deleting an AI Agent without a page reload
* Fixed the issue where a Knowledge Store search with no results showed the onboarding wizard instead of an empty list
* Fixed the issue where slider values of 10,000 or above overflowed the value label across nodes that use a high-range slider, such as the Maximal Tokens field in the LLM Prompt Node. Values of 10,000 or above are now displayed in compact format, for example, `16000` → `16K`, `10100 → 10.1K`. Smaller values are displayed unchanged
* Fixed the issue where goals added via the Code Node using `analyticsData.completedGoals.push()` weren't reflected in the Tasks Completed count in the Conversations explorer
* Fixed the issue where the LLM Prompt Node returned a `400 Bad Request` error when Gemini complied with a system instruction to produce no output in non-streaming Store in Context mode
* Fixed the issue where clicking the token search field in the grounding knowledge configuration broke the UI
* Fixed the issue where streamed AI Agent replies in the Interaction Panel were split across multiple chat bubbles instead of rendering as a single message
* Fixed the issue where creating an Azure OpenAI connection with `baseCustomUrl` failed because inactive fields sent unexpected values to the backend
* Fixed the issue where editing a Flow didn't clear the Interaction Panel chat history
* Fixed the issue where hyperlinks in Interaction Panel chat messages weren't visually distinguished from surrounding text
* Fixed the issue where the `httpRequest` helper inside Transformers ignored the Transformer execution timeout, allowing slow HTTP calls to exceed the configured limit
* Fixed the issue where goals added via a Code Node using `analyticsData.completedGoals.push()` weren't reflected in the Tasks Completed count in the Conversations explorer
* Fixed the issue where Simulator-scheduled runs drifted by 1 hour after DST transitions. Schedules now fire at the configured local time year-round, regardless of daylight saving time changes

### Cognigy Voice Gateway

**Released** June 23, 2026

#### Improvements

* Sorted ElevenLabs legacy voices to the bottom of the voice selection list so that non-legacy voices appear first

#### Bug Fixes

* Fixed the issue where fallback webhook URIs longer than 64 characters caused a database error. The `fallback_webhook_target` and `fallback_webhook_referred_by` columns were widened from 64 to 1,024 characters, and oversized input now returns HTTP 400 with a message identifying the affected column
* Fixed the issue where a carrier `re-INVITE` was incorrectly sent for a `release-media` operation when the session was already in a no-media state, causing a `488 Not Acceptable` error
* Fixed the issue where ElevenLabs `voice_settings` were reset to an empty object when not specified at the task level in a `say` or `gather` verb following a Set Session Config Node. Options now fall back correctly from the task level to the session level and then to credentials
* Fixed the issue where the Azure speech-to-text session didn't restart after a recognizer language change mid-call, causing the original locale to persist for the rest of the call after an IVR language switch
* Fixed the issue where the `NOTIFY` listener in the `SIP REFER` task wasn't removed on all exit paths, which could cause listener leaks when a `REFER` completed, failed, or returned a non-202 response
* Fixed the issue where `asrTimeout` was ignored in continuous ASR mode with Deepgram because an `UtteranceEnd` event arrived before the final transcript was delivered
* Fixed invalid regex patterns in LCR (Least Cost Routing) that could cause route lookup to terminate prematurely, leading to incorrect outbound call routing

### Cognigy Webchat

**Released** June 23, 2026

#### Improvements

* Updated the Webchat v3 widget to [v3.44.0](https://github.com/Cognigy/Webchat/releases/tag/v3.44.0):
  * Added the `webchat/close-button` analytics event, which fires only when the user closes the chat using the header close button
  * Added the `disableMobileScrollLock` option to disable host-page scroll locking on mobile

### Cognigy Insights

**Released** June 23, 2026

#### Bug Fixes

* Fixed the issue where the Conversations explorer (expert mode) showed an incomplete and incorrectly ordered list of flow transitions for a session, causing some flows to appear missing
* Fixed the issue where NLU Performance reports with an hourly custom time range bypassed the cache and used an incorrect time window, resulting in inflated or inconsistent counts
* Fixed the issue where the NLU Performance Top Intents report included conversations at the exact upper boundary of a custom hourly time range, causing counts to be higher than expected
* Fixed the issue where the Conversation Analyzer scheduler Next run hint didn't refresh when switching between Daily and Weekly without saving

### Infrastructure

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2026.14
Source: https://docs.cognigy.com/release-notes/2026.14



<Warning>
  For all Cognigy installations, we are updating Traefik from version 3.4.0 to 3.6.0 to enhance functionality and address known security vulnerabilities in the dependencies of Cognigy, Voice Gateway, and Insights. For on-premises and dedicated SaaS installations, read the [Infrastructure](#infrastructure) section.
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2026.14.3

### Cognigy Voice Gateway

**Released** July 29, 2026

#### Bug Fixes

* Adjusted the default configuration for Deepgram text-to-speech and speech-to-text

## 2026.14.2

Released **July 23, 2026**

### Cognigy

#### Bug Fixes

* Fixed the issue where Extensions reported successful executions as failures because of unexpected API calls in installations where Cognigy Serverless was activated

### Cognigy Voice Gateway

**Released** July 27, 2026

#### Bug Fixes

* Fixed the issue where Deepgram speech-to-text and text-to-speech could use an incorrect configuration

## 2026.14.1

### Cognigy

**Released** July 17, 2026

#### Bug Fixes

* Fixed the issue where Extensions could unexpectedly fail on clusters with Cognigy Serverless enabled. Extension executions now complete successfully instead of showing an error

### Cognigy Voice Gateway

**Released** July 15, 2026

#### Bug Fixes

* Fixed the failover issue with [TTS streaming](/voice-gateway/streaming)

## 2026.14.0

### Cognigy

**Released** July 7, 2026

#### Features

##### Mask and Keep Analytics

The [Delete Profile](/ai/agents/develop/node-reference/analytics/delete-profile) and [Deactivate Profile](/ai/agents/develop/node-reference/analytics/deactivate-profile) Flow Nodes now support a Mask and keep analytics data option. Instead of permanently deleting a contact's analytics data, you can keep it with personal details replaced by a hash. Masked sessions remain counted in Analytics dashboards but are hidden from the Conversations and Message explorers.

##### API-Key Authentication in REST Endpoints

Added API Key authentication to [REST Endpoints](/ai/agents/deploy/endpoint-reference/rest). Inbound requests can now be secured with Endpoint-scoped API keys sent in the `x-rest-endpoint-key` header, generated and revoked from the Endpoint editor.

#### Improvements

* Added the capability to configure the [scenario limit per Project](/ai/agents/test/simulator/overview#limitations)
* Added support for importing [NICE CXone handover providers](/ai/escalate/handover-reference/nice-cxone/nice-cxone-integrated) via Packages and Snapshots
* Added a status indicator to the NiCE CXone (Integrated) handover provider that displays the connection progress between Cognigy and NiCE CXone

#### Bug Fixes

* Fixed the issue where the Code Node editor displayed errors for valid JavaScript, for example, `String.replaceAll`
* Fixed the issue in the API documentation where the paths for `GET /v2.0/handover/provider/{providerId}` and `POST /v2.0/handover/provider` were incorrect
* Fixed the issue where `GET /v2.1/projects/{projectId}/logs/{logEntryId}` returned `404` for IDs returned by `GET /v2.1/projects/{projectId}/logs` when the v2.0 log model was active
* Fixed the issue where the Flow Transition debug message wasn't shown in the Interaction Panel when a Go To Node switched Flows
* Fixed the issue where imported Endpoints through Packages kept the Locale reference from the original Project, causing Insights to report incorrect data about the percentage of users by locale
* Fixed the issue where user inputs and AI Agent outputs were excluded from the Conversations explorer due to a failed write batch caused by analytics records containing null bytes
* Fixed the issue where the Search Extract Output Node returned static instruction examples alongside search results

### Cognigy Voice Gateway

**Released** July 7, 2026

#### Features

##### New TTS Streaming

Introduced the [TTS streaming feature](/voice-gateway/streaming). Use TTS streaming to let your AI Agent deliver synthesized speech as it's generated, rather than waiting for the entire response to finish processing. This feature enables faster, more natural-sounding conversations.

##### Register Status Section

Added the [Register Status](/voice-gateway/webapp/carriers#register-status) section to the Edit Carrier page. When Require SIP Register is enabled, the section shows the latest `REGISTER` attempt details and a timeline of recent registration attempts to help identify registration failures without checking logs.

#### Improvements

* Removed the Use hosted Deepgram service option, making the Region setting always visible on the [Speech Service page](/voice-gateway/webapp/speech-services) in the Voice Gateway Self-Service Portal

#### Bug Fixes

* Fixed the issue where a possible race condition could lead to delayed cleanup of a speech-to-text connection during call transfers
* Fixed the issue where PCAP downloads broke when SIP Call-IDs contained forward slashes (`/`)
* Fixed the issue where trailing commas in contact data caused malformed STT hints that Deepgram Flux rejected with `400`, resulting in silent call failures
* Fixed the issue where atmosphere sounds were reproduced with interference on calls using the Opus audio codec at 48 kHz stereo

### Cognigy Live Agent

**Released** July 7, 2026

#### Bug Fixes

* Fixed the issue where abandoned sessions weren't displayed on the Live Agent dashboard in Insights

### Cognigy Insights

**Released** July 7, 2026

#### Bug Fixes

* Fixed the issue where conversations containing extremely large user input messages triggered a crash loop
* Fixed the issue where the Conversation Analyzer failed to update analysis status after completing a run
* Fixed the issue where the Conversation Analyzer incorrectly showed `Analysis complete` as a success notification when all analyses actually failed, for example, due to an invalid LLM connection

### Infrastructure

#### Cognigy + Insights + AI Copilot + Voice Gateway

Traefik, the Kubernetes ingress controller, has been updated to version 3.6.0 to enhance functionality and address known security vulnerabilities in the dependencies of Cognigy, Voice Gateway, and Insights. The bundled Helm chart is updated accordingly.

##### Traefik Update

Make updates to Cognigy and Voice Gateway separately because they use different Helm charts and `values.yaml` files.

Before installing the Helm chart, follow these steps:

1. In the new Traefik Helm chart, the entry point redirection and TLS settings have moved under a new `http` key. The `traefik.ports.web.redirections` setting is now `traefik.ports.web.http.redirections`. If you've modified the Traefik port redirection in your local `values.yaml` file, update it accordingly.

   Before:

   ```yaml theme={null}
   ports:
     web:
       redirections:
         entryPoint:
           to: websecure
           scheme: https
   ```

   Now:

   ```yaml theme={null}
   ports:
     web:
       http:
         redirections:
           entryPoint:
             to: websecure
             scheme: https
   ```

2. The `globalArguments` key has been deprecated. If you need to pass any arguments to Traefik, use the `additionalArguments` key instead.

3. Deploy the latest CRDs of Traefik by executing the following command:

   ```bash theme={null}
   kubectl apply --server-side --force-conflicts -k https://github.com/traefik/traefik-helm-chart/traefik/crds/?ref=v39.0.9
   ```

For more detailed information, refer to the official Traefik [release notes](https://github.com/traefik/traefik-helm-chart/releases).

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2026.15
Source: https://docs.cognigy.com/release-notes/2026.15



<Warning>
  Deprecations and removals in Cognigy:

  * The `gemini-2.5-pro` model has been deprecated.
  * The `pixtral-large-latest` model has been removed because Mistral AI no longer supports it.
  * The `gpt-4o` and `gpt-4o-mini` models have been deprecated for Microsoft Azure OpenAI. Version 2024-05-13 of `gpt-4o` is retired on October 1, 2026. Versions 2024-08-06 and 2024-11-20 of `gpt-4o`, and `gpt-4o-mini`, are retired on April 14, 2027. Migrate to `gpt-5.1` and `gpt-5-mini` respectively before the retirement date.
  * The `gpt-4.1`, `gpt-4.1-mini`, and `gpt-4.1-nano` models have been deprecated for Microsoft Azure OpenAI. `gpt-4.1` and `gpt-4.1-mini` are retired on April 14, 2027, and `gpt-4.1-nano` on October 14, 2026. Migrate to a more recent model before the retirement date.

  If you use these models in your Flows, migrate to available models from the [Model Support by Feature](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) to ensure your AI Agents remain stable.
</Warning>

<Warning>
  ElevenLabs deprecated and removed the following models used in Cognigy Voice Gateway speech services:

  * The Turbo v2 and Turbo v2.5 models have been deprecated. ElevenLabs recommends migrating to Flash v2 and Flash v2.5.
  * The Monolingual v1 and Multilingual v1 models have been removed. Migrate to Multilingual v2.
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2026.15.4

### Cognigy Voice Gateway

**Released** July 29, 2026

#### Bug Fixes

* Adjusted the default configuration for Deepgram text-to-speech and speech-to-text

## 2026.15.3

### Cognigy Voice Gateway

**Released** July 27, 2026

#### Bug Fixes

* Fixed the issue with calls using dial and refer transfers

## 2026.15.2

### Cognigy

**Released** July 24, 2026

#### Bug Fixes

* Fixed the issue where the logic to display deprecation icons in model lists didn't distinguish deprecation across providers. Now, for models shared across providers, such as `gpt-4o` and `gpt-4o-mini` from OpenAI and Microsoft Azure OpenAI, the deprecation icon only appears for the provider that deprecated the model

### Cognigy Voice Gateway

**Released** July 24, 2026

#### Bug Fixes

* Fixed the issue where Deepgram speech-to-text and text-to-speech could use an incorrect configuration

## 2026.15.1

### Cognigy

**Released** July 23, 2026

#### Improvements

* Improved the tooltip text for the Speech provider failover enabled parameter in the Voice Gateway Endpoint to clarify the behavior of the configured failover

#### Bug Fixes

* Fixed the issue where Extensions reported successful executions as failures because of unexpected API calls in installations where Cognigy Serverless was activated

### Cognigy Voice Gateway

**Released** July 23, 2026

#### Bug Fixes

* Fixed the issue where the configured fallback speech provider for STT wasn't used when the primary speech provider failed, causing calls to go silent and skip the configured call failover. Now, the fallback speech provider is correctly triggered for both STT and TTS when the primary speech provider fails, and the call failover configuration is executed in the following cases:
  * If only a primary speech provider is configured, the call failover occurs after the primary speech provider fails.
  * If both primary and fallback speech providers are configured, the call failover occurs only after both speech providers fail.

## 2026.15.0

### Cognigy

**Released** July 21, 2026

#### Features

##### New Azure OpenAI and OpenAI GPT-5.6 Models

* Introduced support for the [GPT-5.6 model family](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) (`gpt-5.6-sol`, `gpt-5.6-terra`, and `gpt-5.6-luna`), enabling users to leverage the latest AI capabilities. These models are available with the Responses API request type

#### Improvements

* Improved the localization logic in the [Simulator](/ai/agents/test/simulator/overview). Now, both AI Agent and simulated user interact in the selected locale language instead of defaulting to English
* Updated the Conversation Counter in [Admin Center](/ai/administer/access/admin-center/dashboard) and [Management UI](/ai/administer/access/management-ui) to use the v3 pre-aggregated Conversation Counter API. Conversation counts are now calculated overnight and available up to the previous day, with data available from 2025 onward
* Improved [attachment handling for Genesys Open Messaging integrations](/ai/escalate/handover-reference/genesys-cloud-open-messaging#additional-configuration) by sending user chat attachments with the correct media types in Webchat v3 and WhatsApp
* Added Script ID visibility for the [NiCE CXone (Integrated)](/ai/escalate/handover-reference/nice-cxone/nice-cxone-integrated) handover provider. Users can now view and copy the Script ID to distinguish between different configured scripts
* Reordered the settings on the [Project Settings](/ai/administer/access/project-settings) page to reflect priority order

#### Bug Fixes

* Fixed the issue where Playbook assertions executed from the Interaction Panel always returned failed results
* Fixed AWS Bedrock compatibility issues with Claude Sonnet 4.5 and Claude Sonnet 4.6 models caused by conflicting `temperature` and `top_p` parameters in LLM requests from AI Agent Nodes, LLM connection tests, and Simulator
* Fixed the issue that allowed the deprecated handover provider to be selected and saved on newly created Endpoints
* Fixed issues in Knowledge AI lists:
  * Added details for failed Knowledge Sources
  * Restored the Connector Sync button during sync after switching tabs
  * Removed unnecessary hover tooltips for non-truncated titles
* Fixed the issue where analytics events were miscounted in OData exports and reports due to shared analytics payloads between related conversation-turn publishes, causing inconsistent row counts
* Fixed the issue where exported Packages with Endpoints using cloud file storage didn't include the storage provider connection as a dependency or relink it after import
* Fixed the issue where editing the private key of a Google LLM connection broke the connection
* Fixed the issue where oversized log entries were dropped due to MongoDB's 16 MB document limit. Entries are now trimmed before storage to ensure logs are preserved
* Fixed the issue where Node execution logs were missing the session ID, making it difficult to correlate logs by conversation in external tools such as Splunk
* Fixed the issue where selecting messages in the Interaction Panel didn't copy their text into the input field. Selecting another message now replaces the previously copied content
* Fixed the issue where image and image-only PDF Knowledge Sources were stored as unreadable raw bytes due to Azure Document Intelligence OCR failures
* Fixed the issue where WhatsApp attachment captions weren't shown in the chat with human agents in NiCE CXone

### Cognigy Voice Gateway

**Released** July 21, 2026

#### Features

##### Genesys Audio Connector (beta)

Introduced the [Genesys Audio Connector](/voice-gateway/genesys-audio-connector) integration that connects Genesys Cloud to Voice Gateway using the AudioHook protocol over WebSocket, letting Genesys Cloud customers keep using Cognigy's conversational AI without relying on SIP calling.

#### Improvements

* Removed the feature flag for [Call Failover](/ai/agents/deploy/endpoint-reference/voice-gateway#call-failover) in the Set Session Config Node. The feature is now generally available.
* Improved handling of unwanted SIP traffic
* Improved endpoint validation for the call-recording upload path

#### Bug Fixes

* Fixed the issue where calls terminated immediately after a SIP REFER transfer caused unnecessary AI response retries and error messages
* Fixed the issue where ElevenLabs voice speed settings were ignored. Now, Voice Gateway correctly applies the configured speed value for ElevenLabs text-to-speech
* Fixed the issue where the `CALL_COMPLETED` event wasn't triggered after Dial-based call transfers, preventing configured event handling Flows from executing
* Fixed the issue where traces for programmatic outbound calls were incomplete when a notifyUrl was configured. Full call traces are now displayed correctly
* Fixed the issue where calls with long SIP Call-ID values weren't properly terminated after a Hang Up action, leaving calls active after the bot ended the conversation

### Click To Call

**Released** July 21, 2026

#### Bug Fixes

* Fixed the issue where an unused, invalid speech credential blocked Click To Call provisioning

### Webchat

**Released** July 21, 2026

#### Improvements

* Removed the option to create new Webchat v2 Endpoints from the Endpoints page and the Cognigy API. Existing Webchat v2 Endpoints are unaffected and continue to work

#### Bug Fixes

* Fixed the issue where Webchat v3 embedding snippets used the demo URL instead of the correct Endpoint URL

### Agent Copilot

**Released** July 21, 2026

#### Features

##### Enhanced JWT Authentication Options for Agent Copilot

Introduced expanded [JWT authentication options](/agent-copilot/configure/copilot-section) for Agent Copilot. Now, you can configure symmetric or asymmetric JWT verification in the Endpoint settings, including authentication through shared keys or public keys and JWKS endpoints for improved security and key rotation support.

### Cognigy Live Agent

**Released** July 21, 2026

#### Bug Fixes

* Fixed the issue where auto-assigned conversations showed an empty `assigned_at` field in the Conversation collection of the Live Agent OData endpoint

### Cognigy Insights

**Released** July 21, 2026

#### Bug Fixes

* Fixed the issue where opening a listed session from a Contact Profile incorrectly showed `This Session Transcript was deleted` instead of displaying the session
* Fixed the issue where analytics events with the same session, input, and timestamp but different `analyticsEventId` values were incorrectly deduplicated, causing undercounted rows in OData analytics queries
* Fixed the issue where some Flow transitions were missing from the transcript Flow switch history
* Fixed the issue where the Conversations explorer didn't show Flow switches for Flows without output, for example, Go To redirects. Also fixed incorrect Flow reporting and prevented sessions without user input from appearing in conversation reports
* Fixed the issue where the Conversations explorer showed the wrong Flow name for conversations that moved across multiple Flows
* Fixed the issue where Conversation Analyzer failed to analyze conversations with the following models:
  * OpenAI — `gpt-5.5`, `gpt-5.6-luna`, `gpt-5.6-sol`, `gpt-5.6-terra`
  * Azure OpenAI — `gpt-5.5`, `gpt-5.6-luna`, `gpt-5.6-sol`, `gpt-5.6-terra`
  * Anthropic — `claude-sonnet-4-5`, `claude-sonnet-4-6`
  * Google Gemini — `gemini-2.5-flash`, `gemini-2.5-pro`
  * Mistral AI — `mistral-large-latest`
  * AWS Bedrock — `amazon.nova-premier-v1:0`

### Infrastructure

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2026.16
Source: https://docs.cognigy.com/release-notes/2026.16



<Danger>
  Removals:

  * OpenAI's and Microsoft Azure OpenAI's `gpt-5-chat-latest` models have been removed. To avoid any issues with your AI Agent, migrate to a more modern model, for example, `gpt-5.6-sol`.
  * The **Handover Settings** section has been removed from the Endpoint settings. Consequently, the **Legacy** option was removed from the **Handover Provider** list in the Handover to Human Agent, Question, and Check Agent Availability Nodes. If you haven't migrated to the new interface for handover providers, refer to the [migration guide](/ai/escalate/migration).
</Danger>

<Warning>
  Deprecations:

  * [Cognigy Serverless](/ai/for-developers/extensions#cognigy-serverless) is required starting with Cognigy 2026.21.0. The legacy Extension runtime for trusted and untrusted Extensions has been deprecated. The removal is planned for Cognigy 2026.21.0.
    * For SaaS installations, Cognigy performs the migration and no action is required. If Cognigy hasn't contacted you, your Extensions are already running on Cognigy Serverless.
    * For on-premises installations, [install Cognigy Serverless](/ai/administer/installation/serverless-installation) before upgrading to Cognigy 2026.21.0. Otherwise, all Extensions will stop working after the upgrade.
  * As of Cognigy 2026.4.0, the Step and Message explorers in Insights are deprecated. These explorers will remain functional until their full removal in Cognigy 2026.18.0.
  * The [`GET /v2.0/profiles`](/api-reference/contact-profiles/get-contact-profiles) API method has been deprecated. The removal is planned for Cognigy 2026.24.0. Use [`GET /v2.1/profiles`](/api-reference/contact-profiles/get-contact-profiles-performance) instead for improved performance when retrieving Contact Profiles.
</Warning>

## 2026.16.1

### Cognigy Voice Gateway

**Released** August 18, 2026

#### Bug Fixes

* Fixed the issue where the audio connection wasn't properly updated during certain call transfers, causing audio issues after the transfer
* Fixed the issue where DTMF-only barge-in settings weren’t restored after the system finished waiting for caller input. As a result, calls incorrectly accepted speech as an interruption when only key presses were allowed
* Fixed the issue where speech transcription unexpectedly reconnected after a call transfer. This issue also occurred after a speech provider failover
* Fixed the issue where the system reported a media path change as successful while the call continued using the previous path
* Fixed the issue where audio failed to decrypt after one call side disconnected while the other remained active

## 2026.16.0

### Cognigy

**Released** August 4, 2026

#### Features

##### Real-Time Queue Notifications for Genesys Cloud Open Messaging

Introduced real-time queue notifications for the Genesys Cloud Open Messaging handover provider. End users can now receive estimated wait-time updates while waiting for a human agent. This feature improves transparency during handovers. For more information, refer to [Real-Time Queue Notifications](/ai/escalate/handover-features/real-time-queue-notifications#configure-settings-in-genesys-cloud-open-messaging).

#### Improvements

* Standardized the Endpoints-specific API authentication key header for the [REST](/ai/agents/deploy/endpoint-reference/rest) and [Webhook](/ai/agents/deploy/endpoint-reference/webhook) Endpoints. Now, use the `x-cognigy-endpoint-key` request header for both Endpoints. The legacy `x-rest-endpoint-key` and `x-webhook-key` headers are still accepted for backward compatibility, but `x-cognigy-endpoint-key` takes precedence if both are provided
* Added the `isMasked` field to the [GoalEvents](/ai/agents/analyze/odata#goalevents) collection for the Cognigy OData endpoint 2.4
* Added the [`POST /v2.1/projects/{projectId}/logs/package`](/api-reference-logs/logs-v21/package-log-entries-into-a-downloadable-file) and [`GET /v2.1/projects/{projectId}/logs/package/{logsFileId}/download`](/api-reference-logs/logs-v21/download-a-packaged-logs-file) API methods to package log entries into a JSON file you can download. These API methods allow you to avoid paginating through logs when using the [`GET /v2.1/projects/{projectId}/logs`](/api-reference-logs/logs-v21/get-log-entries) API method. The methods are available to some customers via the `FEATURE_PACKAGE_LOGS` feature flag. Contact [Cognigy technical support](/help/get-help) to activate them for your installation
* Removed the beta label from [NiCE CXone handover](/ai/escalate/handover-reference/nice-cxone/overview) providers. They are now displayed as generally available
* Increased the default maximum upload size for UI uploads of [Snapshots](/ai/agents/deploy/snapshots#limitations) and [Packages](/ai/platform-features/packages#limitations) to 8 GB. Direct Cognigy API upload limits remain unchanged
* Increased the default size limit for the [Context](/ai/agents/develop/ai-agent-memory/context#limitations) object from 65 kB to 256 kB

#### Bug Fixes

* Fixed the issue where newly created Locales and Snapshots were displayed in the Locale and Snapshot lists in the Interaction Panel only after the page was reloaded
* Fixed the issue where Cognigy Serverless unexpectedly terminated Extension executions
* Fixed the issue where opening Flows, Endpoints, Locales, and Snapshots returned a `500` error
* Fixed the issue where the status message for the NiCE CXone handover provider showed the incorrect text
* Fixed the issue where the names of API keys created without a name or named `ApiKey` weren't displayed in the API Keys list on the My Profile page. Now, you must enter a name when generating an API key
* Fixed the issue where updating or deleting an AI Agent failed with a `500 Internal Server Error`
* Fixed the issue where fullscreen mode in the mock code editor didn't close when the user pressed `Escape` after typing in the editor
* Fixed the issue where file-size limits weren't enforced for Snapshots and Packages uploaded through the UI
* Fixed the issue where incorrect text was displayed in the Copilot: Set Grid Node
* Fixed the issue where assertions always failed in a Playbook run from the Interaction Panel Test tab
* Fixed the issue where messages from Live Agent to the end user weren't displayed when the Flow included a data-only or an empty Say Node

### Cognigy Voice Gateway

**Released** August 4, 2026

#### Improvements

* Added the Region option to the [Voice Preview settings](/ai/agents/test/voice-preview#set-up-a-speech-provider) for the Deepgram speech provider
* Added the option to opt in to the [Deepgram Model Improvement Program](/ai/agents/test/voice-preview#set-up-a-speech-provider) for Deepgram and Deepgram Flux under **Speech > Speech Service** in the Voice Gateway Self-Service Portal. Opting in allows Deepgram to use call audio and transcripts to improve its models and may reduce per-request costs. Customers can opt out to keep their call data private

#### Bug Fixes

* Fixed the issue where barge-in failed after a previous barge-in was incorrectly handled

### Cognigy Live Agent

**Released** August 4, 2026

#### Improvements

* Improved service stability and performance

### Cognigy Insights

**Released** August 4, 2026

#### Bug Fixes

* Fixed the issue where the `This Session Transcript was deleted` message was displayed instead of the conversation when a session was opened from a Contact Profile
* Fixed the issue where very large conversation messages weren't recorded in Insights
* Fixed the issue where Conversation Analyzer failed for OpenAI's and Microsoft Azure OpenAI's `gpt-5.6` models because an unsupported parameter was sent in the request to the LLM provider

### Infrastructure

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2026.17
Source: https://docs.cognigy.com/release-notes/2026.17



<Warning>
  Deprecations:

  * The [Amazon Alexa](/ai/agents/deploy/endpoint-reference/amazon-alexa) and [Dialogflow](/ai/agents/deploy/endpoint-reference/dialogflow) Endpoints have been deprecated. You can keep using Endpoints of these types, but you can't create new ones. The removal is planned for January 2028.
  * The [Alexa](/ai/platform-features/nlu/external/nlu-connector-reference/alexa) and [Dialogflow](/ai/platform-features/nlu/external/nlu-connector-reference/dialog-flow) NLU connectors have been deprecated. You can still create and use these NLU connectors until the removal date, but no further feature development is planned. Before the removal date, migrate to an alternative NLU connector, such as [Cognigy NLU](/ai/platform-features/nlu/external/nlu-connector-reference/cognigy-nlu). The removal is planned for January 2028.
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2026.17.1

### Cognigy Voice Gateway

**Released** September 2, 2026

#### Bug Fix

* Fixed the issue in the Helm chart that affected new Voice Gateway installations

### Cognigy Insights

**Released** August 19, 2026

#### Bug Fixes

* Fixed the issue where an Insights database migration job failed

## 2026.17.0

### Cognigy

**Released** August 18, 2026

#### Features

##### New Models by Google Gemini

Added support for the `gemini-3.6-flash` and `gemini-3.5-flash-lite` models provided by [Google Gemini](/ai/agents/develop/gen-ai-and-llms/providers/google-gemini). These models provide faster, lower-cost options for high-volume Generative AI tasks.

##### New Models by Anthropic

Added support for the `claude-sonnet-5`, `claude-opus-4-7`, and `claude-opus-4-8` models provided by [Anthropic](/ai/agents/develop/gen-ai-and-llms/providers/anthropic). These models give you access to Anthropic's latest model capabilities.

#### Improvements

* Added the [Use deployment name as model](/ai/agents/develop/gen-ai-and-llms/providers/microsoft-azure-openai) toggle to the Azure OpenAI LLM provider settings for the Responses API. You can now choose whether to use the Deployment Name or Model parameter to specify the model
* Added Custom Model Options and Custom Request Options to the [LLM Entity Extract](/ai/agents/develop/node-reference/other-nodes/llm-entity-extract) and [Question](/ai/agents/develop/node-reference/basic/question) Nodes. You can now pass additional model and request parameters directly to the LLM call
* Added support for multiple Microsoft SSO OAuth connections in [Azure Bot Services](/ai/agents/deploy/endpoint-reference/azure-bot-services) Endpoints. You can now manage tokens for each connection separately using the Overwrite Connection option in the [Get Token](/ai/agents/develop/node-reference/service/get-token), [Invalidate Token](/ai/agents/develop/node-reference/service/invalidate-token), and [Token Status](/ai/agents/develop/node-reference/service/token-status) Nodes
* Improved compatibility with [Anthropic models](/ai/agents/develop/gen-ai-and-llms/providers/anthropic) that don't support sampling parameters. Now, Cognigy removes unsupported `temperature`, `top_p`, and `top_k` parameters before sending the request
* Improved the reliability of the Project Overview dashboard. If the simulation search fails, the Recently Changed panel still displays other resources instead of showing an error
* Added `traceId` and `sessionId` to [transformer error logs](/ai/for-developers/transformers/overview#error-handling-in-transformers), making it easier to trace errors to a specific request and session

#### Bug Fixes

* Fixed the issue where the Logs API generated invalid pagination links with empty next or previous parameters. Now, you can use pagination links to navigate through log entries without receiving a `400 Bad Request`
* Fixed the issue where the platform stopped Knowledge Connector synchronizations but left them in the running state. Now, stopped synchronizations automatically switch to failed, and the connector status updates accordingly
* Fixed the issue where Package imports failed when a Connection contained a long field value, such as a MongoDB connection string. Connection values within the 10,000-character limit now import successfully
* Fixed the issue where an Intent in an unattached Flow triggered a Default Reply and stopped message processing
* Fixed the issue where saving an MCP Server Endpoint with OAuth 2.0 authentication failed when the optional Key Refresh Interval was left empty
* Fixed the issue where handover providers received duplicate AI Agent messages while waiting for a human agent to accept the handover
* Fixed the issue where the Studio script name appeared before a NiCE CXone handover provider was ready. Now, the Studio script name appears only when the provider status is `Ready`
* Fixed the issue where masked `GoalEvents` weren't correctly included or filtered in the OData responses
* Fixed the issue where clicking a conversation from a Contact Profile opened the transcript with an incorrect contact ID and where the Conversations section was inaccessible on small screens
* Fixed the issue where users couldn't open the Node editor for a Say Node configured with the Gallery output type

### Cognigy Voice Gateway

**Released** August 18, 2026

#### Improvements

* Added the `recordProcessor` attribute to the background recording span in the call detail record traces. This attribute identifies whether the audio manager or API server handled the recording

### Cognigy Agent Copilot

**Released** August 18, 2026

#### Bug Fixes

* Fixed the issue where Agent Copilot didn't get session data from a JWT when JWT authentication was deactivated. Agent Copilot now supports session data from JWTs in URLs with [Allow only JWT requests](/agent-copilot/configure/copilot-section) turned off
* Fixed the issue where the Transcript tile didn't show the conversation transcript in the Agent Copilot workspace

### Cognigy Live Agent

**Released** August 18, 2026

#### Improvements

* Renamed AI Copilot Workspace to Agent Copilot Workspace in the [Profile](/live-agent/profile-settings) and [Account](/live-agent/settings/account-settings) settings for consistency with the Cognigy interface

### Cognigy AI Ops Center

#### Bug Fixes

* Fixed the issue where the AI Ops Center accepted invalid notification [Webhook URLs](/ops-center/manage/alerts#webhook). Now, you can use only valid `http://` or `https://` Webhook URLs

### Cognigy Insights

**Released** August 18, 2026

#### Bug Fixes

* Fixed the issue where Conversation Analyzer returned a `502` error when analyzing all new conversations over large date ranges. Now, the analysis completes successfully for these date ranges

### Infrastructure

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2026.18 (latest)
Source: https://docs.cognigy.com/release-notes/2026.18



<Danger>
  Removals:

  * The Step and Message explorers in Insights have been removed. Use the [Conversations](/insights/explorers/conversations) explorer or [OData Endpoint's Steps collection](/ai/agents/analyze/odata#steps) instead.
  * The States tab in the **NLU** settings in the Flow editor and the Set State and Reset State Nodes have been removed. After the removal, existing Flows containing these Nodes will remain fully supported and continue to run as expected. As an alternative, you can use [Intent Conditions](/ai/platform-features/nlu/intents/conditions) for controlling Intent recognition.
</Danger>

<Warning>
  Deprecations:

  * The Blind Mode Node has been deprecated. The removal is planned for March 2027. Use the [Data Redaction](/ai/platform-features/data-redaction) feature instead.
  * The `gpt-5-nano` and `gpt-5-mini` models have been deprecated for Microsoft Azure OpenAI. Migrate to a more recent model, such as `gpt-5-4-nano` or `gpt-5-4-mini`.
  * Use of the **Model** field for deployment selection in [Microsoft Azure OpenAI](/ai/agents/develop/gen-ai-and-llms/providers/microsoft-azure-openai) Responses API requests has been deprecated. The removal is planned for Cognigy 2027.3.0. Use the **Deployment Name** field instead by activating the **Use deployment name as model** parameter.
  * The following parameters have been deprecated:

    * **Maximum Completion Tokens** and **Temperature** in the [AI Agent](/ai/agents/develop/node-reference/ai/ai-agent) Node
    * **Sampling Method**, **Temperature**, **Top Percentage**, **Maximal Tokens**, **Frequency Penalty**, **Presence Penalty**, **Use Stops**, **Stops**, and **Seed** in the [LLM Prompt](/ai/agents/develop/node-reference/service/llm-prompt) Node
    * **Temperature**, **Maximal Tokens**, **Presence Penalty**, **Frequency Penalty**, **Use Stops**, and **Stops** in the [Search Extract Output](/ai/agents/develop/node-reference/other-nodes/search-extract-output) Node
    * **Temperature** in the [LLM Entity Extract](/ai/agents/develop/node-reference/other-nodes/llm-entity-extract) Node
    * **Temperature** in the **LLM Entity Extraction Options** of the [Question](/ai/agents/develop/node-reference/basic/question) Node

    The removal of these parameters is planned for April 2027. For more information, refer to the [Deprecations and Removals](/release-notes/deprecations-and-removals) page.
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2026.18.2

### Cognigy Voice Gateway

**Released** September 8, 2026

#### Bug Fixes

* Fixed the issue where certain IP allowlist entries matched unintended IP addresses and where blocking malicious traffic caused a memory leak by updating `vg-drachtio` to `1.3.5`

## 2026.18.1

### Cognigy Voice Gateway

**Released** September 2, 2026

#### Bug Fixes

* Fixed the issue in the Helm chart that affected new Voice Gateway installations

## 2026.18.0

### Cognigy

**Released** September 1, 2026

#### Features

##### New Model by Google Gemini

Added support for the `gemini-3.7-flash` model provided by [Google Gemini](/ai/agents/develop/gen-ai-and-llms/providers/google-gemini). This model reliably follows instructions and handles multi-step planning and tool calls.

#### Improvements

* Moved the Logs API to its [own service](/api-reference-logs/). The endpoint routes remain the same as those used by the Cognigy API, so no migration is required
* Removed the OAuth2 authentication option as a supported API authentication method from the [Cognigy API](/api-reference/index)
* Added automatic login via SSO when opening a Cognigy URL in [CXone-integrated installations](/ai/administer/installation/access-to-cognigy-ai#log-in-to-cognigy-via-nice-cxone-mpower)

#### Bug Fixes

* Fixed the issue where `GET /v2.0/playbooks/{playbookId}/runs` returned a 500 error when listing the runs of a Playbook that already had at least one run
* Fixed the issue where task notifications showed the internal task identifier instead of the task name
* Fixed the issue where updating connections on the Handover Provider page didn't make the save button clickable
* Fixed the issue where the Custom URL tooltip for LLM and embedding connections displayed a deprecated Azure `/completions` path for the 2023 API version and didn't provide a Responses API example, causing connection errors for customers who followed the example verbatim
* Fixed the issue where editing Gallery cards in a Say or Question Node broke the Node editor in some browsers, for example, Microsoft Edge
* Fixed the issue where opening the Endpoint settings showed an error screen
* Fixed the issue where transient Meta Graph API failures dropped WhatsApp messages, read receipts, or media
* Fixed the issue where a Genesys Cloud Open Messaging conversation reopened and repeatedly reconnected the end user to a human agent after the handover had already ended
* Fixed the issue where empty values for the Update Interval parameter in the Handover to Human Agent Node resulted in failed handovers
* Fixed the issue where users were marked as disconnected in Live Agent while in the handover queue
* Fixed the issue where scheduled simulations targeting a Snapshot crashed in Projects where the Project settings were left unchanged
* Fixed the issue where data redaction didn't apply to structured log payloads, for example, Voice Gateway call transcripts. Now, structured log payloads are redacted, and a redaction failure discards the content
* Fixed the issue where deactivating all Data Redaction settings for a Project displayed a misleading warning and fetched the settings on every message instead of caching them

### Cognigy Voice Gateway

**Released** September 1, 2026

#### Bug Fixes

* Fixed the issue where activating only DTMF barge-in still let the caller's voice input interrupt the AI Agent's output after the first AI Agent output
* Fixed the issue where transferring a call over an encrypted (TLS/SRTP) trunk connected successfully but produced no audio in either direction
* Fixed the issue where a `Speech to text credentials have not been provisioned` alert was incorrectly displayed
* Fixed the issue where changing Deepgram STT settings mid-call didn't take effect until a new session started. Now, Deepgram STT settings changes take effect immediately
* Fixed the issue where AWS STT streaming connections remained open after calls ended, increasing CPU usage and causing calls to hang. Now, streams are always closed when a session ends
* Fixed the issue where a Transfer Node using the Dial transfer type failed to connect calls when the target was a SIP address instead of a phone number, on trunks that require encrypted audio (SRTP)

### Cognigy Webchat

**Released** September 1, 2026

#### Improvements

* Updated Webchat v3 to [v3.50.0](https://github.com/Cognigy/Webchat/releases/tag/v3.50.0)

### Cognigy Live Agent

**Released** September 1, 2026

#### Improvements

* Improved service stability and performance

### Cognigy Insights

**Released** September 1, 2026

#### Bug Fixes

* Fixed the issue where the Conversation Analyzer failed to analyze conversations when a reasoning model, for example, `gpt-5`, was configured
* Fixed the issue where triggering the Analyze all new conversations option failed with a `502` error for date ranges longer than 30 days

### Infrastructure

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2026.2
Source: https://docs.cognigy.com/release-notes/2026.2



<Warning>
  Upcoming removals:

  * The `POST /v2.0/optionsresolver` API method will be removed in release 2026.3 for SaaS installations. If you use this method, update your setup before the removal to avoid disruptions.
  * Anthropic's `claude-3-opus` and Amazon Bedrock's `amazon.titan-text-express-v1` models will be removed in release 2026.3. Switch to the newer models before the removal.
  * As of Cognigy 4.83.0, Webchat v2 is deprecated. The full removal of this product is scheduled for February 2026. Make sure you migrate to [Webchat v3](/webchat/migration) before the removal date.
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2026.2.3

### Cognigy

**Released** January 26, 2026

#### Bug Fixes

* Fixed the issue where users couldn't view the Simulation dashboard due to the `Unable to fetch Simulation Batches` error

## 2026.2.2

### Cognigy

**Released** January 23, 2026

#### Bug Fixes

* Fixed the issue where importing Knowledge Stores via Packages failed

## 2026.2.1

### Cognigy

**Released** January 22, 2026

#### Bug Fixes

* Fixed the issue where Set Session Config Nodes generated incorrect error logs

## 2026.2.0

### Cognigy, Cognigy Voice Gateway, and Cognigy Live Agent

**Released** January 20, 2026

#### Improvements

* Improved service stability and performance

### Infrastructure

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2026.3
Source: https://docs.cognigy.com/release-notes/2026.3



<Warning>
  [Cognigy Live Agent](/live-agent/index) is in maintenance mode. The product will continue to be maintained in the foreseeable future. Users can continue to fully rely on Live Agent being available. In future releases, only security updates and bug fixes will be provided; no new features will be released.
</Warning>

<Warning>
  Removals in Cognigy 2026.3.0:

  * The [`POST /v2.0/optionsresolver`](/api-reference/options-resolver/resolve-options-for-a-field) API for SaaS installations. If you use this method, update your setup to avoid disruptions.
  * Anthropic's `claude-3-opus` and Amazon Bedrock's `amazon.titan-text-express-v1` models. Switch to the newer models to ensure your AI Agents run smoothly.
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2026.3.2

### Cognigy

**Released** February 12, 2026

#### Features

##### Support for the Latest Models from OpenAI, Microsoft Azure OpenAI, and Anthropic

Introduced support for OpenAI's and Microsoft Azure OpenAI's `gpt-5-2` model, and for Anthropic's `claude-sonnet-4-5` and `claude-haiku-4-5` models.

### Cognigy Voice Gateway

**Released** February 11, 2026

#### Bug Fixes

* Fixed the issue where ElevenLabs's STT service didn't correctly route requests to region-specific API endpoints when Europe was configured as the service region

## 2026.3.1

### Cognigy

**Released** February 9, 2026

#### Bug Fixes

* Fixed the issue where the IntelliSense code completion options were limited only to `api.say` in the Code Node editor

### Cognigy Voice Gateway

**Released** February 4, 2026

#### Bug Fixes

* Fixed the issue where Deepgram's STT service failed by enforcing Transport Layer Security (TLS) for Deepgram API endpoints
* Fixed the issue where input was gathered after a `REFER` transfer occurred when event handling for `TRANSFER_REFER_ERROR` was activated

## 2026.3.0

### Cognigy

**Released** February 3, 2026

#### Features

##### New Embedding Model from Google

Introduced the [`gemini-embedding-001`](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) model from Google to support the Knowledge AI feature.

##### OAuth2.0 for AI Agent MCP Tools

Introduced OAuth 2.0 in the MCP tool for [LLM Prompt](/ai/agents/develop/node-reference/service/llm-prompt) and [AI Agent](/ai/agents/develop/node-reference/ai/ai-agent) Nodes. OAuth 2.0 provides more secure and robust authentication options.

##### Cognigy Enhancements

Introduced enhancements in Cognigy to improve AI Agent testing:

* [Simulator Dashboard](/ai/agents/test/simulator/dashboard) – get a clear overview of your simulations at a glance.
* [Simulator Scheduler](/ai/agents/test/simulator/overview#schedule-simulations) – plan and automate simulation runs for easier testing.
* [Mocking](/ai/agents/test/mocking) – simulate responses to speed up development and testing.

#### Improvements

* Added OAuth2 Client Credentials to the Authentication Type section in the [Send SMTP Email Node](/ai/agents/develop/node-reference/data/send-smtp-email), providing a more robust and simpler option for sending emails
* Removed beta from the [Train Flows](/ai/platform-features/nlu/intents/overview#all-flows) feature
* Added visual highlighting for the currently selected option in the simulation transcript. This applies to buttons and list items, making it easier to see which option is active during testing
* Added full support for [handover escalations via the Question Node](/ai/agents/develop/node-reference/basic/question). Handover providers from the Handover Provider settings can now be configured in the Escalation – Intents or Escalation – Wrong Answers sections
* Increased the Knowledge Source tags [limit from 10 to 50](/ai/agents/develop/knowledge-ai/knowledge-source/knowledge-source-tags#limitations) per Knowledge Source
* Improved Cognigy API by returning the version key in `GET v2.0/knowledgestores/{knowledgeStoreId}/descriptors`
* Released the [Data Redaction](/ai/platform-features/data-redaction) feature to all customers following a successful controlled rollout
* Added data redaction support to patterns that are [matched by Slots](/ai/platform-features/data-redaction#restrictions)
* Added support to manage user and global roles exclusively on the NICE CXone platform for integrated environments, and to [log in to Cognigy via the CXone interface](/ai/administer/access/management-ui#create-an-organization)
* Added support for the Live Follow feature when `userId` is updated using the Input Transformer
* Added full CRUD support for the Simulator on the Audit Events page, showing who performed create, view, update, and delete actions

#### Bug Fixes

* Fixed the issue where the Data Redaction Settings section was displayed on the Project settings page when data redaction was deactivated
* Fixed the issue where users with the `fullSupportUser` role couldn't access the Audit Events page
* Fixed the issue where an invalid log displayed the Set Session Config Node content
* Fixed the issue where a required Node property wasn't updated after a version upgrade, which caused the Deepgram model to break
* Fixed the issue where small value percentages in the Success Criteria chart weren't displayed properly
* Fixed the issue where changing AI Agent output timeouts by pressing the `Enter` key broke the UI (`Unable to render the main content`)
* Fixed the issue where Handover Provider settings were deleted when restoring a Snapshot or importing a Package
* Fixed the issue where the xApp link didn't work when real-time translation was enabled
* Fixed the issue where data redaction didn't work when a `[` or `,` appeared in a string sent in the `data` object of the message
* Fixed the issue where an error occurred when the conversation history using Amazon Bedrock models grew over 50 messages
* Fixed the issue where the Custom URL option for OpenAI embedding models wasn't supported by knowledge search
* Fixed the issue where creating a simulation triggered an incorrect error
* Fixed the issue where the Knowledge Store import through Packages failed and threw a validation error

### Cognigy Voice Gateway

**Released** February 3, 2026

#### Features

##### New STT and TTS Provider: Speechmatics

Introduced [Speechmatics](/voice-gateway/references/tts-and-stt-vendors) for both STT and text-to-speech (TTS), making voice interactions more flexible and accurate.

##### ElevenLabs as STT Provider

[ElevenLabs](/voice-gateway/references/tts-and-stt-vendors) is now available as an STT provider, giving you even more choice for your AI voice applications. Also, ElevenLabs is supported by the [Voice Preview](/ai/agents/test/voice-preview) feature in Cognigy.

#### Bug Fixes

* Improved stability and reduced call drops in the Silence Overlay and Atmosphere Sounds features

### Cognigy Webchat

**Released** February 3, 2026

#### Improvements

* Refreshed the GeoIP database (GeoLite2-City) to improve the accuracy of geolocation mapping in Webchat metadata

### Cognigy Live Agent

**Released** February 3, 2026

#### Improvements

* Fixed the issue where inbox-level Notification settings were ignored when account-wide settings were active
* Fixed the issue where the All Conversations tab wasn't visible for human agents
* Fixed the issue where auto-assignment of queued conversations weren't triggered after escalation to a supervisor

### Cognigy Insights

**Released** February 3, 2026

#### Improvements

* Added the [`deleteConversationsv2.1`](/api-reference/conversations-v21/deletes-conversations-records-for-a-sessionid) API method and updated the Cognigy OpenAPI documentation

#### Bug Fixes

* Fixed the issue where the Flow filter in the Transcript Explorer didn't retrieve correct results
* Fixed the issue where OData connections weren't handled correctly
* Fixed the issue where geolocation data was collected despite `maskAnalytics` being enabled. Now, IP addresses and geolocation are excluded
* Fixed the issue where timeline queries in Insights reports were using an incorrect time zone

### Infrastructure

#### Cognigy Voice Gateway

* Rename `dbCreate` to `apiMigrator`: `dbCreate.hookEnabled` -> `apiMigrator.hookEnabled`

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2026.4
Source: https://docs.cognigy.com/release-notes/2026.4



<Warning>
  As of Cognigy 2026.4.0, the Step and Message explorers in Insights are deprecated. These explorers will remain functional until their full removal in Cognigy 2026.18.0.
</Warning>

<Warning>
  As of Cognigy 2026.4.0, the OAuth2 authentication method for the Send SMTP Email Node is deprecated. The full removal of this authentication method is planned for Cognigy 2026.16.0. To avoid disruptions in AI Agents using OAuth2 in this Node, migrate to the [OAuth2 Client Credentials](/ai/agents/develop/node-reference/data/send-smtp-email#oauth2-client-credentials-recommended) authentication method before the removal date.
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2026.4.1

### Cognigy Voice Gateway

**Released** February 18, 2026

#### Bug Fixes

* Fixed the issue where the Recent Calls page in the Voice Gateway Self-Service Portal displayed an infinite loading spinner for users with the Account scope

## 2026.4.0

### Cognigy

**Released** February 17, 2026

#### Improvements

* Added the [Maximum Loops parameter](/ai/agents/develop/node-reference/ai/resolve-tool-action#advanced) to the Advanced section in the Resolve Tool Action Node. This parameter allows users to control how many times the AI Agent can loop back for multistep reasoning
* Added a warning about data security to the [Extensions](/ai/for-developers/extensions) page
* Added deprecation warnings for the [Handover Settings](/ai/escalate/migration) section in the Endpoint settings and set this section to read-only mode
* Added a deprecation tag to the Legacy option of the Handover Providers field in the [Handover to Agent Node](/ai/agents/develop/node-reference/service/handover-to-agent), [Question Node](/ai/agents/develop/node-reference/basic/question), and [Handover to Human Agent Tool](/ai/agents/develop/node-reference/ai/ai-agent#tool)
* Renamed the End Participant Change toggle to End Chat When Agent Leaves in the [Salesforce MIAW handover provider settings](/ai/escalate/handover-reference/salesforce-miaw#configuration-on-the-handover-provider-side)
* Added the Display Agent Details toggle to the [Handover to Human Agent](/ai/agents/develop/node-reference/service/handover-to-agent) and to the [Question](/ai/agents/develop/node-reference/basic/question) Nodes for Genesys handover providers. This toggle allows you to display the human agent's name and profile image in the chat

#### Bug Fixes

* Fixed the issue where the Embedding option was displayed in the Model Type field for LLM providers for which custom embedding models weren't supported
* Fixed the issue where handover providers weren't saved after being created or updated
* Fixed the issue where custom JSON objects sent through handover providers weren't expanded
* Fixed the issue where data assertions for Quick Replies failed during Playbook runs in the Interaction Panel
* Fixed the issue where Extension Nodes were displayed on the tab next to the Extension tab, and the Nodes in the last tab weren't displayed
* Fixed the issue where the Data Redaction settings section was displayed on the Settings page even when the feature was disabled
* Fixed the issue where the Train Flows tooltip was blank
* Fixed the issue where the Live Agent Inbox Id field was empty when the Live Agent Handover provider was selected in a Project with a preconfigured Live Agent account
* Fixed the issue where Nodes prompted users to save changes even if no changes were made
* Fixed a password reset issue that could expose users across multiple organizations
* Fixed the issue where goals weren't completely synced to Insights during Snapshot uploads
* Fixed the issue where the NLU Date Slot outputted an ISO date with incorrect formatting
* Fixed the issue where the API Access Token and API Key fields couldn't be unmasked in the 8x8 and Chatwoot handover provider settings
* Fixed the issue where the Question Node wasn't saved when Genesys Cloud Open Messaging was selected as the handover provider for escalations

### Cognigy Voice Gateway

**Released** February 17, 2026

#### Features

##### Deepgram Flux for STT Services

Introduced support for the Deepgram Flux speech provider for speech-to-text (STT) services. This provider is powered by the Deepgram Flux model, which is optimized for transcription accuracy, real-time streaming, and processing large volumes of audio.

##### NiCE CXone: Preconfigured SIP Trunk for Quick Setup

Added NiCE CXone as a [preconfigured SIP trunk provider](/voice-gateway/references/sip-trunk-providers) in the Voice Gateway Self-Service Portal. By selecting NiCE CXone in the carrier settings, the fields on the Edit carrier page are automatically filled in with the standard SIP trunk values for the NiCE CXone carrier. The feature is available only for installations integrated with the NiCE CXone platform.

#### Bug Fixes

* Fixed the issue where STT hints weren't correctly passed for Deepgram's Nova-3 models
* Fixed the issue where a range error occurred due to misaligned audio buffers
* Fixed the issue where a final ping was sent after a `sip:refer`

### Agent Copilot

**Released** February 17, 2026

#### Improvements

* Added the Accept Only JWT Requests toggle to the Endpoint settings. This toggle allows you to enforce JWT validation for requests to the Agent Copilot URL

### Cognigy Live Agent

**Released** February 17, 2026

#### Improvements

* Implemented conversation-handling logic so that manual assignee changes automatically trigger conversation reassignment
* Removed the Automatically set escalated conversations to Escalation priority option from the Inbox settings
* Updated the Live Agent library to align with current Cognigy components. It now supports proper rendering of AI Agent messages with complex output types, such as Text with Buttons, Galleries, and Text with Quick Replies, in conversations. This update ensures structural consistency across components and full compatibility with all Cognigy output types

#### Bug Fixes

* Fixed the issue where the unread message counts didn't include all unread messages in a conversation that the human agent didn't open
* Fixed the issue where escalated conversations appeared on the Mine tab only after the page was refreshed or the tabs were switched

### Infrastructure

#### Cognigy + Insights + Agent Copilot

* During the Helm chart upgrade, the billing-related CronJob runs automatically in on-premises environments and performs the necessary background operations to maintain data integrity and system consistency. No manual action is required.

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2026.5
Source: https://docs.cognigy.com/release-notes/2026.5



<Warning>
  Anthropic's `claude-3-7-sonnet-latest` model has been removed. Switch to the newer models to ensure your AI Agents run smoothly.
</Warning>

<Warning>
  **Deprecations:**

  * The [`DELETE /v2.0/conversations/{sessionId}`](/api-reference/conversations/deletes-a-conversation) API method has been deprecated. The removal of the API method is planned for Cognigy 2026.20.0. Use the `DELETE /v2.1/conversations/{sessionId}` API method instead.
  * The `createKnowledgeSource` function has been deprecated. For Knowledge Connectors, use [`upsertKnowledgeSource`](/ai/for-developers/knowledge-connector/upsertKnowledgeSource) to create or update Knowledge Sources that also support knowledge synchronization.
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2026.5.3

### Cognigy Voice Gateway

**Released** March 13, 2026

#### Bug Fixes

* Fixed the issue where calls were incorrectly declined with 603 SIP replies due to a rare race condition in speech provider gather tasks

## 2026.5.2

### Cognigy Voice Gateway

**Released** March 10, 2026

#### Bug Fixes

* Fixed the issue where calls using Deepgram as the speech-to-text provider end unexpectedly or didn't recognize some of the caller's inputs

## 2026.5.1

### Cognigy

**Released** March 12, 2026

#### Bug Fixes

* Fixed the issue where Flow execution was reset to the first Entrypoint when simultaneous calls from the same phone number were directed to more than one Flow. The issue affected Endpoints set to a Flow with an Execute Flow Node to handle call events. While the second Flow was executed, the first Flow to receive the call was reset

### Cognigy Voice Gateway

**Released** March 5, 2026

#### Bug Fixes

* Fixed the issue where STT labels set in the Set Session Config Node weren't loaded properly and fell back to `none`. This issue caused speech provider services to fail
* Fixed issue where the Voice Gateway Self-Service Portal displayed a failed status for valid Deepgram Flux speech credentials

## 2026.5.0

### Cognigy

**Released** March 3, 2026

#### Features

##### Knowledge Connector Syncing

Introduced the [Knowledge Connector Syncing](/ai/agents/develop/knowledge-ai/knowledge-connectors) feature to keep Knowledge Stores synchronized with external content sources, such as CMS or documentation systems. This feature reduces manual maintenance and ensures content consistency. The Confluence Extension 6.0.0 and the Diffbot Extension 2.0.0 include support for synchronization.

##### MCP Server Endpoint (Experimental)

Introduced the [MCP Server Endpoint](/ai/agents/deploy/endpoint-reference/mcp-server). It lets external AI clients, such as Claude Desktop and Cursor, discover and call AI Agent tools via the Model Context Protocol (MCP). This feature is experimental and isn't intended for production use.

#### Improvements

* Added support for [hashing user IDs in the transformers](/ai/for-developers/transformers/overview#modify-the-user-id-in-transformers) to enhance data protection and support GDPR compliance
* Added the Custom Options section to configure a custom model or request parameters for the [AI Agent](/ai/agents/develop/node-reference/ai/ai-agent#custom-options) and [Search Extract Output](/ai/agents/develop/node-reference/other-nodes/search-extract-output#custom-options) Nodes
* Updated the [`evaluateRule`](/ai/for-developers/function-reference/other/evaluateRule) function to return a promise, improving reliability in asynchronous workflows
* Added the [CXone Info](/ai/administer/access/management-ui#view-user-details) section, including the NiCE CXone email address and NiCE CXone ID, to the User Details in the Management UI. The section is displayed only if the user has a NiCE CXone account and the environment is integrated with NiCE CXone
* Improved the Salesforce MIAW handover provider stability by ignoring server-sent event (SSE) listeners after the connection was closed
* Improved Token search by prioritizing label-based matching to make Tokens easier to find
* Added the capability to redact the 3- and 4-digit CVV/CVC numbers using the [Credit Card Number redaction](/ai/platform-features/data-redaction#default-data-types) type in the Data Redaction Settings under Project Settings
* Updated the error message shown when no default LLM is found for scheduling a simulation run

#### Bug Fixes

* Fixed the issue where ISO date formatting didn't align with the configured time zone in the Date Picker Node
* Fixed the issue where transformer execution didn't correctly convert a `request-promise` proxy to an Axios proxy
* Fixed the issue where `foreignSessionData` wasn't sent in requests to Salesforce MIAW
* Fixed the issue where the Custom Model field wasn't shown for custom embedding models after being set to empty
* Fixed the issue where triggering an AI Agent without any user input resulted in an error
* Fixed the issue where cloned handover providers referenced non-existent providers after a user created a Snapshot or Package
* Fixed the issue where creating re-referenced handover provider IDs incorrectly caused handover Nodes to reference non-existent providers after creating a Snapshot or Package
* Fixed the issue where not all goals were correctly synced to Insights during Snapshot uploads
* Fixed the issue where merging two different Contact Profiles kept both Profiles active
* Fixed the issue where the Simulator OpenAPI documentation wasn't compatible with the latest OpenAPI version
* Fixed a memory leak in the `service-playbook-execution` service
* Fixed errors thrown when importing Packages containing Knowledge Connectors
* Fixed the issue where CXone-integrated SSO login didn't work in some environments due to an incorrect domain name check and the use of the wrong API URL

### Cognigy Voice Gateway

**Released** March 3, 2026

#### Improvements

* Improved short-answer recognition by starting the listening state before the voice AI Agent finishes speaking
* Improved voice detection accuracy by better aligning Barge-In and listening states
* Added concurrent use of the Continuous ASR and Barge-In features

#### Bug Fixes

* Fixed the issue where multiple `CALL_COMPLETED` events were sent for a single session
* Fixed synchronization issues to ensure the listening mode activates at the correct intervals
* Fixed the issue where synthesized speech from the Microsoft Azure Speech Services TTS provider was cut off. This issue caused missing last syllables or words
* Fixed the ElevenLabs STT provider configuration to correctly support the EU region setting

### Cognigy Live Agent

**Released** March 3, 2026

#### Improvements

* Improved the filter behavior in the Conversations list, including correct handling of AND/OR conditions, filter persistence, and folder navigation. The enhanced filter behavior prevents incorrect or missing conversation results

#### Bug Fixes

* Fixed issues with rendering chat components and updated the Live Agent library to align with current Cognigy components, ensuring proper support for complex AI Agent message types and full compatibility with Cognigy outputs
* Fixed the issue where human agents with status online didn't receive email notifications for new messages in assigned conversations, even when the email notifications were activated in the Profile Settings

### Cognigy Insights

**Released** March 3, 2026

#### Improvements

* Removed the `GET /v1.0/insights/token-resources` API method from the Insights API documentation
* Added the `x-api-key` authentication for Insights APIs, enabling simpler and direct API access without requiring JWT retrieval

#### Bug Fixes

* Fixed the issue where the `PATCH /v2.0/analytics` API method didn't work as expected
* Fixed the issue where the Goals report returned empty results for Snapshot-based Projects due to incorrect data handling in the report logic

### Infrastructure

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2026.6
Source: https://docs.cognigy.com/release-notes/2026.6



<Warning>
  **Deprecations:**

  * Support for the Completions API for custom LLMs has been deprecated. The full removal of this support is planned for December 2026. Refer to your [LLM provider's reference](/ai/agents/develop/gen-ai-and-llms/providers/all-providers) to switch to the appropriate API before the removal date.
  * Support for Google's `gemini-2.0-flash` and `gemini-2.0-flash-lite` models has been deprecated. The full removal of this support is planned for June 1, 2026. Switch to the newer models, such as `gemini-2.5-flash` and `gemini-2.5-flash-lite`, before the removal date.
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2026.6.4

### Cognigy

**Released** April 17, 2026

#### Bug Fixes

* Fixed the issue where users couldn't access the Simulator in a Project due to an internal infrastructure error

## 2026.6.3

### Cognigy

**Released** April 7, 2026

#### Bug Fixes

* Fixed the connection between Cognigy and Voice Gateway, ensuring Mute Speech Input prevents speech barge-in. This update complements the fix from [Voice Gateway patch 2026.6.2](#2026-6-2). To resolve this issue, make sure you update Cognigy to this version and Voice Gateway to 2026.6.2

## 2026.6.2

### Cognigy

**Released** March 25, 2026

#### Improvements

* [Added redaction limit](/ai/platform-features/data-redaction#limitations). Log entries with `data` payloads over 2.5 MB are no longer redacted and now show a placeholder message
* Improved internal log redaction for PCI compliance

### Cognigy Voice Gateway

**Released** April 7, 2026

#### Bug Fixes

* Fixed the issue where activating the Mute Speech Input setting didn't prevent speech barge-in

## 2026.6.1

### Cognigy

**Released** March 18, 2026

#### Bug Fixes

* Fixed the service configuration to ensure correct deployment and operation

### Cognigy Voice Gateway

**Released** March 31, 2026

#### Bug Fixes

* Fixed the issue on the Login page in the Voice Gateway Self-Service Portal. The page displayed v2026.5.x as the current version even after the v2026.6.0 release. Now, the page displays the correct version

## 2026.6.0

### Cognigy

**Released** March 17, 2026

#### Features

##### New Multimodal Agent Widget: Click To Call

[Click To Call](/click-to-call/overview) lets you easily embed a widget in your website or build a custom application that connects to a voice AI Agent for multimodal communication. With Click To Call, users can talk to your AI Agents directly from your website or application with a single click.

##### New Models from OpenAI and Microsoft Azure OpenAI

Introduced [support](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) for the `gpt-5-1` model from [OpenAI](/ai/agents/develop/gen-ai-and-llms/providers/openai) and [Microsoft Azure OpenAI](/ai/agents/develop/gen-ai-and-llms/providers/microsoft-azure-openai).

##### New Models from Anthropic

Introduced [support](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) for the `claude-sonnet-4-6` model from [Anthropic](/ai/agents/develop/gen-ai-and-llms/providers/anthropic).

#### Improvements

* Added a [Language Configuration toggle](#2-speaking-style) to the AI Agent wizard to stop automatic language detection and prevent unintended switching
* Added OAuth2 authentication support to the Cognigy API
* Deactivated Knowledge Connector actions when the required Extension is missing
* Added an error message when LLMs reach the token limit
* Added a safety context preamble to Simulator prompts, reducing content policy errors when using Azure OpenAI with personas or missions that contain terms flagged by Azure's content management filters
* Removed the beta tag from the Data Redaction settings section
* Renamed the Sentences API tag to Example Sentences for clarity

#### Bug Fixes

* Fixed the issue where the Management UI flickered when users were navigating between pages and scrolling content
* Fixed the issue where the error message caused by LLMs reaching the token limit didn't provide the correct information
* Fixed the issue where the Save button in the schedule configuration in the Simulator remained inactive without displaying any validation error message
* Fixed the issue where the `sourceName` value for scheduled sources in the Input object included the `_scheduled` postfix
* Fixed the issue where uploading Snapshots to Projects that didn't include a global LLM caused cache issues
* Fixed the issue where setting the maximum tokens above 4000 in the LLM Prompt Node for Microsoft Azure OpenAI's `gpt-4o` and `gpt-4o-mini` models threw an error, and similar limits for Anthropic's `claude-opus-4-0, claude-sonnet-4-0` and AWS Bedrock's `claude-3-5-sonnet` models
* Fixed the issue where setting the Transcript Turns parameter to 0 in the LLM Prompt Node transcript didn't have any effect
* Fixed the issue where Interaction Panel calls dropped if there was a Set Session Config Node in the Flow setting Microsoft provided by NiCE as TTS or STT vendor
* Fixed the issue where the wrong schedule date was displayed for Knowledge syncs when scheduling was previously deactivated
* Fixed the issue where the legacy `privacy_policy` option was displayed in the Profile Schema field in the Update Profile Node
* Fixed the issue preventing the import of Packages with Knowledge Chunks created using custom models
* Fixed the issue where the Last Edited columns on the Connector Configuration displayed edits that didn't take place
* Fixed the issue where the Knowledge Connectors didn't work in trial environments
* Fixed the issue where the error responses for the `POST /persona/from-script/` API method used `message` instead of `detail` and didn't provide the correct data

### Cognigy Voice Gateway

**Released** March 17, 2026

#### Bug Fixes

* Fixed the issue where the Register status wasn't updated correctly on the Carrier page in the Voice Gateway Self-Service Portal due to a race condition in the outbound authentication process

### Cognigy Live Agent

**Released** March 17, 2026

#### Bug Fixes

* Fixed the issue where the avatar image wasn't aligned properly
* Fixed the issue where notifications for conversations escalated to supervisors weren't translated

### Cognigy Insights

**Released** March 17, 2026

#### Improvements

* Redesigned the search logic in the Transcript Explorer to improve filtering and sorting against data in MongoDB
* Removed the default `order-by` condition in the OData query
* Added a lower timestamp bound to optimize deletion queries and avoid timeout errors

#### Bug Fixes

* Fixed the issue where sorting and filtering in the Transcript Explorer didn't work correctly
* Fixed the issue where the `PATCH /v2.0/analytics` API method failed while updating the PostgreSQL records due to a schema mismatch
* Fixed the issue where the `GET /v2.0/conversations` API method didn't work correctly and returned missing, incorrect, or `null` rating values
* Fixed the issue where messages sent through the WhatsApp Endpoint weren't displayed in the Transcript Explorer when the payload didn't contain top-level text

### Infrastructure

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2026.7
Source: https://docs.cognigy.com/release-notes/2026.7



<Warning>
  Deprecations in Cognigy 2026.7.0:

  * The States in the Interaction Panel, States functions, and States API methods have been deprecated. The removal is planned for Cognigy 2026.12.0. Delete your States and retrain your NLU model before the removal. Use [Intent Conditions](/ai/platform-features/nlu/intents/conditions) as an alternative for controlling Intent recognition.
  * The [`GET /v2.0/projects/{projectId}/logs`](/api-reference-logs/logs/get-log-entries) API method has been deprecated. The removal of this API method is planned for Cognigy 2026.20.0. Use the `GET /v2.1/projects/{projectId}/logs` API method with `start` and `end` parameters instead.
</Warning>

<Note>
  Introduced the [Cognigy MCP Server](/ai/for-developers/developers/cognigy-mcp-server), connecting MCP clients like Claude, Codex, Cursor, and VS Code to the Cognigy REST API. The server enables creation, testing, and iterative improvement of LLM-based AI Agents through a self-improvement loop.
</Note>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2026.7.2

### Cognigy

**Released** April 22, 2026

#### Improvements

* Increased the LLM request timeout for the Simulator from 10,000 ms to 20,000 ms. For on-premises installations, the LLM request timeout is now configurable via the `SIMULATION_LLM_TIMEOUT_MS` environment variable. The default value is 20,000 ms

#### Bug Fixes

* Fixed the issue in the Transcript explorer where session transcripts for Webchat v2 and v3 Endpoints didn't show AI Agent messages that only contained buttons or quick replies
* Fixed the issue where Extensions imported via Snapshot failed with the error `Unable to find extension module`. The stored file path in the Extension referenced the original pre-import ID instead of the new ID generated during the Snapshot import. The issue occurred after upgrading to Cognigy 2026.7.1
* Fixed the issue where users couldn't access the Simulator in a Project due to an internal infrastructure error

## 2026.7.1

### Cognigy

**Released** April 8, 2026

#### Bug Fixes

* Fixed the connection between Cognigy and Voice Gateway, ensuring Mute Speech Input prevents speech barge-in

### Cognigy Voice Gateway

**Released** April 8, 2026

#### Bug Fixes

* Fixed the issue where activating the Mute Speech Input setting didn't prevent speech barge-in

## 2026.7.0

### Cognigy

**Released** April 1, 2026

#### Features

##### Generate Scenarios from Conversation Transcripts in the Simulator

Introduced a powerful alternative for generating scenarios based on Flows or AI Agent Nodes in the Simulator. Now, you can create realistic scenarios from your existing conversation data to test your AI Agents with real-world interactions and improve their accuracy. Upload [conversation transcripts in CSV format](/ai/agents/test/simulator/overview#from-transcripts) to automatically generate scenarios.

##### New Models by OpenAI and Microsoft Azure OpenAI

Added support for new OpenAI and Azure OpenAI models: [`gpt-5.4`, `gpt-5.4-mini`, and `gpt-5.4-nano`](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature), allowing you to choose models based on performance and cost.

##### NiCE CXone Handover Provider (beta)

Introduced the [NiCE CXone](/ai/escalate/handover-reference/nice-cxone/nice-cxone-integrated) handover provider, allowing end users to hand over their conversations from an AI Agent to a human agent in a contact center. This feature is available for Cognigy installations integrated with NiCE CXone.

#### Improvements

* Updated `_.merge()` in Code Nodes so the code editor suggestions now match Lodash's behavior when you combine objects
* Added the capability to select [a time for Knowledge Connector sync](/ai/agents/develop/knowledge-ai/knowledge-connectors#scheduled-synchronizations)
* Removed the `Not understood` option from the filter in the Intent Trainer
* Moved the Connector Label field to the top of the Add Knowledge window. This improvement makes it easier for users to identify and name knowledge connectors before configuring other settings
* Improved OAuth2 handling for Azure OpenAI LLM connections. Custom headers are now sent with both the LLM request and the OAuth2 authentication request

#### Bug Fixes

* Fixed the Playbooks schema and description in the Cognigy OpenAPI documentation
* Fixed the issue where the Track Goal Node editor displayed extra headings and corrected the display text
* Fixed the issue with Contact Profile data deletion. Now, selecting Delete Date correctly clears all Contact Profile data and disables the Data Collection toggle
* Fixed the issue where LLMs returned raw JSON strings instead of parsed objects. This issue occurred with some LLM providers when the Response Format was set to JSON object in the Advanced section of the LLM Prompt Node
* Fixed the Overwrite Analytics Node so Input Data and Slots are no longer prefilled with `{}`. Leaving these fields empty now keeps existing analytics unchanged
* Fixed the issue where field regeneration failed after users added a goal on the New Scenario page in the Simulator
* Fixed the Package import failure caused by Node data containing an invalid `children: null` field
* Fixed the issue where the `<NO_USER_INPUT>` placeholder was included in the LLM context even when a real user input existed. Now, user inputs are sent correctly to the LLM

### Cognigy Live Agent

**Released** April 1, 2026

#### Improvements

* Added support for Microsoft 365 OAuth SMTP

#### Bug Fixes

* Fixed the Conversations Filter API method in the Live Agent OpenAPI documentation:
  * Corrected request body structure
  * Corrected auth header names
  * Fixed `not_equal_to` filter with `query_operator` causing `500 Internal Server Errors`

### Cognigy Insights

**Released** April 1, 2026

#### Bug Fixes

* Fixed the `$filter` behavior for the Sessions collection in the Cognigy OData endpoint when values in the `goals` field were stored as `text[]`. Now, filters using string functions, such as `contains`, `startswith`, and `endswith`, work correctly

### Infrastructure

#### Cognigy + Insights + Agent Copilot

##### Pagination Variable No Longer Required

If you previously added the `FEATURE_USE_NEW_PAGINATION_PLUGIN` variable in `cognigyEnv` or `extraEnvVars`, remove the variable before updating to Cognigy 2026.7.0. This variable is now set by default and no longer needs to be configured manually.

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2026.8
Source: https://docs.cognigy.com/release-notes/2026.8



<Warning>
  Introduced the new `ingress.ingresController` feature flag to controls whether Traefik middleware resources are created. For on-premises installations, refer to the [Infrastructure](#infrastructure) section.
</Warning>

<Note>
  Introduced the [Cognigy MCP Server](/ai/for-developers/developers/cognigy-mcp-server), connecting MCP clients like Claude, Codex, Cursor, and VS Code to the Cognigy REST API. The server enables creation, testing, and iterative improvement of LLM-based AI Agents through a self-improvement loop.
</Note>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2026.8.3

### Cognigy

**Released** May 6, 2026

#### Bug Fixes

* Fixed the issue where the Simulator wasn't accessible for on-premises installations with their own ingress controller implementation
* Fixed the issue where one end user calling AI Agents in different Projects in the same organization resulted in a single shared Contact Profile across Projects. This led to different AI Agents considering the end user as a recurring caller. Now, each Project creates their respective Contact Profiles separately in such cases

## 2026.8.2

### Cognigy

**Released** April 22, 2026

#### Improvements

* Increased the LLM request timeout for the Simulator from 10,000 ms to 20,000 ms. For on-premises installations, the LLM request timeout is now configurable via the `SIMULATION_LLM_TIMEOUT_MS` environment variable. The default value is 20,000 ms

### Cognigy Voice Gateway

**Released** May 15, 2026

#### Improvements

* Improved the stability of Microsoft Azure Speech Services speech-to-text and text-to-speech integrations. This improvement prevents potential connection issues and unexpected outages caused by changes to external services

### Cognigy Insights

**Released** April 22, 2026

#### Bug Fixes

* Fixed the issue in the Transcript explorer where session transcripts for Webchat v2 and v3 Endpoints didn't show AI Agent messages that only contained buttons or quick replies

## 2026.8.1

### Cognigy

**Released** April 16, 2026

#### Bug Fixes

* Fixed the issue where Extensions imported via Snapshot failed with the error `Unable to find extension module`. The stored file path in the Extension referenced the original pre-import ID instead of the new ID generated during the Snapshot import. The issue occurred after upgrading to Cognigy 2026.7.1
* Fixed the issue where users couldn't access the Simulator in a Project due to internal authentication and authorization checks

### Cognigy Voice Gateway

**Released** April 20, 2026

#### Bug Fixes

* Fixed the issue where a Say verb containing only an `<audio>` tag without any text for voice output didn't play the audio
* Fixed the issue where storing Microsoft Azure Speech Services credentials in on-premises installations required an invalid `region` field

### Cognigy Insights

**Released** April 16, 2026

#### Improvements

* Improved Insights stability

## 2026.8.0

### Cognigy

**Released** April 14, 2026

#### Features

##### Retrieve Knowledge Node

Introduced the [NiCE CXone Retrieve Knowledge Node](/ai/agents/develop/node-reference/cxone/retrieve-knowledge). With this Node, you can retrieve knowledge from [NiCE CXone Knowledge Hub](https://help.nicecxone.com/content/knowledgehub/knowledgehub.htm) and provide accurate information to your AI Agent. The Retrieve Knowledge Node is supported only in Cognigy installations hosted in the following environments:

* `cognigy-na1.nicecxone.com`
* `cognigy-na3.nicecxone.com`
* `cognigy-ca1.nicecxone.com`
* `cognigy-uk1.nicecxone.com`
* `cognigy-au1.nicecxone.com`
* `cognigy-jp1.nicecxone.com`

#### Improvements

* For the NiCE CXone handover provider:
  * Added the Display Agent Details toggle to the [Handover to Human Agent](/ai/agents/develop/node-reference/service/handover-to-agent) and [Question](/ai/agents/develop/node-reference/basic/question) Nodes. This toggle allows you to display the human agent's name and profile image in Webchat
  * Added support for the On Resolve Options setting in the [Handover to Human Agent](/ai/agents/develop/node-reference/service/handover-to-agent) Node
  * Added support for [bi-directional typing indicator](/webchat/v3/configuration#webchat-behavior) in Webchat v3
  * Added detection of the `ONRELEASE` action to determine when a [human agent ends the conversation](/ai/escalate/handover-reference/nice-cxone/nice-cxone-integrated#configuration-on-the-handover-provider-side)
* Included legacy API keys on the [My Profile page](/ai/administer/user-menu/my-profile) with the Legacy tag
* Improved error response text for failed [Create](/api-reference/handover-providers/create-a-handover-provider) and [Update](/api-reference/handover-providers/update-a-handover-provider) handover provider API requests
* Improved the detection to determine when human agents end conversations on Salesforce MIAW by using the `CONVERSATION_SESSION_STATUS_CHANGED` event instead of `CONVERSATION_PARTICIPANT_CHANGED`

#### Bug Fixes

* Fixed the issue where the LLM connection set at Project-level wasn't used in attached Flows when the Override Snapshot Connections options were activated in the Endpoint
* Fixed the issue where deactivating a Contact Profile with data deletion didn't set the Contact Profile to inactive
* Fixed the issue where JSON API responses using the HAL convention returned the internal host URL in embedded link URLs (`hrefs`) instead of the configured custom service URL
* Fixed the issue where messages added through the Add Transcript Step Node were ignored in LLM calls
* Fixed the issue where autogenerated scenarios failed when goals were included in the scenario configuration

### Cognigy Voice Gateway

**Released** April 14, 2026

#### Bug Fixes

* Fixed the issue where Say Nodes with text content including only whitespace or empty speech synthesis markup language (SSML) blocked Flow execution

### Cognigy Insights

**Released** April 14, 2026

#### Bug Fixes

* Fixed the issue where assigning `eq null` or `ne null` to the `$filter` query parameter in requests to OData endpoints returned no results. Now, null comparisons correctly filter records with missing values

### Infrastructure

#### Cognigy + Insights + Agent Copilot

For on-premises installations, the `.Values.traefik.enabled` feature flag is no longer used to control whether Traefik middleware resources are created. To control this mechanism, set the new `ingress.ingresController` feature flag in `values.yaml`:

* To activate the creation of Traefik middleware resources, set:
  ```yaml theme={null}
  ingress:
      ingresController: "traefik"
  ```
* To prevent the creation of Traefik middleware resources, set:
  ```yaml theme={null}
  ingress:
      ingresController: ""
  ```
  This option is necessary if you use non-Traefik controllers, such as NGINX, and set `traefik.enabled: false` in earlier releases.

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 2026.9
Source: https://docs.cognigy.com/release-notes/2026.9



<Warning>
  As of Cognigy 2026.8.3, the `ingress.ingresController` feature flag controls whether to create Traefik middleware resources. For on-premises installations, refer to the [Infrastructure section in the Cognigy 2026.8.0 release notes](/release-notes/2026.8#infrastructure).
</Warning>

<Warning>
  As of Cognigy 2026.9.0, the Set State and Reset State Nodes are deprecated and will be removed from the Node Selector in Cognigy 2026.18.0. This deprecation doesn't affect runtime behavior. After the removal, existing Flows containing this Node will remain fully supported and continue to run as expected. As an alternative, you can use [Intent Conditions](/ai/platform-features/nlu/intents/conditions) for controlling Intent recognition.
</Warning>

<Note>
  Introduced the [Cognigy MCP Server](/ai/for-developers/developers/cognigy-mcp-server), connecting MCP clients like Claude, Codex, Cursor, and VS Code to the Cognigy REST API. The server enables creation, testing, and iterative improvement of LLM-based AI Agents through a self-improvement loop.
</Note>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 2026.9.3

### Cognigy Voice Gateway

**Released** May 18, 2026

#### Improvements

* Improved voice call reliability by restarting FreeSWITCH if it becomes unresponsive
* Improved verb validation so unknown fields only warn, not fail, keeping integrations working after updates

## 2026.9.2

### Cognigy

**Released** May 6, 2026

#### Bug Fixes

* Fixed the issue where the Simulator wasn't accessible for on-premises implementations, as in [Cognigy patch 2026.8.3](/release-notes/2026.8#2026-8-3)
* Fixed the issue where end users calling different AI Agents in different Projects of the same organization resulted in a single shared Contact Profile across Projects. This led to different AI Agents perceiving the end user as a recurring caller. Now, each Project creates their respective Contact Profiles separately in such cases

### Cognigy Voice Gateway

**Released** May 15, 2026

#### Improvements

* Improved the stability of Microsoft Azure Speech Services speech-to-text and text-to-speech integrations. The improvement prevents potential connection issues and unexpected outages caused by changes to external services

## 2026.9.1

### Cognigy Voice Gateway

**Released** May 7, 2026

#### Bug Fixes

* Fixed the issue on the Speech Services page in the Voice Gateway Self-Service Portal where the Enable text-to-speech streaming option was displayed for speech vendors for which TTS streaming wasn't supported. Now, the option is displayed only for speech vendors listed in `{{ featureServer.streamingTtsVendors }}` in `values.yaml` that support TTS streaming

### Cognigy Insights

**Released** April 30, 2026

#### Bug Fixes

* Fixed the issue where the menu sidebar was misaligned
* Fixed the issue where the Transcript explorer failed to load due to a database schema mismatch

## 2026.9.0

### Cognigy

**Released** April 28, 2026

#### Features

##### Cognigy Serverless

Introduced [Cognigy Serverless](/ai/for-developers/extensions), providing fully isolated execution environments per organization for running Extensions, improving security, stability, and performance through automatic scaling and inactivity-based shutdown. Untrusted Extensions now run with full functional parity to trusted Extensions. To ensure compatibility with the new execution environment, review the Extensions documentation.

##### New Google GenAI LLMs

Introduced [Google GenAI](/ai/agents/develop/gen-ai-and-llms/providers/google-genai) as a new LLM provider, adding support for the Gemini 3 model family: `gemini-3.1-pro-preview`, `gemini-3-flash-preview`, and `gemini-3.1-flash-lite-preview`. This feature enables more flexible, high-performance AI capabilities across different use cases.

##### NiCE CXone (Non-Integrated) Handover Provider

Introduced the [NiCE CXone (Non-Integrated) handover provider](/ai/escalate/handover-reference/nice-cxone/nice-cxone-non-integrated), enabling handover to human agents in a NiCE CXone contact center without requiring a platform-level integration between Cognigy and NiCE CXone.

##### Send Status to CXone Copilot Node

Introduced the [Send Status to CXone Copilot](/ai/agents/develop/node-reference/cxone/send-copilot-status) Node, allowing Flows to send success or failure status to NiCE CXone Copilot for Agents so human agents can see the result in the CXone Agent Workspace. This feature is only available for Cognigy installations integrated with the NiCE CXone environment.

#### Improvements

* Removed the [End Handover on Participant Change](/ai/agents/develop/node-reference/service/handover-to-agent#salesforce-miaw-settings) setting from the Salesforce MIAW handover provider. Handover ending is now handled automatically on the Salesforce side
* Added NiCE branding by introducing a wordmark logo in Cognigy installations integrated with the NiCE environment. A configurable CXone button was also added to the sidebar and user menu for quick access to the CXone application

#### Bug Fixes

* Fixed the issue with the NiCE CXone handover provider where Package and Snapshot support didn't work correctly, and Snapshot `referenceId` values were overwritten before being stored in the database
* Fixed the issue where the Handover Providers sidebar item didn't stay highlighted when navigating to a Handover Provider Edit page
* Fixed the issue in the MCP Server Endpoint where the MCP server reconnects failed after a restored session. Now, a more stable reconnection behavior is ensured. Also updated the MCP landing page to be compatible with stricter browser [Content Security Policy (CSP) protections](/ai/agents/deploy/endpoint-reference/mcp-server#troubleshooting)
* Fixed the issue where adding invalid Knowledge Source Tags didn't show an error message and blocked the Save button
* Fixed the issue where stopping a simulator run mid-execution created duplicate simulation results in the Insights Transcript explorer due to metrics being recorded twice
* Fixed the issue where special German characters broke the spelled-out alphabet in the Text Cleaner and Question Nodes
* Fixed the issue in the Voice Preview feature where selecting a language not supported by the speech provider caused calls to fail. Now, Cognigy maps to a supported language variant or deactivates the call button if none is available
* Fixed the issue where Knowledge Sources didn't sync when the number of Knowledge Sources in the Knowledge Store exceeded the pagination limit of 25
* Fixed the issue in the Question Node where the Resolve Spelled Out Numbers option in the Answer Pre-Processing section incorrectly modified the last digit of large numbers, such as 18-digit values
* Fixed the issue where Nodes, such as the LLM Prompt Nodes, using tools became unstable when a low number of transcript turns was configured
* Fixed the issue where Flows broke when tool calls exceeded the transcript turns limit in the [LLM Prompt](/ai/agents/develop/node-reference/service/llm-prompt) Node
* Fixed the issue in the Snapshots feature that affected lifecycle operations such as creating, downloading, and restoring Snapshots

### Cognigy Voice Gateway

**Released** April 28, 2026

#### Improvements

* Added support for changing the media path [during an active transferred call via the Voice Gateway API](/ai/agents/develop/node-reference/voice/voice-gateway/transfer). This improvement lets you remove Voice Gateway from the audio stream during sensitive payment exchanges, reducing PCI data exposure.
* Improved search in the Voice Gateway Self-Service portal. Now, results include all records, not just the current page in paginated lists

### Cognigy Insights

**Released** April 28, 2026

#### Bug Fixes

* Fixed the issue in Transcript explorer where sessions were incorrectly ordered, placing sessions with newer last messages under older ones
* Fixed the issue where `time_savings_seconds` was 1,000 times higher and `goal_step_metric_count` was incorrectly calculated for currency and revenue metrics
* Fixed the issue where date range filters in Transcript explorer didn't correctly account for the user's local time zone offset
* Fixed the issue where the Locale filter in the Transcript Explorer only considered the final session Locale. Now, the filter includes sessions where the Locale was set earlier via a Switch Locale Node

### Infrastructure

#### Cognigy + Insights + Agent Copilot

For on-premises installations, introduced a new YAML-based configuration for data redaction in Cognigy 2026.9 and later, replacing environment variables. The new setup supports rollout percentage control and per-organization exclusions via `values.yaml`. For more information, read the [Data Redaction](/ai/platform-features/data-redaction#prerequisites) article.

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 4.100
Source: https://docs.cognigy.com/release-notes/4.100



<Warning>
  [WebSocket and HTTP Webhook](/ai/escalate/handover-reference/genesys-cloud-open-messaging#prerequisites) for Genesys Cloud Open Messaging have been deprecated. Migrate to [Amazon EventBridge](/ai/escalate/handover-reference/genesys-cloud-open-messaging-with-amazon-eventbridge).
</Warning>

<Warning>
  As of April 2025, the AudioCodes Endpoint is deprecated. The full removal of the AudioCodes Endpoint is planned for Q2 2026. Migrate to the [Voice Gateway Endpoint](/ai/agents/deploy/endpoint-reference/voice-gateway) before the removal date.

  Refer to the list of all upcoming removals on the [Deprecations and Removals](/release-notes/deprecations-and-removals) page.
</Warning>

<Tip>
  Help us improve our product documentation on `docs.cognigy.com` by sharing your thoughts in a [quick survey](https://forms.office.com/e/xnqneVasp2). Your feedback shapes the future of our content!
</Tip>

## 4.100.2

### Cognigy Live Agent

**Released** May 22, 2025

#### Bug Fixes

* Fixed the issue where editing custom fields in the Contact Profile details led to an error

## 4.100.1

### Cognigy

**Released** May 13, 2025

#### Bug Fixes

* Fixed the issue where the Steps and Tasks filters in the Transcript Explorer and the filters in the Step Explorer didn't load properly
* Fixed the issue where the Salesforce MIAW `baseUrl` validation failed in the backend, preventing the handover process from completing successfully

### Cognigy Voice Gateway

**Released** May 26, 2025

#### Bug Fixes

* Fixed the issue where users couldn't add phone numbers on the Phone numbers page in the Voice Gateway Self-Service Portal

### Cognigy Live Agent

**Released** May 22, 2025

#### Bug Fixes

* Fixed the issue where editing the default fields in the Contact Profile details led to an error

## 4.100.0

### Cognigy

**Released** April 29, 2025

#### Features

##### Amazon EventBridge Integration for Genesys Cloud Open Messaging

Implemented a new integration using [Amazon EventBridge for Genesys Cloud Open Messaging](/ai/escalate/handover-reference/genesys-cloud-open-messaging-with-amazon-eventbridge) to receive `UserEndEvent` and `AfterCallWorkEvent` events from Genesys, improving the reliability of conversation events.

##### Embedded AI Copilot in Salesforce MIAW

Added support to [embed AI Copilot](/ai/escalate/handover-reference/salesforce-miaw) into the Salesforce MIAW handover provider. Now you can seamlessly embed AI Copilot into your Salesforce MIAW handover workflow for a smoother support experience

#### Improvements

* Added support to [embed Agent Copilot](/ai/escalate/handover-reference/salesforce-miaw) into the Salesforce MIAW handover provider
* Implemented pre-aggregation of conversation counters, phased out of version 2 of the API routes for getting conversation counters, and introduced [version 3 of these API routes](https://api-trial.cognigy.ai/openapi#get-/v3.0/conversationcounter). The new API version improves the performance of conversation count retrieval for billing
* Added handover errors logs to Test > Logs
* Added [backward compatibility](/ai/agents/analyze/goals-and-tasks/tasks#compatibility-alias-to-profile-tasks) for `profile.goals`. Now, `profile.goals` points to `profile.tasks`, ensuring compatibility with all Flows

#### Bug Fixes

* Fixed the issue where Node editors opened in an unsaved state
* Fixed the issue where the debugging label for LLM Entity Extract Node wasn't localized in the Node editor
* Fixed the issue where calls on the Voice Call tab of the Interaction Panel didn't work properly in Safari
* Fixed the issue where intervention closed event from RingCentral Engage didn't trigger a `CALL_COMPLETED` event
* Fixed the issue where Knowledge Chunks with empty text were created when using a web page as a Knowledge Source
* Fixed the issue where clicking outside Node editors incorrectly triggered the `Unsaved changes` pop-up
* Fixed the issue where a postback message from Agent Copilot was forwarded to handover providers
* Fixed the issue where the queue update validation failed for `maximumQueuePosition`
* Fixed the issue where uploading a PDF file with empty table as a Knowledge Source resulted in an error
* Fixed the issue where importing a Package with a Knowledge Store resulted in the `Error while loading Knowledge Source during search ...` error in the Flow execution
* Fixed the issue where merged Contact Profiles were populated after receiving an input message

### Cognigy Voice Gateway

**Released** April 29, 2025

#### Improvements

* Added the [Profanity Filter option](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details) to the Advanced section in the Set Session Config Node and to Activity Parameters. You can use this configuration to handle swear words in transcripts when using Microsoft Azure Speech Services as a speech-to-text (STT) vendor
* Changed the default Profanity Filter option from `masked` to `raw`

#### Bug Fixes

* Fixed the issue where the values on the Phone numbers page in the Voice Gateway Self-Service Portal were changed after the page was fully loaded
* Fixed the issue where the Barge In function was disabled when a new `gather` input was received

### Cognigy Webchat

**Released** April 29, 2025

#### Improvements

* Added capability to configure [aria labels](/webchat/v3/accessibility#aria-labels) for the Webchat v3 components to improve accessibility
* Updated Webchat v3 to [v3.19.0](https://github.com/Cognigy/Webchat/releases/tag/v3.19.0)

### Cognigy Live Agent

**Released** April 29, 2025

#### Bug Fixes

* Fixed the issue where clicking Cognigy in the Profile settings menu redirected users to the Cognigy login page, but assigned `undefined` to `organisationId` in the URL
* Fixed the issue where some actions in Automation Rules were using incorrect validations
* Fixed the issue where Live Agent stopped working due to Live Agent connecting to a read-only Redis replica instead of to `master`

### Cognigy Insights

**Released** April 29, 2025

#### Bug Fixes

* Fixed the issue where Step Explorer didn't display correctly when using pagination
* Fixed the issue where OData queries that were canceled by client were leaving broken connections to the database
* Fixed the issue where CognigyScript didn't parse `goals` in the Profile object correctly
* Fixed the issue where `endpointUrlToken` wasn't in the `stepevents` table
* Fixed the issue where setting Slots in an Overwrite Analytics Node broke the Transcript Explorer
* Fixed the issue where the Time Span filters reset when users switched Projects. Now, the Time Span filters stay consistent across Projects

### Infrastructure

#### Cognigy

* Introduced the `pre-aggregate-conversations` cron job that is scheduled to run daily to pre-aggregate billable conversations

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Cognigy Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 4.30
Source: https://docs.cognigy.com/release-notes/4.30



<Warning>
  With release 4.30.0, we have officially marked our Kubernetes repository as `deprecated`. If you are an `on-premises` customer and you are running Cognigy/Cognigy Insights yourself, have a look at our [Helm Chart](https://github.com/cognigy/cognigy-ai-helm-chart), which we have crafted for Cognigy/Cognigy Insights! Our Kubernetes repository will still receive updates `until 31st December 2022`.
</Warning>

## 4.30.0

### Cognigy

**Released** July 18, 2022

#### Features

##### WhatsApp Cloud API endpoint

This release of Cognigy contains the new WhatsApp Cloud API endpoint,
which allows Cognigy customers to seamlessly deploy their AI Agents on WhatsApp.

#### Improvements

* Improved by updating the Webchat Widget to version 2.41.2
* Improved by adding the possibility to delete orgs using the UI through a multistep process
* Send the channel type to Live Agent
* Improved Live Agent synergy by adding an additional step to the Agent Creation Wizard
* Improved by upgrading the Helmet package version, so it includes the security header "Content-Security-Policy"
* Improved UX of the Management-UI for managing multi-org users
* Improved UX of Logs page by adding reverse infinite scroll for log entries
* Improved UX by preserving resource selection in Run Playbook Dialog across Agent routes
* Improved UX by hiding the override snapshot connections toggle when no snapshot is selected

#### Bug Fixes

* Fixed a bug where:
  * A "Submit" button was attached to the handover accepted message in MS Teams
  * The handover accepted message was not displayed to slack users
  * Handover to Meta Messenger is only done if a cancel button was added
  * The handover accepted message was not displayed to WhatsApp and Meta Messenger if no cancel button was added
* Fixed security vulnerabilities in service-endpoint, service-handover, service-alexa-management and service-webchat
* Fixed a bug where notify and inject transformers should be able to log info, debug and error logs
* Fixed a bug where deleting the user from Cognigy will also delete the user from LiveAgent
* Fixed a bug where the openAPI document connections API needed to be corrected
* Fixed a bug where handover settings and transformer error notifications used to show wrong title
* Fixed a bug where an error message wasn't displayed when a Lexicon Slot had more than 30 characters
* Fixed a bug where Optional Question nodes using the "Stop Flow execution after Answer" setting would cause an error
* Fixed a bug where the understood property would be incorrect after using executeFlow or Go To

### Cognigy Insights

**Released** July 18, 2022

#### Improvements

* Improved by adding dropdown to Transcript Explorer with options to create a Playbook and redirect to the Playbook editor in Cognigy AI

#### Bug Fixes

* Fixed a bug: Remove default timezone warning
* Fixed a bug: Fixed webpack config

### Cognigy Live Agent

**Released** July 18, 2022

#### Improvements

* Add Cognigy channel to the conversation information section and also add the possibility to filter conversations based on this field

#### Bug Fixes

* Fixed a bug where the Cognigy user profile in the conversation was emptied while receiving new incoming messages into the conversation


# 4.31
Source: https://docs.cognigy.com/release-notes/4.31



<Warning>
  With release 4.30.0, we have officially marked our Kubernetes repository as `deprecated`. If you are an `on-premises` customer and you are running Cognigy/Cognigy Insights yourself, have a look at our [Helm Chart](https://github.com/cognigy/cognigy-ai-helm-chart), which we have crafted for Cognigy/Cognigy Insights! Our Kubernetes repository will still receive updates `until 31st December 2022`.
</Warning>

## 4.31.0

### Cognigy

**Released** August 2, 2022

#### Features

##### WhatsApp Cloud rich media support

This release of Cognigy adds support for rich media elements for the WhatsApp Cloud channel. You will now have the chance to output Locations, Contacts and Rich Links as part of our Say-Node in the Flow-Editor when specifying outputs for the WhatsApp Cloud channel.

#### Improvements

* Improved by exposing errors related to CognigyScript parsing on the logs page
* Improved by ensuring that the cognigy\_ai\_brain\_in\_memory\_count metric shows the correct state
* Improved by making the demo webchat scalable
* Support for adaptive cards on the Webchat tab of the Say Node
* Updated the Webchat Widget to version 2.41.2
* Improved by adjusting the agent assist flow node

#### Bug Fixes

* Fixed a bug that leads to a crash as soon as there was an executeFlow node with parseSlots enabled
* Fixed a bug where there is wrong usage of icon for AudioCodes and voice-gateway / also wrong output for voice-gateway text in chat history
* Fixed a bug where you could not delete multiple agents one after another
* Fixed a bug where we were not able to save multiple changes in say node
* Fixed a bug where the older question nodes in an agent are not displaying the field data on selecting it the first time after the handover properties are released for the node, thereby "last updated user" is displayed as zeros
* Fixed a bug when unsupported WhatsApp input data was not passed to the flow

### Cognigy Insights

**Released** August 2, 2022

#### Improvements

* Improved by adding Highlighting the transcripts by changing the background color of the transcript list table to white

#### Bug Fixes

* Fixed a bug: insights not loading

### Cognigy Live Agent

**Released** August 2, 2022

#### Improvements

* Improved by redesigning the Whisper agent assist and adding new info cards to it
* Improved by making the conversation Cognigy Channel type field human readable
* Add the possibility to edit contact profiles by agents directly in the conversation information section, the change will be reflected in Cognigy as well
* Improved by adding user ability to set their status to "away"
* Improved by adding an "Unassign" command in place of the "None" agent in the conversation action section, so that users can more easily unassign conversations
* Improved by adding inbox avatar to conversation list and having both upload and delete operations propagate immediately without needing a page refresh
* Improved by allowing rendering of info cards using the new JSON payload proportioned by Cognigy
* Improved by adding human-readable channel mapping to conversation filters and conversation information section
* Improved by fixing a memory leak while querying conversations
* Improved by making info card images' size more dynamic. It adds preview text for agent assist messages and the possibility to use line breaks in info card bodies. It fixes floating arrows when hovering and galleries not working correctly when hovered

#### Bug Fixes

* Fixed a bug where availability status colors were not propagating correctly
* Fixed a bug where user who clicks the assigned agent unassigns the conversation
* Fixed a bug where agent availability was set back to online after reloading the page


# 4.32
Source: https://docs.cognigy.com/release-notes/4.32



<Warning>
  With release 4.30.0, we have officially marked our Kubernetes repository as `deprecated`. If you are an `on-premises` customer and you are running Cognigy/Cognigy Insights yourself, have a look at our [Helm Chart](https://github.com/cognigy/cognigy-ai-helm-chart), which we have crafted for Cognigy/Cognigy Insights! Our Kubernetes repository will still receive updates `until 31st December 2022`.
</Warning>

## 4.32.1

### Cognigy

**Released** August 24, 2022

#### Bug Fixes

* Fixed a bug where merging a package was deleting NLU models

## 4.32.0

### Cognigy

**Released** August 17, 2022

#### Improvements

* Remove "powered by Cognigy" link from webchat when whitelabeling flag is enabled
* Add default, webchat, plugin Adaptive cards rendering to Interactive Panel
* Improved by adjusting the agent assist flow node
* Improved by making the conversation counter timezone aware
* Improved the error message when executing a Flow Node failed

#### Bug Fixes

* Fixed a bug where the Salesforce Handover Integration doesn't provide the full conversation history
* Adjusted log levels from 'error' to 'debug' for anonymous callers
* Fixed a bug where two identical user IDs would create different Contact Profiles when containing a "+" sign
* Fixed a bug where the agent assist output appeared in channels other than live-agent
* Fixed a bug where Initializing a second handover request in live agent using a different inbox would fail
* Fixed a bug where packages with intents created with a name starting/ending with white space were not accepted
* Fixed a bug where the password reset functionality would not work for users which are part of multiple organizations
* Fixed a bug where AI channel preview for Adaptive Card output type was not working

### Cognigy Insights

**Released** August 17, 2022

#### Improvements

* Improved by adding the info Tooltips in Insights dashboards
* Improved by including 'today' under 'Last X days'
* Improved by Making Contact ID copyable in the Transcript explorer
* Changed duration formatting for the number charts
* Changed background color in transcript explorer
* Fixed sidebar animation
* Added Search field for Flows + order of flows for the Transcript Explorer

#### Bug Fixes

* Fixed a bug: Message explorer search filter doesn't work
* Fixed a bug: The user has to refresh their Insights dashboard to see the Actual Active contacts
* Fixed a bug: Close dropdown icon in the Insights dashboards left sidebar doesn't work
* Fixed a bug: Close dropdown icon in the Insights dashboards left sidebar doesn't work

### Cognigy Live Agent

**Released** August 17, 2022

#### Improvements

* Improved by adding the new status "Aborted" to conversations that didn't get a customer response. It is configurable under the Account Settings section
* Improved by adding a link to inbox settings in the assist bot sidebar and persisting selected flow in conversation when the sidebar is closed

#### Bug Fixes

* Fixed a bug where the text of user messages in assist bot was not correctly aligned
* Fixed a bug where the input box was not staying at the bottom on big screens, and assist bot chat messages were misaligned

### Infrastructure (AI + IN)

#### Stateful services

To improve the functionality and fix known security vulnerabilities in the dependencies Cognigy/Cognigy Insights uses, we have updated the following dependencies:

* Redis, from `v5.0.8` to `v5.0.14`
* RabbitMQ, from `v3.8.3` to `v3.9.20`

In addition, we have released a new version of our `MongoDB Helm Chart` which you can use to deploy a 3-node MongoDB replica-set. We have increased the MongoDB version from `v4.2.5` to `v4.2.21` in this new Helm Chart. For more information, refer to the [MongoDB Helm Chart](https://github.com/Cognigy/cognigy-mongodb-helm-chart)  repository.

Updating to this release requires you to update these `stateful services`. Doing this will require restating all application Pods and will come with a couple of seconds of downtime. In order to restart all Pods after you have updated MongoDB, RabbitMQ, Redis and Redis-persistent, you can use the following command:

```bash theme={null}
for i in $(kubectl get deployment --namespace <target_namespace> --template '{{range .items}}{{.metadata.name}}{{"\n"}}{{end}}'|grep service-)
do
kubectl --namespace <target_namespace> rollout restart deployment $i
done
```

Be sure to replace the `<target_namespace>` with e.g. something like `cognigy-ai`.

#### NLU services - optional

In order to avoid that you need to provision unnecessary resources in your Kubernetes clusters, we now require you to select which `NLU languages` you actually need in your Cognigy setup. If you, for instance, want to only build bots in English language, you can activate English NLU by doing the following in your `values.yaml` file:

```yaml theme={null}
## Enable your choice of NLU service by following the below example
serviceNlpScoreEn:
  enabled: true
serviceNlpTrainEn:
  enabled: true
```

#### New services

You will also see that this release introduces two new microservices:

* service-insights-ui
* service-insights-api

We are slowly separating Cognigy.Insights from Cognigy in order to improve scalability, stability and performance of both products. To prepare your installation, follow these steps closely in order to make sure that the upgrade is smooth:

Create a new DNS entry for `service-insights-api` and point it towards your loadbalancer which was provisioned when you have installed Cognigy. A potential example would be the following:

Let's assume that your `Cognigy API` is available on the DNS name `api-trial.cognigy.ai` a good name for the new Insights API service would be `insights-api-trial.cognigy.ai` - so we e.g. just prefix the Cognigy API address with `insights-`.

Using a command like `dig` on Linux should reveal that both DNS names point to the same Loadbalancer address:

```bash theme={null}
dig api-trial.cognigy.ai

; <<>> DiG 9.10.6 <<>> trial.cognigy.ai
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 26015
;; flags: qr rd ra; QUERY: 1, ANSWER: 4, AUTHORITY: 0, ADDITIONAL: 1

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 1232
;; QUESTION SECTION:
;trial.cognigy.ai.              IN      A

;; ANSWER SECTION:
trial.cognigy.ai.       300     IN      CNAME   ae79c0ec73eca4d64a0509f7a75fe42a-1344350581.eu-central-1.elb.amazonaws.com.
ae79c0ec73eca4d64a0509f7a75fe42a-1344350581.eu-central-1.elb.amazonaws.com. 60 IN A 52.29.55.202
ae79c0ec73eca4d64a0509f7a75fe42a-1344350581.eu-central-1.elb.amazonaws.com. 60 IN A 3.121.18.227
ae79c0ec73eca4d64a0509f7a75fe42a-1344350581.eu-central-1.elb.amazonaws.com. 60 IN A 3.64.138.183

;; Query time: 30 msec
;; SERVER: 192.168.0.1#53(192.168.0.1)
;; WHEN: Wed Aug 10 18:37:42 CEST 2022
;; MSG SIZE  rcvd: 181
```

Performing the same `dig` command for the new Insights API server should return similar data.

The last step is to make sure that you add the following key/value pair to the `cognigyEnv` section of your `values.yaml` file:

```yaml theme={null}
FEATURE_USE_SERVICE_INSIGHTS_UI: "true"
```


# 4.33
Source: https://docs.cognigy.com/release-notes/4.33



<Warning>
  With release 4.30.0, we have officially marked our Kubernetes repository as `deprecated`. If you are an `on-premises` customer and you are running Cognigy/Cognigy Insights yourself, have a look at our [Helm Chart](https://github.com/cognigy/cognigy-ai-helm-chart), which we have crafted for Cognigy/Cognigy Insights! Our Kubernetes repository will still receive updates `until 31st December 2022`.
</Warning>

## 4.33.1

### Cognigy

**Released** September 7, 2022

#### Improvements

* Improved by storing response data from Salesforce into the Input object after starting a Handover

#### Bug Fixes

* Fixed a bug where the ChatRequestFailed event from Salesforce would cause the Handover status to be set to "completed" instead of "error"
* Fixed a bug where context updates during agentInject where lost when the Handover was finished

## 4.33.0

### Cognigy

**Released** September 2, 2022

#### Features

* Added the feature to be able to use the intent model of a fallback locale

#### Improvements

* Improved the Suggested Reply card of the Agent Assist Card flow node
* Verify the integrity of translation strings
* Improved by keeping the Agent NLU Settings section as collapsed by default
* Improved by refining the order of settings in Interaction Panel
* Improved by adding missing events to the Interfaces and Endpoints
* Display Adaptive cards in Transcripts in Contact profiles and Insights

#### Bug Fixes

* Fixed a bug where the input object contained the wrong nlu properties. This happened when an intent-confirmation-answer triggered a new intent instead
* Add traceId to readConversationRpc
* Remove logging on product logs page for "user-connected" and "user-disconnected" events in the endpoint

### Cognigy Insights

**Released** September 2, 2022

#### Features

* Improve TE performance (MongoDB iops and DB) for large data
* Tooltip on Insights metrics with link to Cognigy documentation

#### Improvements

* Show Insights dashboard dropdown options (and keep the uparrow active)

#### Bug Fixes

* Not all steps available in starting steps dropdown
* filter-does-not-work-for-emails

### Cognigy Live Agent

**Released** September 2, 2022

#### Features

* Improved by adding the possibility to Auto Set Away agents from the account settings
* Improved by adding an option to set max conversation limit in account settings. This limit gates any new assignment of conversations (via bulk assign, manual single assign or auto-assign/auto-reassign) if an agent has already reached the conversation limit. As part of this, the conversation action section was reworked to include an "unassign" action for teams
* Improved by adding the ability to clear the conversation chat history with a new delete button
* Improved by adding the ability to remove the account conversations automatically after resolving them. Enabling this option under the account settings will remove the already resolved

#### Improvements

* Improved by adding a coherent information message when data for reports is insufficient
* Improved by changing the availability colors for online, busy, away, and offline statuses to be better recognized

#### Bug Fixes

* Fixed a bug where conversations were not filtering correctly when the cognigy channel type was a camel-case string (e.g. filters for microsoftTeams were returning no results, while filters for "alexa" were working)


# 4.34
Source: https://docs.cognigy.com/release-notes/4.34



<Warning>
  With release 4.30.0, we have officially marked our Kubernetes repository as `deprecated`. If you are an `on-premises` customer and you are running Cognigy/Cognigy Insights yourself, have a look at our [Helm Chart](https://github.com/cognigy/cognigy-ai-helm-chart), which we have crafted for Cognigy/Cognigy Insights! Our Kubernetes repository will still receive updates `until 31st December 2022`.
</Warning>

## 4.34.0

### Cognigy

**Released** September 13, 2022

#### Features

##### French user interface

This release of Cognigy adds French as a new language in our Cognigy user interface

#### Improvements

* Improved by adding white background to all lists
* Improved the datepicker in a question node by adding the possibility to provide a function to enable or disable dates
* Improved by replacing pickle storage format of Intent models with mixed plain text / byte file hierarchy
* Forward whisper and live agent assist messages to Live Agent using a dedicated channel
* Add support for forwardable whisper and live agent assist messages in third-party endpoints
* Create a new Webchat release 2.42.0

#### Bug Fixes

* Fixed a bug where the "index logs" API which is also used by our logs page would not properly work when using filters
* Fixed a bug where "an dem 18.6." was not recognized as a date in German
* Fixed a bug where some invalid fields in audio preview settings did not allow updating the agent settings
* Fixed a bug where merging a package was deleting NLU models

### Cognigy Insights

**Released** September 13, 2022

#### Bug Fixes

* Fixed an issue where data in our OData analytics feed was double JSON stringified
* Fix date range filtering option

### Cognigy Live Agent

**Released** September 13, 2022

#### Improvements

* Improved by switching the Live Agent wording for the conversation status from "Aborted" to "Abandoned"
* Improved by adding status colors for notifications shown while actions are performed, such as updating settings or assigning conversations

#### Bug Fixes

* Fixed a bug where the reports number are inconsistent depending on the Live Agent client timezone and date range applied
* Fixed a bug where the agent's name was cut off in the Reports Overview


# 4.35
Source: https://docs.cognigy.com/release-notes/4.35



<Warning>
  With release 4.30.0, we have officially marked our Kubernetes repository as `deprecated`. If you are an `on-premises` customer and you are running Cognigy/Cognigy Insights yourself, have a look at our [Helm Chart](https://github.com/cognigy/cognigy-ai-helm-chart), which we have crafted for Cognigy/Cognigy Insights! Our Kubernetes repository will still receive updates `until 31st December 2022`.
</Warning>

## 4.35.1

### Cognigy

**Released** September 26, 2022

#### Bug Fixes

* Fixed a bug where ongoing handover requests would fail for the "Salesforce Handover Provider" due to an incompatible data-format

## 4.35.0

### Cognigy

**Released** September 23, 2022

#### Features

##### Audit Events dashboard

This release of Cognigy adds a new Audit Events dashboard. This new dashboard is only usable for users with the "admin" role assigned. The dashboard shows all operations which have manipulated objects in Cognigy such as "create"/"update" & "delete" operations.

##### Attachment Handling for WhatsApp Cloud Endpoint

This release of Cognigy adds support for handling attachments such as video, audio and file-uploads through the Whatsapp Cloud Endpoint.

#### Improvements

* Create a new Webchat release 2.42.1
* Improved by adding a metric that shows if a service is connected to dependencies such as MongoDB
* Improved by adjusting the Liveness Probe to align more with industry best practices
* Improved by adding a startup check for whether NFS is working
* Improved reconnect handling for RabbitMQ connections
* Improved reconnect handling for Redis connections
* Improved by the error message for failures in an Extension
* Improved by storing response data from Salesforce into the Input object after starting a Handover

#### Bug Fixes

* Fixed a bug where RCE handover was not working when the "Forward Only Handover Conversations" toggle is OFF for first time in the chat session (works again on next message) in the handover settings section of the endpoint
* Fix Line structured content converter issue
* Fixed a bug where manually triggered intents (cIntents) got no property intentLevel in the input object
* Fixed a bug where uploading intents in secondary locales that are localized, would not work
* Fixed a bug where the Agent Assist Cards would not be sent to LA properly if no Handover Accepted Message was defined in the Handover to Agent Node
* Fixed a bug where activating a deactivated Profile did not work
* Fixed a bug where the ChatRequestFailed event from Salesforce would cause the Handover status to be set to "completed" instead of "error"
* Fixed a bug where context updates during agentInject where lost when the Handover was finished
* Fixed a bug where intent trainer records would not be stored if the triggered Intent had a Default Reply
* Fixed security vulnerabilities in the Runtime services for Mongoose
* Fixed a bug where there could be a possible attack vector on "indexUsers" RESTful API. The API is now restricted to users with admin/userManager/userDetailsViewer global roles or with roles projectAdmin/memberManager in any assigned project
* Fixed a bug where `<speak>` tag was mandatory to work with Voice Preview SSML
* Fixed a bug where training out of date error badge will not hide immediately after training
* Fixed a bug where html is not rendered in a standard message in the Interaction panel
* Fixed a bug where index trainer records timed out

### Cognigy Insights

**Released** September 23, 2022

#### Bug Fixes

* Fixed a bug where Step explorer was crashing
* App crash if Step Explorer API call returns invalid data
* App crash if the custom theme could not be loaded
* Transcript Explorer - Channel Filtering

### Cognigy Live Agent

**Released** September 23, 2022

#### Improvements

* Removed the OAuth pop-up window from the login process. Users are now redirected and authenticated within the same browser window

#### Bug Fixes

* Fixed a bug where the conversations tab counters (me, unassigned, total) are sometimes incorrect
* Fixed a bug where mentions were not being removed when clearing chat history

### Infrastructure (AI + IN)

#### New Services

This release contains a new microservice:

* service-runtime-file-manager
* clamd

This new service provides functionality which is currently only used by our new
WhatsApp Cloud Endpoint and allows end-users to upload file attachments such as
images, audio, video and e.g. files like PDF. Once a file has been received, it
will be checked for infections with the "clamd" component.

##### DNS record

You will have to create a new DNS record for `service-runtime-file-manager` as
this is a new microservice which will also be fully exposed to the external/public
network. The DNS record needs to be pointed to your loadbalancer which has been
provisioned when Cognigy was installed initially. The following would be an
example of what this can look like:

Let's assume that your `Cognigy API` is available on the DNS name `api-trial.cognigy.ai`

* a good DNS name for the new microservice would e.g. be `files-api-trial.cognigy.ai` -
* so we have essentially just prefixed the `Cognigy API` address with `files-`.

Using a command like `dig` on Linux needs to reveal that both DNS names (old one and
the new one after you have created it) should point to the same loadbalancer address:

```bash theme={null}
dig api-trial.cognigy.ai

; <<>> DiG 9.10.6 <<>> trial.cognigy.ai
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 26015
;; flags: qr rd ra; QUERY: 1, ANSWER: 4, AUTHORITY: 0, ADDITIONAL: 1

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 1232
;; QUESTION SECTION:
;trial.cognigy.ai.              IN      A

;; ANSWER SECTION:
trial.cognigy.ai.       300     IN      CNAME   ae79c0ec73eca4d64a0509f7a75fe42a-1344350581.eu-central-1.elb.amazonaws.com.
ae79c0ec73eca4d64a0509f7a75fe42a-1344350581.eu-central-1.elb.amazonaws.com. 60 IN A 52.29.55.202
ae79c0ec73eca4d64a0509f7a75fe42a-1344350581.eu-central-1.elb.amazonaws.com. 60 IN A 3.121.18.227
ae79c0ec73eca4d64a0509f7a75fe42a-1344350581.eu-central-1.elb.amazonaws.com. 60 IN A 3.64.138.183

;; Query time: 30 msec
;; SERVER: 192.168.0.1#53(192.168.0.1)
;; WHEN: Wed Aug 10 18:37:42 CEST 2022
;; MSG SIZE  rcvd: 181
```

Make sure that the `dig` command shows similar information after you have provisioned
the new DNS record.

##### Values modifications

Make sure that you also update your `Helm values` and implement the following changes in them:

In the `ingress` section, make sure that you define a `host` for the runtime file manager:

```yaml theme={null}
serviceRuntimeFileManager:
  host: "files-api-trial.cognigy.ai"
```


# 4.36
Source: https://docs.cognigy.com/release-notes/4.36



<Warning>
  With release 4.30.0, we have officially marked our Kubernetes repository as `deprecated`. If you are an `on-premises` customer and you are running Cognigy/Cognigy Insights yourself, have a look at our [Helm Chart](https://github.com/cognigy/cognigy-ai-helm-chart), which we have crafted for Cognigy/Cognigy Insights! Our Kubernetes repository will still receive updates `until 31st December 2022`.
</Warning>

## 4.36.3

### Cognigy Insights

**Released** October 27, 2022

#### Bug Fixes

* Fix a bug where certain session information could not be stored due to schema validation problems
* Fix an issue where syncing database indices would remove custom indices customers have created

## 4.36.2

### Cognigy

**Released** October 14, 2022

**Important Instruction**:

For on-premises installation, to enable login to Cognigy with SSO, the environment variable `CLIENT_ID_COGNIGY_LIVE_AGENT=cognigy-live-agent` should be set.

#### Bug Fixes

* Fixed a bug where SSO sign-in was not possible for Cognigy Live Agent

## 4.36.1

### Cognigy

**Released** October 7, 2022

#### Improvements

* Improved by increasing the timeout for fetching the conversation counter

### Cognigy Live Agent

**Released** October 14, 2022

#### Bug Fixes

* Fixed all email transcript requests broken
* Fixed broken sso login

## 4.36.0

### Cognigy

**Released** October 5, 2022

#### Features

##### Google Cloud Storage Provider

This feature of Cognigy adds support for "Google Cloud" as a storage provider for the "File Upload" functionality located in the WhatsApp Cloud Endpoint

#### Improvements

* Improved by making Analytics Step field searchable during Node search & searching Nodes that have an analytics step set searchable
* Improved by redesigning the input mode selection in Interaction panel
* Improved by adding resource names to the delete confirmation prompts
* Improved by changing name of Agent Assist endpoint
* Improved by adding indices to collections of service-resources and service-trainer. This improves performance while retrieving intents, feedbackreports, and trainer-records
* Improved by redirecting to "Insights Transcript Viewer" when clicking on a conversation item in Cognigy
* Displaying of ratings in the conversation history improved, they are now in the correct order

#### Bug Fixes

* Fixed bug where timestamps in Live Agent were incorrect.
* Fixed bug where implementation of "send transcript as first message" for Salesforce was broken.
* Fixed a bug where "1111€." was not recognized as a number, because of the following dot after the euro sign
* Fixed a bug where lexicon api OpenAPI specifications would show an invalid field
* Fixed a bug where we got an error when uploading a package with duplicate lexicon slots
* Fixed a bug where logs screen gives a timeout for the basic filtering using type
* Fixed a bug where the tooltip description for the Subscription Region under Voice Settings was misleading
* Fixed a bug where password reset and change password gives an error, but the actual operation worked.

### Cognigy Insights

**Released** October 5, 2022

#### Features

##### Heatmap as new Report type

This release of Cognigy Insights adds a "heatmap" as a new report type on the Engangement dashboard visualizing the number of handovers per time-interval

##### Detailed Views for Reports

This release of Cognigy Insights adds the ability to view more details for each report type. The detailed view either just increases the visual space on which a report will be visualized or even adds additional filtering capabilities and displays a larger value-range

#### Improvements

* Implementing Dashboard print styles in Google Chrome
* Implementing ACL for Insights
* Adding the logic for handling rating as input and output

#### Bug Fixes

* Fixed expanded path calculation in Step Explorer

### Cognigy Live Agent

**Released** October 5, 2022

#### Features

* Assist bot is the new way to get assistance while chatting with your customers. It is now available on the right-hand side of the conversation. Assist bot endpoints to need to be configured in the Cognigy Live Agent Inbox settings

* Improved by adding a new section under settings called "Attachments", where attachments contained in conversation messages can be reviewed and deleted

* Improved by adding the possibility to automatically delete attachments by specifying a TTL per account. The setting is available under the account settings

* Improved by adding a new feature, called 'Agent Masking'. Enable this feature in account settings in order to have all agent names, email addresses, and images randomized (some agent lists will still be unmasked for administrative purposes - e.g. to see which agents are in the account, add agents to the correct inbox, etc.).

#### Improvements

* Improved by renaming the unattended property to queued for the reports page

#### Bug Fixes

* Improved by removing a bottleneck while retrieving the list of conversations


# 4.37
Source: https://docs.cognigy.com/release-notes/4.37



<Warning>
  With release 4.30.0, we have officially marked our Kubernetes repository as `deprecated`. If you are an `on-premises` customer and you are running Cognigy/Cognigy Insights yourself, have a look at our [Helm Chart](https://github.com/cognigy/cognigy-ai-helm-chart), which we have crafted for Cognigy/Cognigy Insights! Our Kubernetes repository will still receive updates `until 31st December 2022`.
</Warning>

## 4.37.0

### Cognigy

**Released** October 18, 2022

**Important Instruction**:

For on-premises installation, to enable login to Cognigy with SSO, the environment variable `CLIENT_ID_COGNIGY_LIVE_AGENT=cognigy-live-agent` should be set.

#### Improvements

* Improved by adding an option to also display a system message below the login form
* Added an option to auto hang up after all input retries in AudioCodes endpoints

#### Bug Fixes

* Fixed a bug where flow path highlighting was incorrect during live follow and voice call
* Fixed a bug where password reset and change password would give an error, but the actual operation worked.
* Fixed a bug where users see a generic license error notification on the login screen
* Fixed a bug where SSO logout was not working for the identity providers Ping and Azure
* Fixed bug where service-handover crashed due to broker message double acking.

### Cognigy Insights

**Released** October 18, 2022

#### Improvements

* Disabling scroll bar on print view on other devices than Mac
* Adding grid to the detailed view of line chart, legend bottom and left

#### Bug Fixes

* Fixed a bug where xAxis on lineCharts was jumping with query grain param change after timeframe filter change
* Fixed a bug where detailed view charts height wasn't resizing properly
* Fixed a bug with redirect from insights to cognigy when there is a query string in the URL (redirect from transcript explorer to playbook editor)
* Fixed a bug where only the first 25 steps were loaded for the dropdown list filter

### Cognigy Live Agent

**Released** October 18, 2022

#### Improvements

* Improved by adding timestamps to messages in the conversation transcript

#### Bug Fixes

* Fixed a bug where forwarding whisper assist or assist bot messages as private notes should be private, and the end user should not receive it
* Fixed a bug where teams assignments were not refreshing correctly on live changes


# 4.38 - Happy Halloween 🎃
Source: https://docs.cognigy.com/release-notes/4.38



<Warning>
  With release 4.30.0, we have officially marked our Kubernetes repository as `deprecated`. If you are an `on-premises` customer and you are running Cognigy/Cognigy Insights yourself, have a look at our [Helm Chart](https://github.com/cognigy/cognigy-ai-helm-chart), which we have crafted for Cognigy/Cognigy Insights! Our Kubernetes repository will still receive updates `until 31st December 2022`.
</Warning>

## 4.38.0

### Cognigy

**Released** October 31, 2022

#### Features

##### Amazon Lex Endpoint

This release of Cognigy officially introduces the Amazon Lex Endpoint which allows our customers to integrate with Amazon Lex NLU.

##### Genesys Cloud Handover Provider

This release of Cognigy officially adds support for Genesys Cloud as a handover provider.

#### Improvements

* Improved by adding file upload and attachment handling to service-endpoint for contacts, location and sticker type of whatsapp messages
* Improved by adding new organization wide data privacy settings that can be set in the management-UI
* Improved by displaying the speech providers logo in voice preview dialogue
* Improved by adding a togglable XML editor to the Voice Preview feature
* Improved by showing an additional error message for json editor fields containing invalid json
* Improved by adding a JSON validator for updating Audio Settings and Translate Settings
* Improved by filtering and searching Audit Events using type and user at the same time
* Improved by adding new role "Live Agent Supervisor"

#### Bug Fixes

* Fixed a bug where long script texts overflow in the email content field
* Fixed a bug where the session timeout duration was too long. Reduced the default session timeout duration to 7 days if "Remember Me" check box is selected during login
* Fixed a bug where voice preview was passing the JWT as a query param
* Fixed a bug where the Interaction panel call failed if the optional region field for the Microsoft voice provider was not set
* Fixed a bug where login with SSO throws an error if cognigy client is not configured
* Fixed a bug where sentence types could not be used in rules and conditions
* Fixed a bug where userId schema validation is getting failed due to type change of userId field after JSON parsing it in "getHandoverMetaInRedis"
* Fixed a bug where agent inject did not work with Facebook (Meta)
* Fixed a bug where the TTL for being able to do notify/inject requests would not reset after sending a notify/inject request
* Fixed a bug where Live Follow would not work for the Socket Endpoint
* Fixed a bug where Live Follow would not include the data of the input message
* Fixed a bug where live agent redirection did not work if live-agent env variable is configured

### Cognigy Insights

**Released** October 31, 2022

#### Features

##### Configurable TTL for entire Insights data

This release of Cognigy Insights adds the ability to configure a TTL for all data we collect as part of Cognigy Insights. The TTL can either be configured on a system level or via the Cognigy Management UI

#### Improvements

* Detailed view for Pie Chart graph showing all fractions (and reducing the default to only show the largest 10 fractions)
* Improved by adding a TTL to all analytics services, for document expiration after a predefined time

#### Bug Fixes

* Remove tooltip clipping
* Fixed a bug where sorting does not work on orderOfIntentsByAverageNluScore query
* Fixed a bug where Transcript explore filtering is not working
* Fixed a bug where users with liveAgentSupervisor role were not able to access insights

### Cognigy Live Agent

**Released** October 31, 2022

#### Features

* Add the first part of the supervisor dashboard to the Live Agent UI. This includes the ability to view the current status of all agents, an overview of the conversation handling, account and agent based. The supervisor dashboard is currently only available to users with the `supervisor` role

#### Improvements

* Improved by fixing several Helm chart issues opened on GitHub

  * [https://github.com/Cognigy/cognigy-live-agent-helm-chart/issues/1](https://github.com/Cognigy/cognigy-live-agent-helm-chart/issues/1)
  * [https://github.com/Cognigy/cognigy-live-agent-helm-chart/issues/2](https://github.com/Cognigy/cognigy-live-agent-helm-chart/issues/2)
  * [https://github.com/Cognigy/cognigy-live-agent-helm-chart/issues/3](https://github.com/Cognigy/cognigy-live-agent-helm-chart/issues/3)
  * [https://github.com/Cognigy/cognigy-live-agent-helm-chart/issues/4](https://github.com/Cognigy/cognigy-live-agent-helm-chart/issues/4)
  * [https://github.com/Cognigy/cognigy-live-agent-helm-chart/issues/5](https://github.com/Cognigy/cognigy-live-agent-helm-chart/issues/5)
  * [https://github.com/Cognigy/cognigy-live-agent-helm-chart/issues/6](https://github.com/Cognigy/cognigy-live-agent-helm-chart/issues/6)
  * [https://github.com/Cognigy/cognigy-live-agent-helm-chart/issues/7](https://github.com/Cognigy/cognigy-live-agent-helm-chart/issues/7)
  * [https://github.com/Cognigy/cognigy-live-agent-helm-chart/issues/8](https://github.com/Cognigy/cognigy-live-agent-helm-chart/issues/8)
  * [https://github.com/Cognigy/cognigy-live-agent-helm-chart/issues/9](https://github.com/Cognigy/cognigy-live-agent-helm-chart/issues/9)
    > Note: `values.configmap.SECRET_KEY_BASE` has been replaced by `cookiesIntegrity` to reference a secret. It is important to create a new secret and set the previous value. Otherwise, users will need to do logout and login

* Improved by adding validation for date ranges and restricting queries to one year in the past for the reports section

#### Bug Fixes

* Fixed a bug where the UI was not validating files properly while uploading profile pictures and inbox avatars
* Fixed a bug where the agent role was able to access unauthorized endpoints
* Fixed a bug where the forward arrow in agent assist bot messages was overlapping the messages when hovered in a reduced window
* Fixed a bug where Galleries could not be scrolled when the screen was smaller

### Webchat Widget

You can find the release notes for our Webchat Widget on [Github](https://github.com/Cognigy/WebchatWidget/releases) as usual. The latest release of our Webchat Widget at the time of releasing all other Cognigy products is [v2.42.1](https://github.com/Cognigy/WebchatWidget/releases/tag/v2.42.1).


# 4.39
Source: https://docs.cognigy.com/release-notes/4.39



<Warning>
  With release 4.30.0, we have officially marked our Kubernetes repository as `deprecated`. If you are an `on-premises` customer and you are running Cognigy/Cognigy Insights yourself, have a look at our [Helm Chart](https://github.com/cognigy/cognigy-ai-helm-chart), which we have crafted for Cognigy/Cognigy Insights! Our Kubernetes repository will still receive updates `until 31st December 2022`.
</Warning>

## 4.39.1

### Cognigy

**Released** November 18, 2022

#### Bug Fixes

* Fixed a bug where service-session-state-manager would stop processing incoming requests if a database write failure was encountered during a bulk update operation. The error is now logged, but processing continues normally
* Fixed a memory leak in service-ai that was related to incorrectly caching Flows in-memory
* Prevent sending the entire conversation history to Cognigy Live Agent when entering a handover multiple times in a row

### Cognigy Live Agent

**Released** November 18, 2022

#### Bug Fixes

* Fixed a bug where the translations in the success message for updating agents inside an inbox was not displayed
* Fixed a bug where previous conversations were not refreshed in realtime

## 4.39.0

### Cognigy

**Released** November 15, 2022

#### Improvements

* Improved by making image attachments sent from live agent downloadable
* Improved by updating the Webchat Widget to version 2.43.0
* Improved by enabling the Inject API for REST Endpoints
* Improved by creating a new wizard for agent assist.
* Improved by adding the feature to clone a flow and all its sub-resources
* Improved by deleting the insights data during delete organization
* Improved by adding Test Translation Configuration endpoint with a Test Connection button

#### Bug Fixes

* Fixed a bug where the layout in the interaction panel is broken when "show data" input toggle is enabled
* Fixed security vulnerabilities in service-api by updating socket.io
* Fixed a bug where Login shows a 'network error' when trying to log in with an invalid password
* Fixed a bug where lexicon entries are not fetched with accept type application/hal+json
* Fixed a bug where users could keep trying to log in just by changing their original IP
* Fixed a bug where handovers with Live Agent would not show the messages from previous agents correctly.
* Fixed a bug where a Cognigy script condition in a button inside a default reply would have no effect.
* Fixed a bug where custom properties which were added to the input object are removed when moving between Flows using executeFlow Nodes
* Fixed a bug where the intent level would not be defined when using an Execute Flow Node
* Fixed a bug where the intent level would not be defined when using Attached Flows
* Fixed a bug where introducing an Overwrite Analytics node to a flow would prevent sessions from being accessible through the OData endpoint
* Fixed a bug where the "overwrite" setting of the intent upload feature would not completely replace existing intents and their settings
* Fixed a bug where intent entrypoints would not work when using Attached Flows
* Fixed a bug where some (Belgian) French number slots would not be recognized.
* Fixed a bug where German money slots would sometimes be recognized as date slots.
* Fixed a bug where German number slots would sometimes be recognized as date slots.
* Fixed a bug where German date slots would sometimes be wrong or not recognized.
* Fixed a bug where intents would not be predicted due to wrong data being forwarded to the runtime

### Cognigy Insights

**Released** November 15, 2022

#### Features

##### Selectable grain for detail view

This release of Cognigy Insights adds the ability to select a grain for detail views for some of our reports on the quantitive dashboards

#### Improvements

* Adding a metric that shows if a service is connected to dependencies such as RabbitMQ & MongoDB
* Adding OData Endpoint for New Live Agent Collections
* Added a new screenshot and documentation for the analytics, insight's session and steps expiration values

#### Bug Fixes

* Fixed a bug where the Conversations line chart in detailed view was crashing on the "Today" time filter
* Fixed a bug where a custom scroll bar didn't work
* Fixed a bug with Insights / OData: 2.3 authentication failures
* Fixed a bug with adding correct index on analytics collection to support update analytics records

### Cognigy Live Agent

**Released** November 15, 2022

#### Features

* Add the supervisor dashboard graph and global inbox filtering in the overview dashboard

#### Improvements

* Improved by adding a button to scroll down when you scroll up to the messages and avoids auto scrolling when new messages are coming
* Improved by adding default alphabetical sorting for inboxes across live agent

#### Bug Fixes

* Fixed a bug where conversation filters were broken

### Webchat Widget

You can find the release notes for our Webchat Widget on [Github](https://github.com/Cognigy/WebchatWidget/releases) as usual. The latest release of our Webchat Widget at the time of releasing all other Cognigy products is [v2.43.0](https://github.com/Cognigy/WebchatWidget/releases/tag/v2.43.0).


# 4.40
Source: https://docs.cognigy.com/release-notes/4.40



<Warning>
  With release 4.30.0, we have officially marked our Kubernetes repository as `deprecated`. If you are an `on-premises` customer and you are running Cognigy/Cognigy Insights yourself, have a look at our [Helm Chart](https://github.com/cognigy/cognigy-ai-helm-chart), which we have crafted for Cognigy/Cognigy Insights! Our Kubernetes repository will still receive updates `until 31st December 2022`.
</Warning>

## 4.40.3

### Cognigy

**Released** December 5, 2022

#### Bug Fixes

* Reverted NodeJS upgrade for three additional microservices sending outbound HTTP calls as the newer NodeJS version can break compatibility with certain APIs

## 4.40.2

### Cognigy

**Released** November 30, 2022

#### Bug Fixes

* Reverted NodeJS upgrade for two microservices sending outbound HTTP calls as the newer NodeJS version could break compatibility with old APIs which still use old TLS versions (e.g. TLS v1.0)
* Fixed a problem related to Voice Activity Detection always being active

## 4.40.1

### Cognigy

**Released** November 29, 2022

#### Bug Fixes

* Fixed a bug where internal state could be reset in certain situations which leads to e.g. the Context being reset in the middle of conversations

## 4.40.0

### Cognigy

**Released** November 29, 2022

#### Improvements

* Improved the UX of the Voice Preview Dialog
* Improved by updating the Webchat Widget to latest version (v2.44.0)
* Improved by allowing to navigate Positive/Negative confirmation words using Enter key
* Improved by adjusting the generic Voice Flow Nodes "Barge In" and "Play"

#### Bug Fixes

* Fixed a bug where when using a Think node in a flow with Complete Goal nodes defined, these goals were not available in the OData Sessions endpoint.
* Fixed a bug where an Execute Cognigy NLU node would not trigger an intent's default reply. A new option Process Intent Default Reply has been added to the Execute Cognigy NLU node. This option is disabled by default, which means the default reply is not processed. However, when it is enabled, the default reply is processed in the same manner as when triggered by means other than the Execute Cognigy NLU node.
* Fixed a bug where execution count for realtime translation with DeepL did not work properly for English and Portuguese.
* Fixed a bug where English money slots would sometimes be recognized as date slots.
* Fixed a bug where service-parser would crash when importing Keyphrases with an empty data object.
* Fixed a bug where CognigyScript in lookup Nodes would log errors.
* Fixed a bug where the media file attached from WhatsApp endpoint is not visible in the interaction panel during a live-follow
* Fixed a bug where voice name and language gets reset after audio provider is changed

### Cognigy Insights

**Released** November 29, 2022

#### Improvements

* Changed Step Explorer visualization in order to support larger amounts of data during visualization phase

#### Bug Fixes

* Fixed a bug where live agent escalations records would not get inserted because of corresponding session information
* Fixed a bug where Transcript Explorer would show an incorrect number of sessions whenever the channel name differs for the same sessionID
* Fixed a bug where the conversations line chart in detailed view was crashing on the "Today" time filter

### Cognigy Live Agent

**Released** November 29, 2022

#### Improvements

* Improved by adding a raw data modal for the whisper assist and assist bot forwardable messages - allowing our customers to inspect the data they will forward for all channels Cognigy supports

#### Bug Fixes

* Fixed a bug where the email transcript was missing descriptions for audio, video, image and list messages
* Fixed a bug where the translations in the success message for updating agents inside an inbox was not displayed correctly
* Fixed a bug where navigating to a nonexistent inbox in inbox settings will let you edit it, even if it does not exist. Resulting in an error when updating

### Webchat Widget

You can find the release notes for our Webchat Widget on [GitHub](https://github.com/Cognigy/WebchatWidget/releases) as usual. The latest release of our Webchat Widget at the time of releasing all other Cognigy products is [v2.44.0](https://github.com/Cognigy/WebchatWidget/releases/tag/v2.44.0).

### Infrastructure (AI + IN)

#### Stateful services

To improve the functionality and fix known security vulnerabilities in the dependencies of Cognigy/Cognigy Insights uses, we have updated the following dependencies:

* RabbitMQ, from `v3.9.20` to `v3.9.24`

**Important: Once you have upgraded to the new release containing the new RabbitMQ version, run the following command to rollout (restart) all Pods that run in your cluster - be sure to replace the `<target_namespace>` accordingly:**

```bash theme={null}
for i in $(kubectl get deployment --namespace <target_namespace> --template '{{range .items}}{{.metadata.name}}{{"\n"}}{{end}}'|grep service-)
do
kubectl --namespace <target_namespace> rollout restart deployment $i
done
```

#### New service

This release also adds a new service called `service-collector`. This microservice will be used in the neare future to collect analytics information for Cognigy Insights so we can also ingest data from other Cognigy products such as Cognigy Live Agent and Cognigy Voice Gateway.

**Important: In case you have overwritten the replica-count (or other aspects) for specific services in your own `values` file, make sure that you also apply the same configuration for the new `serviceCollector` section in your `values`.**

#### Config change

In case you are already using the `call capabilities within the Cognigy Interaction Panel`, you have to make sure that you change the following ENV variable in the `cognigyEnv` configmap section:

```text theme={null}
VOICE_GATEWAY_PREPARE_CALL_API
```

from value: `/api/v1/call/prepare` to value: `/api/v2/call/prepare`.

In addition, you have to set the following ENV variable so `Cognigy Live Agent` will be able to display attachments that have been uploaded through Cognigy endpoints:

```text theme={null}
RUNTIME_FILE_MANAGER_CORS_WHITELIST
```

the value should be at least set to the `frontend DNS name` of the `Cognigy Live Agent` installation - an example value would be: `https://dev-live-agent.cognigy.ai`


# 4.41 - Merry X-Mas 🎅🏻🎄
Source: https://docs.cognigy.com/release-notes/4.41



<Warning>
  With release 4.30.0, we have officially marked our Kubernetes repository as `deprecated`. If you are an `on-premises` customer and you are running Cognigy/Cognigy Insights yourself, have a look at our [Helm Chart](https://github.com/cognigy/cognigy-ai-helm-chart), which we have crafted for Cognigy/Cognigy Insights! Our Kubernetes repository will still receive updates `until 31st December 2022`.
</Warning>

## 4.41.2

### Cognigy

**Released** December 23, 2022

#### Bug Fixes

* Fixed a bug where the metric showing message processing times did not properly work in the Cognigy monitoring solution

## 4.41.1

### Cognigy

**Released** December 22, 2022

#### Improvements

* Improved by handling "user disconnect" event for Ring Central Engage handover provider. This event will be handled for supporting endpoints like our Webchat Widget

## 4.41.0

### Cognigy

**Released** December 13, 2022

#### Improvements

* Improved by adding "Copy JSON Path" & "Create Token from JSON path" context menu options to JSON editor fields - for example within the Input Object inspector in the Interaction Panel
* Improved by updating the Webchat version to version 2.45.0
* Improved by adding a new Node field type "adaptivecard" to the Extension Framework
* Improved the date-picker in Cognigy Management-UI to be more consistent across different web browsers
* Improved by releasing a new version of our rest-api-client package to npm (v0.13.1)
* Improved layout by adding a header to Cognigy Live Agent setup section in Agent Creation Wizard
* Improved by renaming Whisper Assist to Agent Assist for the Cognigy Live Agent Handover Provider

#### Bug Fixes

* Fixed a bug where the reparseSlots option would not work for the Amazon Lex Endpoint
* Fixed a bug where invalid Slot annotations would prevent downloading Intents
* Fixed a bug where case values with a trailing new line could not be triggered in Flows
* Fixed a bug where sessionId is undefined in input/output/inject/notify Transformers on inject/notify API calls
* Fixed a bug where phone number quick reply button is not working as expected
* Fixed a bug where the response of an HTTP request would not be stored in the input/context in case of an error

### Cognigy Insights

**Released** December 13, 2022

#### Improvements

* Changed the data foundation we are using to drive the "Conversation Count" and "Conversations linechart" report and make sure they are based on session data to make those reports more accurate
* Renamed "Average Conversation length" and "Max Conversation Length" to "Average Session Length" and "Max Session Length" as the reports were actually based on session data
* Improved the performance of certain analytics reports and our OData analytics interface by adding missing database indices

#### Bug Fixes

* Fixed a bug where the "Users by Locale" report on the overview dashboard did not indicate a timeframe in the tooltop of the detailed view
* Fixed a bug where the timestamp filter was calculated incorrectly
* Fixed a bug where the "percentage bar-chart tooltip" was showing a timestamp
* Fixed a bug where duration was formatted wrongly in multiline charts
* Fixed a bug where the Contact ID was hashed multiple times in certain situations. This led to problems where certain data could not be retrieved from the system as we had a mismatch in the given Contact ID and the hashed value stored in our data stores

### Cognigy Live Agent

**Released** December 13, 2022

#### Features

##### Searchable Attachments

This release of Cognigy Live Agent extends the Attachments section on the Account level by providing additional search capabilities so attachments can be search by "Cognigy session id", "Cognigy user id" and filename

#### Bug Fixes

* Fixed a bug where the "Supervisor Overview Graph" would show incorrect data in certain conditions
* Fixed a bug where no error feedback was given when uploading a file with a non-supported file format
* Fixed a bug where messages could appear twice in the conversations

### Webchat Widget

You can find the release notes for our Webchat Widget on [Github](https://github.com/Cognigy/WebchatWidget/releases) as usual. The latest release of our Webchat Widget at the time of releasing all other Cognigy products is [v2.45.0](https://github.com/Cognigy/WebchatWidget/releases/tag/v2.45.0).

### Infrastructure (AI + IN)

#### New secret

We have introduced a new Kubernetes secret which is mounted into the `service-endpoint` Deployment. You don't have to do anything as the Cognigy Helm Chart will automatically create the secret in your respective Kubernetes namespace and provide a random value for it.


# 4.42 - Happy New Year ✨
Source: https://docs.cognigy.com/release-notes/4.42



<Warning>
  Live Agent Lite, also known as Livechat, is an outdated product that is no longer receiving updates or support. As such, all references to this product will be removed in future versions of Cognigy. We strongly recommend upgrading to Cognigy Live Agent, the successor of Live Agent Lite.
</Warning>

<Warning>
  The Cognigy OData interfaces 2.0 and 2.2 have been deprecated. These interfaces will be removed in Cognigy 4.48.0. Migrate to OData 2.3 before the removal date.
</Warning>

## 4.42.1

### Cognigy

**Released** January 19, 2023

#### Bug Fixes

* Fixed a bug where Salesforce Handover functionality would sometimes not work properly due to a problem in structured content conversion (e.g. galleries)
* Fixed a security vulnerability in service-api
* Fixed a security vulnerability in service-endpoint
* Fixed a security vulnerability in insights-api (this is an API-server only fix)

## 4.42.0

### Cognigy

**Released** January 11, 2023

#### Improvements

* Added the opportunity to handle the "user disconnect" event for the RCE handover provider. This event will be handled for supporting endpoints like Webchat with the RCE handover provider.
* Extended the "Handover to Agent" Flow Node with the capability to pass "additional category IDs" to RingCentral Engage
* Added the `api-ban` service to block calls from prohibited IP addresses
* Added a warning when a Transfer Node is called from the Interaction Panel
* Added the capability to receive `conversationcounter` via Open API for the `voice-gateway2` channel
* Improved the Set Session Config Node by adding advanced configuration for Azure TTS and STT

#### Bug Fixes

* Removed the required reloading of a page with lexicon attachments after switching locales
* Fixed the service-resources error when the service is shut down in the 3rd Runtime entry
* Fixed the incorrect number of handover requests in Insights
* Fixed the behavior of the Set Session Config Node, and its values are now correctly updated
* Fixed the issue with the Play Node crashing when configuring the Set Activity Parameters

### Cognigy Insights

**Released** January 11, 2023

#### Improvements

* Added a [Data Expiration indicator](/insights/data-management/data-expiration) in the upper-right corner of the Insights interface

#### Bug Fixes

* Removed restrictions on displaying a limited number of agents in the Agent drop-down list in the upper-left corner of the Insights interface
* Removed restrictions on displaying a limited number of flows in the Flow drop-down list on the Transcript Explorer page

### Cognigy Live Agent

**Released** January 11, 2023

#### Bug Fixes

* Fixed time codes overlapping conversation messages in the Previous Conversations menu
* Added a customer name in the conversation transcript sent by email

### Cognigy Voice Gateway

**Released** January 11, 2023

#### Improvements

* Added the `api-ban` service to block calls from prohibited IP addresses
* Added a warning when a Transfer Node is called from the Interaction Panel
* Added the capability to receive `conversationcounter` via Open API for the `voice-gateway2` channel
* Improved the Set Session Config Node by adding advanced configuration for Azure TTS and STT

#### Bug Fixes

* Fixed the behavior of the Set Session Config Node, and its values are now correctly updated
* Fixed the issue with the Play Node crashing when configuring the Set Activity Parameters

### Webchat Widget

You can find the release notes for Webchat Widget on [GitHub](https://github.com/Cognigy/WebchatWidget/releases) as usual. The latest release of our Webchat Widget at the time of releasing all other Cognigy products is [v2.46.0](https://github.com/Cognigy/WebchatWidget/releases/tag/v2.46.0).

### Infrastructure (AI + IN)

#### Horizontal Pod Autoscaling

We have introduced additional HPA objects for additional microservices. Don't use them in production, yet.


# 4.43 - Happy National Pie Day 🥧
Source: https://docs.cognigy.com/release-notes/4.43



*New updates, fresh out of the oven - our release is hot and ready to serve!*

## 4.43.1

### Cognigy Insights

**Released** January 25, 2023

#### Bug Fixes

* Fixed an issue in the Transcript Explorer were input data would not be displayed properly

### Cognigy Live Agent

**Released** January 25, 2023

#### Bug Fixes

* Fixed the issue where the Helm Chart OData service configmap pointed to the wrong values for the PostgreSQL dependency API, which caused an error while upgrading to version 4.43.0

## 4.43.0

### Cognigy

**Released** January 23, 2023

#### Features

* Once an Intent is found with high confidence, the mapping between it and its child Intents may still be ambiguous. The [Intent Mapping Priority](/ai/platform-features/nlu/intents/intent-hierarchy/intent-mapping-priority) functionality allows users to specify whether it should be more likely that the parent or child Intent is predicted.

#### Bug Fixes

* Fixed the issue in the Cognigy NLP scoring microservices, where compute-intensive tasks in specific API calls would block the health check server from responding to the Kubernetes livenessProbe. That would lead to container restarts
* Fixed the issue in reading profile while sending a message to the RCE handover provider with the "Forward Only Handover Conversations" toggle in handover provider settings is turned off
* Fixed the issue when data was not returned during the reconfirmation process of a Question Node, if no negative word was used and the Skip if Answer in Input setting was enabled
* Fixed the date format from `month/day/year` to `day/month/year` for the Polish locale
* Fixed the mismatch between the First/Last Name fields for contact attachments and the ones defined in the code structure

### Cognigy Insights

**Released** January 23, 2023

#### Improvements

* Added WhatsApp attachments in the chat for Transcript Explorer and Message Explorer
* Added the Goals filter in the Additional filters section for Transcript Explorer

#### Bug Fixes

* Fixed the inconsistent results while using a different set of filters in Transcript Explorer

### Cognigy Live Agent

**Released** January 23, 2023

#### Improvements

* Added an opportunity to preview a message with raw data in the chat

#### Bug Fixes

* Fixed the issue related to the second handover between bot and human agent in a single chat session
* Removed the empty Goals field from the Edit Contact menu when the contact is updated
* Fixed the display of First Name and Last Name in the Contact Details attachment sent from WhatsApp to the Live Agent chat

### Cognigy Voice Gateway

**Released** January 23, 2023

#### Improvements

* Added a voice warning message for Interaction Panel Calls when using a Transfer Node

#### Bug Fixes

* Fixed non-functioning automatic translation in the voice-gateway Channel
* Fixed the Barge-in Config Node behavior that caused interruption during a call

### Webchat Widget

You can find the release notes for our Webchat Widget on [GitHub](https://github.com/Cognigy/WebchatWidget/releases) as usual. The latest release of our Webchat Widget at the time of releasing all other Cognigy products is [v2.47.0](https://github.com/Cognigy/WebchatWidget/releases/tag/v2.47.0).


# 4.44 - Happy Get Up National Day 🧗 🌟
Source: https://docs.cognigy.com/release-notes/4.44



*Reach for the Stars: Embrace the spirit of Get Up National Day and rise to new AI heights with our release!*

## 4.44.1

### Cognigy

**Released** February 9, 2023

#### Bug Fixes

* Fixed a bug where Question Node values could not be updated in certain situations

## 4.44.0

### Cognigy

**Released** February 1, 2023

#### Features

* Added the [Adaptive Card](/ai/agents/develop/node-reference/basic/say) type to the Output Type list for the Say and Question Nodes
* Added [Generative AI](/ai/agents/develop/gen-ai-and-llms/generative-ai) (available only for Beta program participants)

#### Improvements

* Added the option to create descriptions as optional property to Flows, Lexicons, and Intents

#### Bug Fixes

* Fixed duplication of messages that were displayed in Genesys Cloud handover provider when running multiple service-handover replicas
* Removed content transformation to Markdown in Live Agent handover client
* Updated the description for the Chart Node API documentation that did not have a proper example and model
* Fixed duplicates in Yes/No sentences
* Fixed the issue where UI was requesting snapshots resources with a projectId
* Fixed the issue where a `validation failed` error was logged when chatting from the Interaction panel for the first time after refreshing the UI
* Fixed the issue where `onlyAcceptEscalationIntents` in Question Nodes would not work correctly with attached Flows

### Cognigy Insights

**Released** February 1, 2023

#### Improvements

* Created print styles for detailed views of charts for the Google Chrome browser
* Added animation on changing top tiles in dashboards
* Replaced tile "Max. Session Length" with "Containment rate" in the Engagement dashboard

#### Bug Fixes

* Fixed the Y-Axis margin-left for the Conversation Summary chart (Live Agent dashboard) and Number of Inputs chart (Overview dashboard)
* Fixed rendering inputData in Transcript Explorer

### Cognigy Live Agent

**Released** February 1, 2023

#### Features

* Added the opportunity to create [multiple roles](/live-agent/roles) for users

#### Improvements

* Added the opportunity to view Location messages with full information from WhatsApp
* Improved the synchronization between Live Agent and Insights for sending conversation events based on the `cognigy_project_id` instead of the Inbox one
* Added the description about the Supervisor role in the Settings > Agents page
* Upgraded PostgreSQL and Redis dependencies to the latest version in the Helm Chart (does not apply to external instances). For more information, see [migration guide](/live-agent/installation/migration/helm-dependencies)

#### Bug Fixes

* Fixed the issue when text bubbles with newline characters are not visualized correctly in the chat on the Agent side
* Fixed the issue when Agents with the Supervisor role and their customers couldn't view sending and receiving messages

### Cognigy Voice Gateway

**Released** February 1, 2023

#### Features

* Added the [Dial Transfer Type](/ai/agents/develop/node-reference/voice/voice-gateway/transfer) to the Voice Gateway Transfer Node. If you want to use this update, add a new Voice Gateway Transfer Node in the Flow Editor and manually transfer the required settings from the old Node.

#### Improvements

* Added the Reason field to the Voice Gateway Transfer Node

#### Bug Fixes

* Fixed the issue where STT Hints were being applied globally when using Set Activity Parameters

### Webchat Widget

You can find the release notes for our Webchat Widget on [GitHub](https://github.com/Cognigy/WebchatWidget/releases) as usual. The latest release of our Webchat Widget at the time of releasing all other Cognigy products is [v2.48.0](https://github.com/Cognigy/WebchatWidget/releases/tag/v2.48.0).

### Infrastructure (AI + IN)

#### Improvements

* Added the option to use predefined `storageClass` for flow-modules, function, and redis-persistent
* Added the option to define custom value for the PVC size of flow-modules, function, and redis-persistentAdded option to define custom value for the PVC size of flow-modules, function and redis-persistent
* Added the option to choose whether to deploy or not deploy the EFS provisioner for the AWS cloud provider

#### Deprecation warning

This is applicable only if you are using AWS EFS for flow-modules and functions. Previously you could define the AWS EFS IDs as

```yaml theme={null}
efs:
  flowModules:
    id: ""
  functions:
    id: ""
```

The above-mentioned format of defining EFS IDs will be deprecated in the future release of the helm chart. If you have defined the EFS IDs in the above-mentioned format, then migrate to the new format

```yaml theme={null}
flowModules:
  persistence:
    aws:
      efs:
        id: ""

functions:
  persistence:
    aws:
      efs:
        id: ""
```


# 4.45 - Happy Family National Day 👨‍👩‍👧‍👦
Source: https://docs.cognigy.com/release-notes/4.45



*Just like families, our Cognigy team is built on strength and support - Happy National Family Day from us and our latest release!*

## 4.45.2

### Cognigy Live Agent

#### Bug Fixes

* Fixed the issue where the OData service was generating the `Cannot convert literal due to expected type Edm.Date` error during query execution

## 4.45.1

### Cognigy Insights

**Released** March 6, 2023

#### Bug Fixes

* Ensure that masked-data string (\*\*\*\*) will only be rendered if no other data needs to be visualized for a message in the Transcript- and Message-Explorer
* Ensure that masked-data string (\*\_\*\*) won't be rendered as an object with individual "\_" fields

### Cognigy Live Agent

**Released** March 3, 2023

#### Bug Fixes

* Fixed the issue where the service was generating a 503 error during query execution

## 4.45.0

### Cognigy

**Released** February 20, 2023

#### Features

* [Generative AI Beta](/ai/agents/develop/gen-ai-and-llms/generative-ai) is publicly available for all Cognigy users
* Added the opportunity that allows Webchat users to upload attachments easily. Users can drag and drop the file or click the new attachment symbol. The functionality is configurable from the [Webchat Endpoint](/ai/agents/deploy/endpoints/file-storage) settings

#### Improvements

* Marked the [user input as ignored](/ai/agents/develop/ai-agent-memory/input#properties) so that they are neither counted as understood nor misunderstood in Analytics
* Added formatting for code in the Code Nodes, which are created via Generative AI Flow
* Added the opportunity to download an Intent Trainer export package from the Task Menu

#### Bug Fixes

* Set the [understood property](/ai/for-developers/code/analytics-data) in Analytics to Null for messages in an active handover
* Fixed the issue where userId is getting lost from the messages in the RingCentral Engage chat and is displayed with the AI Agent name. Now the AI Agent and user messages are properly distinguished in the chat
* Fixed the error which caused issues with the functioning of the Agent Availability Node
* Fixed the issue where events are sent to the Live Agent when analytics masking was activated
* Fixed the issue where time Slots were incorrect for the pattern `from <hour> until <hour>` in English and German
* Fixed the issue where the first message of a session with sensitive data is not masked in the Logs
* Fixed the issue where all user data are being masked with `PERSON_NAME` in the Intent Trainer. Now only `creditCardNumber` will be masked

### Cognigy Insights

**Released** February 20, 2023

#### Improvements

* Changed the calculation of the ratio of understood messages. Now the `null` values don't consider in the calculation. This change affects charts, including [NLU performance](/insights/reports/nlu-performance)

#### Bug Fixes

* Fixed the issue where the chart's detailed view was not displaying correctly in print mode on web browsers
* Fixed the issue where Webchat attachments were rendered as unknown attachment types in the Message and Transcript Explorers
* Fixed the issue where getting indexConversations via API would sometime timeout for a large dataset

### Cognigy Live Agent

**Released** February 20, 2023

#### Features

* Moved the push and email notifications from [Profile Settings](/live-agent/profile-settings#notifications) to [Account Settings](/live-agent/settings/account-settings#notifications) and can only be controlled by administrators. If the administrator has not yet changed the push and email notification settings, a human agent still has the opportunity to change them in the Profile Settings
* Added the opportunity for human agents to [create hyperlinks](/live-agent/conversation/send-reply#send-replies-to-users) and send the hyperlinked text in the conversation chat

#### Improvements

* Improved the design of messages sent by a human agent in the conversation chat

### Cognigy Voice Gateway

**Released** February 20, 2023

#### Features

* Added the [Call Recording](/ai/agents/develop/node-reference/voice/voice-gateway/call-recording) Node
* Added the generic [Send Metadata](/ai/agents/develop/node-reference/voice/generic/send-metadata) Node

#### Improvements

* Added a maximum limit of 999 retries to the No user input retries field in the [User Input Timeout - Config](/ai/agents/develop/node-reference/voice/generic/user-input-timeout-config) Node
* Added the opportunity to process [call completion](/voice-gateway/references/events/overview) events in the Flow

#### Bug Fixes

* Fixed missing the .wav format support in the [Transfer Node](/ai/agents/develop/node-reference/voice/voice-gateway/transfer) (Dial type > Dial Music)

### Infrastructure (AI + IN)

#### Redis v7

In this release, we are officially updating the version of Redis we use to Redis 7, which will impact the following two deployments in the Helm Chart:

* statefulRedis
* statefulRedisPersistent

Both Redis instances will restart during the upgrade, and the application might not be available for a few minutes.

<Warning>
  If you are not using the built-in Redis deployments but run Redis outside your Cognigy Helm Chart, make sure to start migrating to Redis 7. Support for Redis 5 has ended, and you will no longer receive any security patches.
</Warning>


# 4.46 - Happy National Be Heard Day 🔊🤝
Source: https://docs.cognigy.com/release-notes/4.46



*Customer-centric communication made easy with Cognigy Conversational AI: Celebrating National Be Heard Day with our updates that help you connect with your customers in a way that builds trust and fosters loyalty!*

## 4.46.1

### Cognigy Live Agent

**Released** April 4, 2023

#### Improvements

* Improved performance by fixing spikes that occurred while querying user roles

## 4.46.0

### Cognigy

**Released** March 7, 2023

#### Features

* Added [Yes/No Intents](/ai/platform-features/nlu/intents/yes-no-intents) in Beta, allowing to detect Yes/No using a machine learning based Intent model. This feature is available to all Cognigy users

#### Improvements

* Added an API for tracking analytics in the Code Node
* Renamed the Complete Text Node to the [GPT Prompt](/ai/agents/develop/node-reference/service/llm-prompt) Node
* Added an error notification in UI if the GPT Prompt Node fails
* Changed the behavior of full-screen mode. Now the expanded code editor no longer spans the entire screen but only the browser window
* Added the option to [trust an Extension](/ai/for-developers/extensions#make-extensions-trusted) via UI
* Improved by trusting the extension after the update if it is already trusted
* Improved by rendering xApps submit payloads in the Live Follow session of the Interaction panel and Insights UI. To configure that, go to [Set HTML xApp State](/ai/agents/develop/node-reference/xApp/set-html-xApp-state#description) or [Set AdaptiveCard xApp state](/ai/agents/develop/node-reference/xApp/set-AdaptiveCard-xApp-state#description) Nodes

#### Bug Fixes

* Fixed the issue where users were getting an error on login because `organisationId` was included in the URL
* Fixed the issue where the `amazon-login` window did not close after successful login
* Fixed the issue where extensions using the newest extension tools might fail installation
* Fixed the issue where the reset password link was not working for users with Single Organization Access when `organisationId` is not correct
* Fixed the issue where the `LocaleId` parameter was not included in the body request of the patch flow endpoint in the OpenAPI documentation
* Changed the description on the 8x8 Endpoint card. The article was added: `Platform Connect your Agent to 8x8 Platform` → `Connect your Agent to the 8x8 Platform`.
* Fixed the issue where sometimes context could be lost during a conversation
* Fixed the issue where the `positiveAnswer` attribute would appear in `input.nlu.intentMapperResults`, even though it was not used. This attribute has now been removed from `input.nlu.intentMapperResults`
* Fixed the issue where `userId` is getting lost from the messages in RCE and is displayed with the "Bot", instead of the "Bot", and user messages should appear distinct in the RCE agent chat window
* Fixed the issue where the fullscreen mode for the JSON Code editor did not properly work for Lexicon data
* Disabled the live logging when Cognigy gets an error for an invalid cursor value

### Cognigy Insights

**Released** March 7, 2023

#### Improvements

* Added an error notification when the CSV report couldn't be downloaded
* Rendered the masked data strings only when there is no other content to show in messages in Transcript Explorer and Message Explorer
* Changed the naming of Conversation Details and Conversation Id to Session Details and Session Id in Conversation Details of Transcript Explorer

#### Bug Fixes

* Fixed the issue where goals were not shown in the Session details of the Transcript Explorer
* Fixed the issue where WhatsApp media attachments didn't have a caption text
* Fixed the issue where filtering by rating was crashing Message Explorer
* Fixed the issue where the CSV report was unavailable to load due to the wrong API call
* Fixed the issue where HTML was rendered as a plain string in the Transcript and Message Explorers messages
* Fixed the issue where masked message data were rendering like an object with `"\*"` fields

### Cognigy Live Agent

**Released** March 7, 2023

#### Features

* Attachments received from Webchat are treated similarly to those received from WhatsApp
* Supervisor can now [override the status of any Agent](/live-agent/reports/supervisor-overview)

#### Improvements

* Upgraded the OpenSSL version to fix security vulnerabilities

#### Bug Fixes

* Fixed the text from "Webchat Endpoint" to "LA Assist Bot Endpoint" in the Create new Assist Bot section to accurately reflect the endpoint being referenced
* Fixed the issue where files uploaded through Webchat or WhatsApp were not displaying correctly in the conversation
* Fixed the issue where the handover to Live Agent was failing when the Cognigy Session ID was longer than 44 characters

### Cognigy Voice Gateway

**Released** March 7, 2023

#### Features

* Added the Nuance speech provider to the Voice Gateway web application and to Cognigy Voice Nodes
* Removed one second of a post-dial delay call from the Voice Gateway client. That allows answering calls faster

#### Improvements

* Updated speech provider voice lists for VG and AudioCodes endpoints

#### Bug Fixes

* Fixed the Session Config node Barge-In feature where it only allows users to barge in once
* Fixed the behavior for the AudioCodes Call Recording Node when the Node performed pause/resume/stop operations without sending a payload

### Webchat Widget

You can find the release notes for our Webchat Widget on [GitHub](https://github.com/Cognigy/WebchatWidget/releases) as usual. The latest release of our Webchat Widget at the time of releasing all other Cognigy products is [v2.49.0](https://github.com/Cognigy/WebchatWidget/releases/tag/v2.49.0).

### Infrastructure

#### Cognigy + Insights

##### Kubernetes version

In this release, Cognigy Helm Chart officially supports Kubernetes 1.24, which is the version most cloud providers offer for fully managed Kubernetes clusters.

##### Redis memory eviction policy

In this release, `maxmemory` and `maxmemory-policy` have been made configurable. The default value for `maxmemory` is `512mb`, and `maxmemory-policy` is `volatile-ttl`.

Redis instances will restart during the upgrade, and the application might not be available for a few minutes.

#### Voice Gateway

In this release, we added support to use static public IP on the SBC-RTP nodes in an EKS cluster.

We have replaced the `spec.values.sbc.ec2EipAllocator.eipGroupRole` with `spec.values.sbc.ec2EipAllocator.sipEipGroupRole`. If you have defined the tags of EIP for `voip_nodes` in `spec.values.sbc.ec2EipAllocator.eipGroupRole`, then use the `spec.values.sbc.ec2EipAllocator.sipEipGroupRole` instead. We have also introduced a new var `spec.values.sbc.ec2EipAllocator.rtpEipGroupRole`.

For more information, read [AWS EKS Cluster Setup](https://github.com/Cognigy/cognigy-vg-helm-chart/blob/main/docs/003-aws-eks-cluster-setup.md) on GitHub.


# 4.47 - Happy Math Day 🧮
Source: https://docs.cognigy.com/release-notes/4.47



## 4.47.1

### Cognigy

**Released** March 28, 2023

#### Bug Fixes

* Fixed the issue where audio at the beginning of calls from the Interaction Panel was clipped

### Cognigy Voice Gateway

**Released** March 28, 2023

#### Bug Fixes

* Fixed the issue where audio at the beginning of calls from the Interaction Panel was clipped

## 4.47.0

### Cognigy

**Released** March 23, 2023

#### Improvements

* Added an Upload Progress Dialog to the [Snapshot Uploading process](/ai/agents/deploy/snapshots#working-with-snapshots)
* Added preventing unnecessary builds when the Yes/No logic settings are changed
* Added detected Yes/No Intent name and score to the [Expert mode](/ai/agents/test/interaction-panel/overview) in the Interaction panel
* Added a disclaimer to the Yes/No Intent editor in Manage > Localization: `Please enable Yes / No Intents in your Project / Flow Settings to use them`
* Added using the last [conversation entries](/ai/agents/develop/node-reference/service/llm-prompt) in the GPT Prompt Node. To do that, specify `@cognigyRecentConversation` or `@cognigyRecentUserInputs` in the Prompt field
* Renamed xApp user-facing labels from `app/apps` to `xApp/xApps`, and the names of the [xApp Nodes](/ai/agents/develop/node-reference/xApp/overview) was also changed
* Added support for Single-Tenant Teams bots in the Azure Bot Services Endpoint. To learn more, read [Azure Bot Services: Advanced Deployment with SSO (Microsoft Teams)](https://support.cognigy.com/hc/en-us/articles/6417377133596)

#### Bug Fixes

* Fixed the issue where Cognigy did not save the Intercom Endpoint settings
* Fixed the issue where the Copy JSON path feature from JSON editors added an unnecessary dot in the path notation
* Fixed the issue where conversation handover is not resolved, and AI Agent messages are sent twice when the handover provider and endpoint channel are both 8x8
* Fixed the issue where exact matching could fail if a user input matches an example sentence in an Intent. This issue occurred when the Intent contains an annotated lexicon keyphrase with whitespace or punctuation characters
* Fixed the issue where error message details were not displayed correctly in the UI when exporting/importing an Intent
* Fixed the issue where German date ranges in the `dd.mm.yy` format were not be recognized

### Cognigy Insights

**Released** March 23, 2023

#### Bug Fixes

* Renamed the Contacts line chart to the [Unique Contacts](/insights/reports/engagement#unique-contacts) chart
* Added Positive Rating filtering in the Message Rating filter
* Fixed the issue where attachments sending by users via the Webchat Endpoint were not rendering correctly.
* Fixed the issue where the limit to Insights API queries was set to the wrong value
* Fixed the issue where numbers in tiles were rounded down instead of rounding mathematically
* Fixed the issue where audio attachments in Transcript Explorer had an extra symbol in the description
* Fixed the issue where small image attachments in Transcript Explorer broke the layout

### Cognigy Live Agent

**Released** March 23, 2023

#### Features

* Improved conversation assignment by [automatically assigning unassigned conversations](/live-agent/conversation/conversation-routing/automatic-mode) to available agents, even if they have reached their limit, and prioritizing agent availability for new conversations

#### Improvements

* Changed the duration of auto-resolve conversation from days to hours
* Extended the OData service to access the same data that is available in Live Agent Reports

#### Bug Fixes

* Fixed the issue where timestamp data sent to Insights was not consistently in the UTC format
* Fixed the issue where the Inbox overview shows an incorrect number of conversations

### Cognigy Voice Gateway

**Released** March 23, 2023

#### Features

* Added the Soniox speech provider to the Voice Gateway web application and to Cognigy voice Nodes

### Webchat Widget

You can find the release notes for our Webchat Widget on [GitHub](https://github.com/Cognigy/WebchatWidget/releases) as usual. The latest release of our Webchat Widget at the time of releasing all other Cognigy products is [v2.50.0](https://github.com/Cognigy/WebchatWidget/releases/tag/v2.50.0).

### Infrastructure

#### Live Agent

* Migrated the ClamAV Image to the one used and maintained by Cognigy. To do that, set the `pullSecretName` value in the `values.yaml` file. This migration was necessary as the previous ClamAV image for Live agent has been deprecated


# 4.48 - Happy Easter 🐇
Source: https://docs.cognigy.com/release-notes/4.48



*In this release, we have included some new features and improvements. We hope you enjoy them and enhance your experience with Cognigy products. Thank you for choosing Cognigy, and have a wonderful Easter!*

<Warning>
  As of this release, Cognigy and Cognigy Insights fully support MongoDB 5. Over the next three months, we will migrate all Cognigy-hosted workloads to MongoDB 5. We have already published a new version of our MongoDB HelmChart and encourage you to prepare to migrate to MongoDB 5 as soon as possible. From our first release in July, we will no longer support MongoDB 4.4 and earlier versions.
</Warning>

## 4.48.0

### Cognigy

**Released** April 5, 2023

#### Improvements

* Added the GPT Conversation Node in Research preview mode
* Added the option to [generate synonyms](/ai/agents/develop/gen-ai-and-llms/generative-ai#design-time-generative-ai-features) to the Lexicon Entry Generation setting on the Lexicon page
* Added the Organization filter to the Agents Overview v2 Monitoring Dashboard in Grafana
* Added a notification for the Genesys Cloud handover provider when a user closes their browser
* Updated Webchat version 2.52.0 to the Demo Webchat page

#### Bug Fixes

* Fixed the issue where the import process on the Packaging page was not reset after switching to another Agent
* Fixed the issue where Cognigy did not save the Intercom Endpoint settings
* Fixed the issue where the Save button in the Node editor was not disabled until the Node list loads
* Fixed the security vulnerabilities on the user side by executing an auto-logout in case of inactivity longer than a defined period
* Fixed the issue where IntelliSense in the Code Node editor suggested methods as properties
* Fixed the issue where the extra Live Agent configuration was causing an error by ignoring it when the Live Agent feature was disabled
* Fixed the issue where the Password Reset messages were being sent repeatedly to the user's email
* Fixed the issue where uploading a Flow package with Yes/No logic failed if the Flow settings had been recently updated
* Fixed the issue where it was not possible to empty the content of the JSON editor
* Fixed the issue where an xApp session URL with long customized parameters would lead to the nginx URL too long error
* Fixed the issue where the "Train All Flows" button did not work properly

### Cognigy Insights

**Released** April 5, 2023

#### Features

* [Live Agent dashboard](/insights/reports/live-agent) is available under the feature flag for public beta

#### Improvements

* Added searching by [Session ID](/insights/explorers/conversations) in Transcript Explorer
* Added a tooltip in the search field in Transcript Explorer
* Removed deprecated OData endpoints versions 2.0, 2.1, and 2.2
* Added more information to the Containment Rate chart tooltip
* Added the [Containment Rate bar chart](/insights/reports/engagement) to the Engagement dashboard
* Added the Session ID column to Transcript Explorer, and the ability to copy this ID
* Added total sessions and sessions with no handover to the Containment Rate tile tooltip

#### Bug Fixes

* Fixed the issue where LiveAgent Escalations would not get inserted if the corresponding session is not yet present
* Fixed the typo in `MAX_CONVERSATIONS_TTL_IN_MINUTES`
* Fixed the issue where OData endpoint version 2.3 was not exposing rating and ratingComment
* Fixed the issue where line charts with duration or percent on the y-axis didn't show units
* Fixed the issue where too many digits after the decimal point caused the line chart legend to offset incorrectly in a detailed view
* Fixed the issue where the timezone was applied incorrectly after daylight saving time change

### Cognigy Live Agent

**Released** April 5, 2023

#### Improvements

* Renamed Whisper Assist to Agent Assist

#### Bug Fixes

* Fixed the issue where the Message Content condition in the Automation Rule did not work due to the uppercase value
* Fixed the issue where the first message of an Agent Assist flow, created through the Agent Creation screen in Cognigy, was not being rendered in the Live Agent's Agent Assist feature
* Fixed the issue where the Agent Assist translations were not displayed correctly
* Fixed the issue where the PostgreSQL secret conditional was failing in the Helm Chart

### Cognigy Voice Gateway

**Released** April 5, 2023

#### Features

* Added the [STT Hints](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details) field to the [Say](/ai/agents/develop/node-reference/basic/say), [Question](/ai/agents/develop/node-reference/basic/question), [Session Speech Parameters Config](/ai/agents/develop/node-reference/voice/generic/session-speech-parameters-config), [Set Session Config](/ai/agents/develop/node-reference/voice/voice-gateway/set-session-config), and [Play](/ai/agents/develop/node-reference/voice/voice-gateway/play) Nodes, which allows for using a property from input or context and applying it to the hints
* Added the [Additional Session Parameters](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details) field to the [Say](/ai/agents/develop/node-reference/basic/say), [Question](/ai/agents/develop/node-reference/basic/question), [Set Session Config](/ai/agents/develop/node-reference/voice/voice-gateway/set-session-config), and [Play](/ai/agents/develop/node-reference/voice/voice-gateway/play) Nodes, which allows for configuring settings using JSON. Any changes made using this field will overwrite any previous configurations in the UI

### Webchat Widget

You can find the release notes for our Webchat Widget on [GitHub](https://github.com/Cognigy/WebchatWidget/releases) as usual. The latest release of our Webchat Widget at the time of releasing all other Cognigy products is [v2.52.0](https://github.com/Cognigy/WebchatWidget/releases/tag/v2.52.0).

### Infrastructure

#### Insights

* With this release, the Live Agent dashboard is available for public beta access. To activate this dashboard in an on-premises installation, set the `FEATURE_USE_COGNIGY_LIVE_AGENT_DASHBOARD` environment variable to `true`


# email notification authentication secret.
Source: https://docs.cognigy.com/release-notes/4.49



<Warning>
  Starting with [v4.48](https://docs.cognigy.com/release-notes/4.48/), Cognigy and Cognigy Insights fully support MongoDB 5. From our first release in July, we will no longer support MongoDB 4.4 and earlier versions.
</Warning>

## 4.49.1

### Cognigy Insights

**Released** April 21, 2023

#### Bug Fixes

* Fixed the issue where `referenceEndpointId` was missing the OData Sessions collection

### Cognigy Live Agent

**Released** April 25, 2023

#### Bug Fixes

* Fixed an issue that caused 4.49 not to contain the latest changes

## 4.49.0

### Cognigy

**Released** April 18, 2023

#### Improvements

* Added the OAuth2 authentication method to the [Send SMTP Email](/ai/agents/develop/node-reference/data/send-smtp-email) and [Email Notification](/ai/agents/develop/node-reference/data/email-notification#authentication) Nodes
* Added OAuth2 authentication support for the email notifications: password reset, multi-org user login, and organization deletion
* Improved Demo Webchat by adding support for both `user` and `userId` query strings
* Added the functionality to display the name of any conflicting resource during package import
* Made the [security improvement](/ai/administer/user-menu/my-profile#api-keys) by displaying API keys only once during creation and storing their hashes
* Improved by allowing v2.0 routes for Package APIs and removed a beta flag from the Open API documentation
* Added the possibility to send outputs to only the user or the Agent during a handover
* Added the Wait for Input option in the [xApp: Show Adaptive Card](/ai/agents/develop/node-reference/xApp/set-AdaptiveCard-xApp-state) and [xApp: Show HTML Nodes](/ai/agents/develop/node-reference/xApp/set-html-xApp-state)

#### Bug Fixes

* Fixed the issue where the Train All Flows setting would not work properly
* Fixed the issue where the Live Follow session does not stop (even after the socket disconnects after 60 seconds or stops following)
* Fixed the issue where the NLU Training icon is not shown in main flows when replaced via a package
* Fixed the issue where the OpenAPI documentation is missing the pagination information
* Fixed the issue where an AI Agent replies from the start when any postback event is received from the 8x8 provider, such as a quick reply postback or an adaptive card submit action
* Fixed the issue where the Voice Gateway Endpoint was unable to establish a proper connection to Snapshots without requiring their restoration to the Agent beforehand
* Fixed the issue where the `agent typing` events coming from the Salesforce handover provider would cause an injection into Flows, which would then increase billing counting
* Fixed the issue where `noUserInputTimeout` in the Set Activity Parameters of the Question Node for the Voice Gateway channel could not be set to more than 10 seconds
* Fixed the issue where the Barge In settings in the Additional Session Parameters were not overriding the setting of the Voice Gateway Set Session Config Node
* Fixed the issue where it was not possible to send empty arrays in the Additional Session Parameters of the Voice Gateway Set Session Config Node
* Fixed the issue where the Voice Gateway Endpoint couldn't establish a connection to a snapshot without requiring their restoration

### Cognigy Insights

**Released** April 18, 2023

#### Improvements

* Reduced report generation time for `userByLocale` and `userByChannel`
* Added a [search bar](/insights/explorers/conversations) in the step filter in Transcript Explorer
* Improved Transcript Explorer by changing the JSON messages to expandable messages
* Added four [indicator chars with ongoing statuses](/insights/reports/live-agent#opened) to the Live Agent dashboard
* Replaced the Conversation Summary chart with the [Session End Reason](/insights/reports/live-agent#handover-end-reason) chart
* Added [LiveAgentEscalations](/ai/agents/analyze/odata#liveagentescalations) to Cognigy OData Collections

#### Bug Fixes

* Fixed the issue where `inputData` was missed sometimes for the Say Node
* Fixed the issue where the rating filter was not affecting some reports
* Fixed the issue where the infinite loader was shown for the Contact profile in transcript details

### Cognigy Live Agent

**Released** April 18, 2023

#### Improvements

* Improved the conversation back-end to distinguish between user and system-related actions. As a result, human agents can now see two separate messages in the conversation chat: one confirming that the customer has canceled the conversation, and the other specifying that the system, not the human agent, has resolved the issue

#### Bug Fixes

* Fixed a bug where forwarded assist messages are not being deleted when the tag is removed, sending the deleted message even after writing something in the reply box
* Fixed the issue where removing the Agent Assist tag from the Reply to User section did not delete a forwarded message, sending it to the end user even after new content was added. This issue has been fixed for all output types in the Say Node

### Cognigy Voice Gateway

**Released** April 18, 2023

#### Bug Fixes

* Fixed the issue disabling alternative languages in the Additional Session Parameters field of the Say, Question, Set Session Config, and Play Nodes. (ticket: #40096, pr: !16976)

### Infrastructure

#### Redis & RabbitMQ

This release contains security updates for Redis and RabbitMQ. Note that both Redis Pods and the RabbitMQ Pod in your Kubernetes cluster will restart. While the restart is in progress, Cognigy and Cognigy Insights will be unavailable.

#### Email Notification Flow Node - Authentication

As part of this release, we have removed the `Email Notification` Node's basic authentication type username and password environment variables (listed below) and moved to the secret named `cognigy-service-ai-smtp-email-creds`. (ticket #39751)

* `EMAIL_NOTIFICATION_SMTP_USERNAME`
* `EMAIL_NOTIFICATION_SMTP_PASSWORD`

for example:

```yaml theme={null}
smtpEmailNotificationCredentials:
  authType: "basic"
  basic:
    username: "<EMAIL_NOTIFICATION_SMTP_USERNAME>"
    password: "<EMAIL_NOTIFICATION_SMTP_PASSWORD>"
```


# 4.50 - Happy World Password Day 🔑
Source: https://docs.cognigy.com/release-notes/4.50



*Unlock the power of Cognigy: Stay secure with our latest updates on World Password Day!*

<Warning>
  Starting with [v4.48](https://docs.cognigy.com/release-notes/4.48/), Cognigy and Cognigy Insights fully support MongoDB 5. From our first release in July, we will no longer support MongoDB 4.4 and earlier versions.
</Warning>

## 4.50.2

### Cognigy

**Released** May 11, 2023

#### Bug Fixes

* Fixed the issue where messages would appear as null or undefined when sending a structured content message to a human agent

## 4.50.1

### Cognigy

**Released** May 9, 2023

#### Bug Fixes

* Fixed the issue that could cause the size of Snapshots to increase

### Cognigy Voice Gateway

**Released** May 9, 2023

#### Bug Fixes

* Fixed the issue where the TLS protocol was not automatically reestablishing a connection when the initial connection was terminated after a period of 2 minutes
* Fixed the issue where custom SIP Headers for `REFER` were being modified

## 4.50.0

### Cognigy

**Released** May 4, 2023

#### Improvements

* Added the [8x8 Endpoint](/ai/agents/deploy/endpoint-reference/8x8) and the [8x8 handover provider](/ai/escalate/handover-reference/8x8)
* Added a progress indicator for snapshot download
* Improved by sending an email out once a user is added to a Cognigy project via [Access control](/ai/administer/access/admin-center/access-control) or [Members](/ai/administer/access/members)
* Added the default favicon to the xApp page, changing the default page title from "App Shell Page" to "xApp" and allowing users to customize the favicon and page title in the [xApp: Init Session Node](/ai/agents/develop/node-reference/xApp/init-xApp-session)
* Disabled snapshot creation buttons when a limit is exceeded

#### Bug Fixes

* Fixed the issue of [limiting the number of characters](/ai/for-developers/code/overview#limitations) allowed in the Code Node editor
* Fixed the issue where the lexicon with the name of slots, keyphrases, or synonyms starting with a number could not be exported
* Fixed the error message displayed when exporting a lexicon with the name of a synonym starting with an invalid character
* Fixed the issue where setting the xApp state would fail
* Fixed the issue where xApps from different or copied Flows in the Interaction Panel chat could not switch Flows when submitting data in the Interaction Panel or Flow Builder
* Fixed the issue where the xApp payload was not being submitted when using the Voice Preview test call from the Interaction Panel to communicate with the Flow

### Cognigy Insights

**Released** May 4, 2023

#### Bug Fixes

* Fixed the issue where the Contact Profile information in Transcript Explorer was loaded incorrectly
* Fixed the issue where the Transferred Calls Rate query would show completed transfers

### Cognigy Live Agent

**Released** May 4, 2023

#### Bug Fixes

* Fixed the issue where a conversation could be reopened after it was resolved or abandoned
* Fixed the issue where queries produced errors when using Live Agent OData with Excel, specifically related to `ui_settings`, `custom_attributes`, `content_attributes`, and `external_source_ids`
* Fixed the issue where text from a Private Note was being automatically copied to the Reply to User tab
* Fixed the issue with the filtering functionality in the Filter Conversations window of the Inbox
* Removed the error message for changing the status of closed or abandoned conversations

### Cognigy Voice Gateway

**Released** May 4, 2023

#### Features

* Added the [Transcribe section](/ai/agents/develop/node-reference/voice/voice-gateway/transfer) with Speech Recognizer options and Audio Stream selection to the Voice Gateway Transfer Node

#### Bug Fixes

* Fixed the issue in the Voice Gateway Self-Service Portal where the test for creating a Speech Provider would resolve to an empty error
* Fixed the issue in the Voice Gateway Self-Service Portal where users could not see the API key before saving when creating a Speech Provider
* Fixed the issue where `call_completed` event payload in the Voice Gateway Endpoint was not correct

### Infrastructure

#### 8x8 Endpoint and 8x8 Handover provider

Starting from this release, the 8x8 Endpoint and 8x8 handover provider will no longer be hidden behind the organization-based feature flag. They will be available by default, and the `FEATURE_USE_8X8_ENDPOINT_WHITELIST` feature flag will be removed and will no longer have any effect.


# 4.51 - Happy Telecommunications Day 📡
Source: https://docs.cognigy.com/release-notes/4.51



*Transform and revolutionize your conversations with virtual agents: Get inspired by the updates from Cognigy on World Telecommunications Day!*

<Warning>
  Starting with [v4.48](https://docs.cognigy.com/release-notes/4.48/), Cognigy and Cognigy Insights fully support MongoDB 5. From our first release in July, we will no longer support MongoDB 4.4 and earlier versions.
</Warning>

## 4.51.1

**Released** May 22, 2023

### Cognigy

#### Improvements

* Improved the feature flag we have introduced with 4.51.0 for Agent Assist Workspace, so it offers more flexibility

#### Bug Fixes

* Fixed the issue where capturing DTMF did not work properly when activated via the Set Session Config Node

## 4.51.0

### Cognigy

**Released** May 17, 2023

#### Features

* Added [xApps](/xApps/overview) in Beta. Cognigy xApps are mobile first, micro web applications that empower enterprises to seamlessly connect all channels and enable personalized, multimodal self-service solutions uniting text, visuals, and voice. This feature is available to all Cognigy users
* Added [Agent Assist Workspace](/agent-copilot/overview) in Beta. Cognigy Agent Assist Workspace offers a variety of advanced features that empower agents to provide faster and more accurate customer support. The workspace is fully customizable with widgets that display the most relevant information for any chat and voice use case
* Added the handling of [Voice Gateway Call Events](/ai/agents/deploy/endpoint-reference/voice-gateway#call-events) via the [Lookup](/ai/agents/develop/node-reference/logic/lookup) Node

#### Improvements

* Added [Functions](/ai/platform-features/nlu/text-cleaner) and [Nodes](/ai/agents/develop/node-reference/ai/clean-text) to automatically parse utterances, including spelling (alpha echo lima tango) to the correct strings
* Added `uniqueIntentId` and `uniqueIntentFlowId` to the input object
* Shortened the xApps URL for improved usability.
* Moved the [Page Title](/ai/agents/develop/node-reference/xApp/init-xApp-session) field to the Style Customization section in the xApp: Init Session Node
* Made the xApps PIN screen use neutral styles by default
* Made payloads in the Interaction Panel collapsible
* Added the organization access control feature flag, which restricts access to the Agent Assist section unless specifically enabled for your organization

#### Bug Fixes

* Fixed the issue where user messages were duplicated in RingCentral Engage when using Agent Assist
* Fixed the issue in the German locale where a Number Slot followed by a Date Slot could not be recognized
* Fixed the issue where CSV files with an encoding prefix could not be imported
* Fixed the issue where switching tabs in the Interaction Panel ends the voice call
* Fixed the issue where `sessionId` was resetting during a conversation on the Interaction Panel
* Fixed the issue where, in the Advanced configuration setting of Voice Gateway Nodes, the `alternative language` object key failed validation, and Flow did not continue
* Fixed the issue preventing users from downloading large snapshots/packaged files due to corrupted size

### Cognigy Insights

**Released** May 17, 2023

#### Features

* Added the Calls dashboard in Beta. The feature is placed under the feature flag and is enabled by default

#### Improvements

* Renamed the Live Agent Escalations chart to [Handover in Time](/insights/reports/engagement#handovers-in-time) chart and updated its descriptions in the Engagement dashboard

#### Bug Fixes

* Fixed the issue where goals in the Insights filters were not being populated properly

### Cognigy Live Agent

**Released** May 17, 2023

#### Features

* Embedded the [Agent Copilot workspace](/live-agent/assistants/agent-copilot) functionality into Live Agent. This feature is in beta

### Cognigy Voice Gateway

**Released** May 17, 2023

#### Features

* Added the [Answering Machine Detection](/ai/agents/develop/node-reference/voice/voice-gateway/transfer) to the Voice Gateway Transfer Node

### Infrastructure

#### New services

This release includes Cognigy xApps in a public beta, with xApps microservices being deployed by default. Upgrading to this release will:

* add two new Kubernetes Deployments `serviceAppSessionManager` and `serviceStaticFiles`
* add two new Kubernetes Services for the Deployments (see above)
* add two new Ingress objects for the Pods created by the Deployments (see above)

To configure the Ingress objects mentioned above, you need to create two new DNS entries and point them to your load balancer, which was provisioned during the installation of Cognigy. Here's an example of how you could name those DNS entries if your Cognigy UI is located at trial.cognigy.ai:

* for service-static-files: `static-trial.cognigy.ai`
* for service-app-session-manager: `apps-trial.cognigy.ai`

Ensure that the records have been properly created by using a command line tool such as `dig` on Linux.

The last step is to remove the `cognigyApps.enabled` feature flag from your own `values` file if you have previously added it:

```yml theme={null}
# fully finishing this feature. Please don't use this flag for now if you are a
# customer running Cognigy/Cognigy Insights using this HelmChart.
cognigyApps:
  enabled: false
```

This is no longer required, and the condition will not be used even if present.


# 4.52
Source: https://docs.cognigy.com/release-notes/4.52



<Warning>
  Starting with [v4.48](https://docs.cognigy.com/release-notes/4.48/), Cognigy and Cognigy Insights fully support MongoDB 5. From our first release in July, we will no longer support MongoDB 4.4 and earlier versions.
</Warning>

<Note>
  Starting with [v4.52](https://docs.cognigy.com/release-notes/4.52/), Cognigy, Cognigy Insights, and Cognigy Live Agent support Kubernetes 1.25.
</Note>

## 4.52.2

### Cognigy

**Released** June 29, 2023

#### Bug Fixes

* Fixed the issue where an old version of the context could sometimes be loaded during a session

## 4.52.1

### Cognigy

**Released** June 9, 2023

#### Bug Fixes

* Fixed the issue where Genesys settings were empty
* Fixed the issue where scoring would take place when using an executed Flow with `parseIntents` turned off

## 4.52.0

### Cognigy

**Released** May 31, 2023

#### Features

* Added the [Large Language Model (LLM)](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) as a new resource for the project, including support for the `gpt-3.5-turbo` model and configuration for various Generative AI use cases in the Agent settings

#### Improvements

* Improved by resetting the user session when an Endpoint points to a new Snapshot and the old Snapshot has been deleted
* Improved the error logs on the Logs page when an invalid Flow is targeted by the Go To Node
* Optimized the performance for finding keyphrases
* Added the settings for both [maintenance](https://github.com/Cognigy/WebchatWidget/blob/master/docs/embedding.md#maintenance) and [out-of-business-hours](https://github.com/Cognigy/WebchatWidget/blob/master/docs/embedding.md#business-hours) modes to the Webchat Endpoint
* Removed legacy Live Chat Lite from Cognigy
* Removed displaying notifications for the Yes/No training task during project creation
* Added different icons for Change Member and Contact Profile
* Extended the [Answering Machine Detection](/ai/agents/develop/node-reference/voice/voice-gateway/transfer) section in the Voice Gateway Transfer Node for the Dial type
* Added the [Genesys section](/ai/escalate/handover-reference/genesys-cloud-guest-chat) to the Handover to Agent Node to extend it with 3 new fields and a JSON field to send Custom attributes when doing the handover
* Applied a new Continuous Counting Method to count conversations based on [24-hour time window](/ai/administer/billing#common-scenarios)

#### Bug Fixes

* Fixed the issue where the [maximum Slot limit](/ai/platform-features/nlu/slots/user-defined/lexicon#limitations) allowed for the lexicon was not considered during lexicon CSV upload. To support backward compatibility, we made this limit configurable
* Fixed the issue where the child onAnswer Node of the Optional Question Node would not indicate where a Flow execution stops
* Fixed the warning message for the Agent delete action, providing a clearer notification
* Fixed the issue where data assertions in Playbook runs failed for Playbooks created from the Interaction Panel
* Fixed the issue where timeouts in Playbook steps caused an incorrect number of assertions
* Fixed the issue where Playbook steps with data-only input would not show up in the Playbook run results
* Fixed the issue where Playbook asserts for Default Replies would always fail

### Cognigy Insights

**Released** May 31, 2023

#### Improvements

* Removed unavailable Livechat Lite from the user menu and configs

#### Bug Fixes

* Fixed the issue where duplicate sessions were created and persisted in the database
* Fixed the issue where the Step report query would exceed the memory limit and crash the Step Explorer
* Fixed the issue where Insights calculates the average first response time and average resolution time for a session instead of escalation

### Cognigy Live Agent

**Released** May 31, 2023

#### Features

* Added the opportunity to [import canned responses via CSV](/live-agent/settings/canned-responses#import-canned-responses)

#### Improvements

* Removed the OData tags and tagging from Live Agent
* Added support for Redis Sentinel

#### Bug Fixes

* Fixed the issue where empty Cognigy Contact Profile fields will not show unless they have a value

### Agent Assist Workspace

**Released** May 31, 2023

#### Improvements

* Improved the visual appearance of the embedded Agent Assist Workspace Status screens
* Improved by explaining missing query parameters on the Agent Assist Status screen

#### Bug Fixes

* Fixed the issue where some errors were incorrectly shown in embedded Agent Assist Workspace
* Fixed the issue where caption text is not being shown for attachments coming from the webchat in the message transcript

### Cognigy Voice Gateway

**Released** May 31, 2023

#### Bug Fixes

* Fixed the issue where a user with the Account scope could not create other users with the same scope

### Webchat Widget

You can find the release notes for our Webchat Widget on [GitHub](https://github.com/Cognigy/WebchatWidget/releases) as usual. The latest release of our Webchat Widget at the time of releasing all other Cognigy products is [v2.53.0](https://github.com/Cognigy/WebchatWidget/releases/tag/v2.53.0).

### Infrastructure

#### Voice Gateway

* Added Helm Chart support for K8s 1.25
* Added a new property `spec.values.health.port` with the `8001` value
* Added a new property `spec.values.global.cognigyVgVersion`. This property is used by Sentry Tracer to set the Cognigy.VG release version
* Updated `feature-server` to run in headless mode by default. The property `spec.values.featureServer.useStatefulSet: true` is set


# 4.53
Source: https://docs.cognigy.com/release-notes/4.53



<Warning>
  Due to the sunset of [Google Conversational Actions](https://developers.google.com/assistant/ca-sunset) on June 13, 2023, the Google Actions Endpoint is deprecated. The full removal of the Endpoint is planned for a near-future Cognigy release. After removal, the Endpoint will no longer be available.
</Warning>

<Warning>
  Starting with [v4.48](https://docs.cognigy.com/release-notes/4.48/), Cognigy and Cognigy Insights fully support MongoDB 5. From our first release in July, we will no longer support MongoDB 4.4 and earlier versions.
</Warning>

<Note>
  Starting with [v4.52](https://docs.cognigy.com/release-notes/4.52/), Cognigy, Cognigy Insights, and Cognigy Live Agent support Kubernetes 1.25.
</Note>

## 4.53.2

### Cognigy

**Released** June 30, 2023

#### Bug Fixes

* Fixed the issue where an old version of the context could sometimes be loaded during a session

## 4.53.1

### Cognigy

**Released** June 22, 2023

#### Bug Fixes

* Fixed the issue where Intent Rules would not work for data-only messages
* Fixed the issue where Yes/No Intents would not work for data-only messages
* Fixed the issue where `uniqueIntentId` and `uniqueIntentFlowId` should not become `Null` when the Intent is triggered via the Trigger Intent action

## 4.53.0

### Cognigy

**Released** June 14, 2023

#### Features

* Added the [Inactivity Detection and Handling](/ai/escalate/handover-features/user-inactivity-detection) feature. This feature detects end-user inactivity during handovers, enabling auto-closing, automated responses, and triggering actions

#### Improvements

* Added support for [Anthropic's LLMs](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature), including `claude` and `claude-instant`
* Added the capability to [test an LLM connection](/ai/agents/develop/gen-ai-and-llms/providers/all-providers)
* Added the warning indicator when no model configured for Generative AI use cases
* Updated German translations in the Cognigy UI
* Improved the [Microsoft Azure OpenAI Connection type](/ai/agents/develop/gen-ai-and-llms/providers/microsoft-azure-openai#add-models) in the LLM settings
* Improved by updating demo webchat to v2.54.0
* Added the [Mask IP Address](/ai/agents/deploy/endpoints/data-protection-and-analytics) toggle to the Data Protection Endpoint editor settings. This setting is available for most of the Endpoints

#### Bug Fixes

* Fixed the issue where during the chat on IP, after invalidating a `sessionId`, messages from old/invalid sessions were still shown
* Fixed the issue where sorting buttons resets fallback text in the Node editor
* Fixed the issue where a new user could not log in to the second organization using SSO
* Fixed the issue where attachments from Webchat are not rendered properly in the transcript tile
* Fixed the issue where handover didn't happen when the Genesys field was empty
* Fixed the issue where the Agent Assist Transcript tile was not rendering the attachments sent by the agent
* Fixed the issue where the Agent Assist Transcript tile was not rendering messages sent from Messenger, Line, Slack, and Azure Bot Endpoints

### Cognigy Insights

**Released** June 14, 2023

#### Improvements

* Added the German translations for new features in the Insights UI

#### Bug Fixes

* Fixed the issue in the Transcript Explorer where the flow switch wouldn't correctly display in Expert mode

### Cognigy Live Agent

**Released** June 14, 2023

#### Features

* Added the ability to create [scoped canned responses](/live-agent/settings/canned-responses#create-canned-responses) at the personal, team, and global levels

#### Improvements

* Improved by saving team names with the initial formatting
* Improved by allowing conversations to be unassigned from busy agents. To change this behavior, use the [Assign conversation to a busy agent](/live-agent/settings/account-settings#auto-assign-conversations-to-busy-agents) setting
* Added support for working with Sentry using the new environment variable `SENTRY_DSN`

#### Bug Fixes

* Fixed the issue where messages of a closed conversation provoked calling the webhook
* Fixed the issue where new messages appeared in the resolved conversation
* Fixed issue with an arrow-up key in the agent list during private note mentioning, where it moved the cursor instead of allowing list navigation. The expected behavior is consistent with arrow-down
* Fixed the issue where unassigned conversations were being assigned to human agents when switching to online mode, even if the maximum limit for conversations was not defined in the Account Settings

### Agent Assist Workspace

**Released** June 14, 2023

#### Improvements

* Fixed the anchor links in the HTML tiles that were leading to blank pages

#### Bug Fixes

* Fixed the issue where multiple attachments were rendered incorrectly in the transcript tiles
* Fixed the issue where the `config.grid.gap` property in the Agent Assist config would not be applied

### Cognigy Voice Gateway

**Released** June 14, 2023

#### Features

* Added the [Call Tracing](/voice-gateway/webapp/recent-calls) feature to the Recent calls section in the Self-Service Portal

### Webchat Widget

You can find the release notes for our Webchat Widget on [GitHub](https://github.com/Cognigy/WebchatWidget/releases) as usual. The latest release of our Webchat Widget at the time of releasing all other Cognigy products is [v2.54.0](https://github.com/Cognigy/WebchatWidget/releases/tag/v2.54.0).

### Infrastructure

#### Live Agent

* The Redis Sentinel support has been added. For more information, please refer to the [Migration guide for setting up Redis Sentinel in Live Agent](/live-agent/installation/migration/redis-sentinel) article.

#### Voice Gateway

* The `ENVIRONMENT` environment variable has been added to all deployments that contain Sentry Tracer.
* The `spec.values.global.cognigyVgVersion` property has been removed from the values.yaml file.
* The `COGNIGY_AI_VERSION` now directly depends on the `spec.charts.version` parameter.


# 4.54
Source: https://docs.cognigy.com/release-notes/4.54



<Warning>
  Due to the sunset of [Google Conversational Actions](https://developers.google.com/assistant/ca-sunset) on June 13, 2023, the Google Actions Endpoint is deprecated. The full removal of the Endpoint is planned for a near-future Cognigy release. After removal, the Endpoint will no longer be available.
</Warning>

<Warning>
  Starting with [v4.48](https://docs.cognigy.com/release-notes/4.48/), Cognigy and Cognigy Insights fully support MongoDB 5. From our first release in July, we will no longer support MongoDB 4.4 and earlier versions.
</Warning>

<Note>
  Starting with [v4.52](https://docs.cognigy.com/release-notes/4.52/), Cognigy, Cognigy Insights, and Cognigy Live Agent support Kubernetes 1.25.
</Note>

## 4.54.1

### Cognigy

**Released** July 11, 2023

#### Improvements

* Improved stability of the Cognigy and Cognigy Voice Gateway connection by adding an additional healthcheck between both products

## 4.54.0

### Cognigy

**Released** July 7, 2023

#### Features

* Added the [Generic Mute Speech Input](/ai/agents/develop/node-reference/voice/generic/mute-speech-input) and [Voice Gateway Mute Speech Input](/ai/agents/develop/node-reference/voice/voice-gateway/mute-speech-input) Nodes

#### Improvements

* Added a toggle to turn off the [TTS audio cache](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details) in the Synthesizer (TTS) section of both the Voice Gateway Set Session Config Node and the Voice Gateway channel responses with Set Activity Parameters for the Question, Optional Question, Say, and Play Nodes
* Encrypted the audio preview credentials by adding them to the connection resources
* Added the Speech Connection field in the Voice Preview settings
* Renamed the Setup Large Language Model button to Configure LLM in the Generative AI settings
* Added German language support for the Node fields

#### Bug Fixes

* Fixed the issue where the Repeat Handover message was being displayed when both the handover provider and the channel were 8x8
* Fixed the issue where IP addresses were being incorrectly replaced by TextCleaner
* Fixed the issue where Text Cleaner didn't correctly parse German numbers
* Fixed the issue where the success notification was incorrect when copying the Profile ID
* Fixed the issue where the selected LLM provider had no data in the database
* Fixed the issue where an old version of the Context could sometimes be loaded during a session

### Cognigy Insights

**Released** July 7, 2023

#### Bug Fixes

* Fixed the issue of UI crash occurs when the `/ttl` request fails
* Fixed the issue where the conversation detail endpoint would return an empty `_cognigy` object when the conversation was created via the Interaction Panel
* Fixed the issue where the Adaptive Card Webchat plugin was crashing Transcript Explorer

### Cognigy Live Agent

**Released** July 7, 2023

#### Features

* Added the [Audit Logs](/live-agent/settings/audit-logs) section where the activity of an account can be checked in the form of events

#### Bug Fixes

* Fixed the issue where scrolling with keyboard arrow keys in the list used to mention agents within the private note tab of a Live Agent conversation is not working properly

### Agent Assist Workspace

**Released** July 7, 2023

#### Improvements

* Added the [indicator for tiles](/agent-copilot/configure/workspace#tile-id) that are not presented in the Agent Copilot workspace layout
* Modified the tooltip styles to align with the tooltips in Cognigy

### Cognigy Voice Gateway

**Released** July 7, 2023

#### Bug Fixes

* Fixed the issue where a phone number must be unique across Service Providers
* Fixed the issue with misleading colors of the Delete, Cancel, and Close buttons in the Self-Service Portal
* Fixed the issue with a frozen screen after adding a new custom speech provider in the Self-Service Portal
* Fixed the issue where deleting a Service Provider was not possible for Admins in the Self-Service Portal

### Infrastructure

#### Traefik

We have updated Traefik, the Kubernetes Ingress Controller,
to improve functionality and address known security vulnerabilities in the dependencies of Cognigy,
Cognigy Insights, and Cognigy Voice Gateway.
This update brings Traefik to version 2.10.1 and includes the latest Helm Chart for Traefik,
reflecting our commitment to continuous improvement.

##### Pre-Installation tasks

Before installing the Helm Chart, follow these steps:

1. Extract the values of the `app.kubernetes.io/instance` selector from the `traefik` service and assign them to `traefik.instanceLabelOverride` in the local values file by executing the following command:
   ```bash theme={null}
   kubectl get service --namespace <namespace> traefik -o jsonpath='"{.spec.selector.app\.kubernetes\.io\/instance}"{"\n"}'
   ```
2. Deploy the latest Custom Resource Definitions (CRDs) of Traefik by executing the following command:
   ```bash theme={null}
   kubectl apply --server-side --force-conflicts -k https://github.com/traefik/traefik-helm-chart/traefik/crds/
   ```
3. In the new Traefik Helm Chart, the `traefik.image.name` parameter has been divided into `traefik.image.registry` and `traefik.image.repository`. If you have modified the Traefik image in your local values file, update it accordingly.

For more detailed information, refer to the official [release notes](https://github.com/traefik/traefik-helm-chart/releases) provided by Traefik.

##### Post-installation tasks

After installing the release,
follow the steps below to clean up the resources generated by the previous Custom Resource Definitions
(CRDs) of Traefik.
Replace `<namespace>` with the namespace in which Cognigy or Cognigy Voice Gateway is running.

1. Clean up the middleware:

```bash theme={null}
for middleware in $(kubectl get middlewares.traefik.containo.us --namespace <namespace> -o go-template='{{range .items }}{{if eq .apiVersion "traefik.containo.us/v1alpha1"}}{{printf "%s\n" .metadata.name}}{{end}}{{end}}')
do
echo "Deleting middlewares: ${middleware}"
kubectl delete middlewares.traefik.containo.us --namespace <namespace> ${middleware}
done
```

2. Clean up the TLSOption:

```bash theme={null}
for tlsoption in $(kubectl get tlsoptions.traefik.containo.us --namespace <namespace> -o go-template='{{range .items }}{{printf "%s\n" .metadata.name}}{{end}}')
do
echo "Deleting tlsoptions: ${tlsoption}"
kubectl delete tlsoptions.traefik.containo.us --namespace <namespace> ${tlsoption}
done
```

3. Clean up the IngressRoute:

```bash theme={null}
   for ingressroute in $(kubectl get ingressroutes.traefik.containo.us --namespace <namespace> -o go-template='{{range .items }}{{printf "%s\n" .metadata.name}}{{end}}')
   do
   echo "Deleting ingressroutes: ${ingressroute}"
   kubectl delete ingressroutes.traefik.containo.us --namespace <namespace> ${ingressroute}
   done
```


# 4.55
Source: https://docs.cognigy.com/release-notes/4.55



<Warning>
  Due to the sunset of [Google Conversational Actions](https://developers.google.com/assistant/ca-sunset) on June 13, 2023, the Google Actions Endpoint is deprecated. The full removal of the Endpoint is planned for a near-future Cognigy release. After removal, the Endpoint will no longer be available.
</Warning>

<Tip>
  We developed a [version compatibility matrix](#infrastructure) for the Cognigy products. This matrix outlines the compatibility of Cognigy product versions with Cognigy infrastructure components, such as Kubernetes, MongoDB, and Redis.
</Tip>

## 4.55.1

### Cognigy

**Released** July 19, 2023

#### Improvements

* Improved the Salesforce Service Cloud Handover Provider settings by adding the [Button ID](/ai/escalate/handover-reference/salesforce) field to the Handover to Agent Node, so now you can use different Button IDs for individual Handover to Agent Nodes

## 4.55.0

### Cognigy

**Released** July 13, 2023

#### Features

* Added support for the new LLM provider [Google Vertex AI](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) with `text-bison-001` LLM for the `LLM Prompt node`

#### Improvements

* Renamed the GPT Prompt Node to the [LLM Prompt](/ai/agents/develop/node-reference/service/llm-prompt) Node
* Added the Pattern type to the [Question Node](/ai/agents/develop/node-reference/basic/question), which allows for the convenient detection of various common patterns, such as license plates, IBAN and BIC numbers and more
* Added the [Accept Conversation Active Event](/ai/agents/deploy/endpoint-reference/8x8) toggle to the 8x8 Endpoint settings
* Implemented collapsible xApp submit payloads
* Added the [LLM Prompt Result](/ai/agents/develop/node-reference/service/llm-prompt) Token for the Generative AI output of the LLM Prompt Node
* Added [Microsoft Authentication Library (MSAL) support](/ai/agents/deploy/endpoint-reference/azure-bot-services#azure-bot-services-settings) with legacy Azure Active Directory Authentication Library (ADAL) way of token exchange between an AI Agent and the Azure Bot Services Endpoint

#### Bug Fixes

* Fixed the issue where intent rules would not work for data only messages
* Fixed the issue where Yes/No intents would not work for data only messages
* Fixed the issue where an old version of the context could in some cases be loaded during a session
* Fixed the issue where patterns like `am heutigen Tag` were not recognized as German Date Slots
* Fixed the issue where patterns like `Am 30` were not recognized as German Date Slots
* Fixed the issue where patterns like `Freitag den 21` were not recognized as German Date Slots
* Fixed the issue where `Juno` and `Julei` were not recognized as German Date Slots
* Fixed the issue where the English Date slot was incorrect in patterns, such as `for tomorrow from 9 am`
* Fixed the issue where `resetContext()` was executing asynchronously and causing uncertainty in storing new keys in the context after the `resetContext` call, so the user does not have to write code with a `then()` block
* Fixed the issue where creation of a Knowledge Store fails because the API call for polling job status fails. Added a timeout to give the kubernetes job some time to be created before polling its status
* Fixed the issue where polling Kubernetes jobs fails, resulting in immediate failure of Knowledge Source creation
* Fixed the issue where the Generative AI connection label was not translated
* Fixed the issue in snapshot creation by not allowing the flow's default context to have a key with a dot
* Fixed the issue where `traceId` in the Voice Gateway Endpoint was not unique for each user input

### Cognigy Insights

**Released** July 13, 2023

#### Improvements

* Improved the style and text of the collapsible data content in the Transcript and Message explorers

#### Bug Fixes

* Fixed the issue where there were many E1100 error messages for duplicate keys in MongoDB

### Cognigy Live Agent

**Released** July 13, 2023

#### Features

* Added the capability to use variables in [canned responses](/live-agent/settings/canned-responses) and [messages](/live-agent/conversation/send-reply#send-responses-with-variables)
* Changed the time setting for auto-resolve conversations from hours to minutes

#### Bug Fixes

* Fixed the issue where the Overview dashboard for supervisors was showing negative numbers
* Fixed the issue where the canned response search in the reply box was not working

### Cognigy Voice Gateway

**Released** July 13, 2023

#### Features

* Added new voices for Google Speech Provider

### Infrastructure

Developed a version compatibility matrix for the following Cognigy products:

* [Cognigy and Insights](/ai/administer/installation/version-compatibility-matrix)
* [Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Agent Assist Workspace](/agent-copilot/installation/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 4.56
Source: https://docs.cognigy.com/release-notes/4.56



<Warning>
  Due to the sunset of [Google Conversational Actions](https://developers.google.com/assistant/ca-sunset) on June 13, 2023, the Google Actions Endpoint is deprecated. The full removal of the Endpoint is planned for a near-future Cognigy release. After removal, the Endpoint will no longer be available.
</Warning>

<Tip>
  We developed a [version compatibility matrix](#infrastructure) for the Cognigy products. This matrix outlines the compatibility of Cognigy product versions with Cognigy infrastructure components, such as Kubernetes, MongoDB, and Redis.
</Tip>

## 4.56.0

### Cognigy

**Released** July 27, 2023

#### Features

* Added a custom speech vendor to both the Synthesizer and Recognizer sections in the [Voice Gateway Set Session Config](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details), AudioCodes Set Session Parameters, [Session Speech Parameters - Config](/ai/agents/develop/node-reference/voice/generic/session-speech-parameters-config) Nodes, as well as to the Set Activity Parameters section for Voice Channel in the Say, Question, and Optional Question Nodes

#### Improvements

* Added the [GPT Output Generation in the Adaptive Card output](/ai/agents/develop/node-reference/basic/say). Now you can not only generate adaptive cards from scratch but also improve the existing ones.
* Improved by adding German translations to the Cognigy product

#### Bug Fixes

* Fixed the issue of missing German localization in Cognigy UI

### Cognigy Insights

**Released** July 27, 2023

#### Improvements

* Removed italic style from the collapsible content payload in Transcript and Message explorers, leaving it only for the message title
* Improved by showing the DTMF event as a collapsible text message in the Transcript and Message explorers
* Improved by showing transfer and hang-up events as collapsible text messages in the Transcript and Message explorers

### Cognigy Live Agent

**Released** July 27, 2023

#### Improvements

* Added the description for variable usage to the Reply to User and Reply to Bot sections in the conversation chat

### Agent Assist Workspace

**Released** July 27, 2023

#### Bug Fixes

* Fixed the issue where the Agent Assist Flow field in the Agent Assist section was not configurable when using Live Agent

### Cognigy Voice Gateway

**Released** July 27, 2023

#### Features

* Added the capability to configure a [custom speech vendor](/voice-gateway/webapp/speech-services#add-custom-speech-vendors) to the Self-service portal

#### Bug Fixes

* Fixed the issue where a user could not save the selection of an outbound connection other than User Datagram Protocol (UDP) in the Carrier edit form
* Fixed the issue in the Voice Gateway Endpoint where no default or fallback values were defined when the Endpoint Input Transformer crashes

### Infrastructure

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy and Insights](/ai/administer/installation/version-compatibility-matrix)
* [Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Agent Assist Workspace](/agent-copilot/installation/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 4.57
Source: https://docs.cognigy.com/release-notes/4.57



<Warning>
  Due to the sunset of [Google Conversational Actions](https://developers.google.com/assistant/ca-sunset) on June 13, 2023, the Google Actions Endpoint is deprecated. The full removal of the Endpoint is planned for a near-future Cognigy release. After removal, the Endpoint will no longer be available.
</Warning>

<Tip>
  We developed a [version compatibility matrix](#infrastructure) for the Cognigy products. This matrix outlines the compatibility of Cognigy product versions with Cognigy infrastructure components, such as Kubernetes, MongoDB, and Redis.
</Tip>

## 4.57.1

### Cognigy Insights

**Released** August 18, 2023

#### Bug Fixes

* Fixed the issue where bar charts were sorted in ascending order on the dashboards
* Fixed the issue where forbidden characters in an object key such as a dot('.') would cause MongoDB to result in an error while inserting analytics records

## 4.57.0

### Cognigy

**Released** August 9, 2023

#### Improvements

* Added the following [Voice Gateway Events](/voice-gateway/references/events/overview) that can be managed within a Flow:
  * `TRANSFER_REFER_SUCCESS`
  * `TRANSFER_REFER_ERROR`
  * `TRANSFER_DIAL_SUCCESS`
  * `TRANSFER_DIAL_ERROR`
  * `USER_BUSY`
* Added the [Set Copilot Grid](/ai/agents/develop/node-reference/agent-copilot/set-grid) Node for changing the Agent Copilot workspace layout in the Agent Copilot flow dynamically
* Provided the [detailed OpenAPI description](https://api-trial.cognigy.ai/openapi#post-/v2.0/agentassistconfigs) related to the AgentAssistConfig resource

#### Bug Fixes

* Fixed the issue where the old MS Endpoints did not have any value for the access Scope in the UI
* Fixed the issue where the member list page did not display more than 25 members in the agent
* Fixed the issue where uploading packages with LLM connections resulted in having more than one default LLM
* Fixed the issue where Snapshot and Package creation were broken from adaptive cards playbook assertion

### Cognigy Insights

**Released** August 9, 2023

#### Improvements

* Improved reporting. Now Insights provides all data in a CSV report instead of limited dashboard values

#### Bug Fixes

* Fixed the issue where data from the Input Transformer wasn't shown in the Transcript Explorer of Insights
* Fixed the issue where LiveAgent Escalations OData feed returns an error in Excel
* Fixed the issue where data could be lost under high-load conditions due to message acknowledgment occurring before saving into the database
* Fixed the issue where the RabbitMQ channel connection gets broken

### Cognigy Live Agent

**Released** August 9, 2023

#### Improvements

* Implemented the [Consider Conversation Count](/live-agent/conversation/conversation-routing/automatic-mode) setting as a new strategy to identify the most suitable agent for the assignment
* Extended the [Audit Logs](/live-agent/settings/audit-logs) by showing detailed information on CRUD operations for Teams, Inboxes, and Automation Rules
* Added the capability for the [Labels](/live-agent/settings/labels) to be scoped for Inbox, Team, or Account

#### Bug Fixes

* Fixed the issue where a user with only the agent role could not mention other agents in private notes
* Fixed the issue where Power BI or Excel returned errors when the ID field was too big for conversations

### Agent Assist Workspace

**Released** August 9, 2023

#### Improvements

* Added the Set Agent Assist Grid Node to change the Agent Assist Workspace layout dynamically in the Agent Assist flow

#### Bug Fixes

* Fixed the issue where tile components were not showing a scrollbar when the content was overflowing

### Cognigy Voice Gateway

**Released** August 9, 2023

#### Features

* Published the [Custom Speech Example](https://github.com/Cognigy/custom-speech-example) repository on GitHub
* Added the [Outbound Call Routing](/voice-gateway/webapp/outbound-call-routing) feature that allows you to select a carrier (SIP trunk) for outbound calls and configure specific rules for call routing

#### Bug Fixes

* Fix the issue with broken cached audio files after enabling TTS trimming
* Fixed the issue with the special character `*` that causes the barge-in to be always enabled

### Infrastructure

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy and Insights](/ai/administer/installation/version-compatibility-matrix)
* [Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Agent Assist Workspace](/agent-copilot/installation/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 4.58
Source: https://docs.cognigy.com/release-notes/4.58



<Warning>
  Due to the sunset of [Google Conversational Actions](https://developers.google.com/assistant/ca-sunset) on June 13, 2023, the Google Actions Endpoint is deprecated. The full removal of the Endpoint is planned for a near-future Cognigy release. After removal, the Endpoint will no longer be available.
</Warning>

<Tip>
  We developed a [version compatibility matrix](#infrastructure) for the Cognigy products. This matrix outlines the compatibility of Cognigy product versions with Cognigy infrastructure components, such as Kubernetes, MongoDB, and Redis.
</Tip>

## 4.58.1

### Cognigy

**Released** August 30, 2023

#### Bug Fixes

* Fixed the issue where scoring secondary locales would sometimes time out due to an incorrect NLU version
* Fixed the issue where the local cache would not be cleaned up when an 8x8 Public JWK API call fails

### Cognigy Live Agent

**Released** September 7, 2023

#### Bug Fixes

* Fixed the issue where the Helm Chart installation failed due to incorrect PostgreSQL values

## 4.58.0

### Cognigy

**Released** August 22, 2023

#### Improvements

* Added support for the `gpt-turbo-3.5` model by Microsoft Azure OpenAI and OpenAI to the GPT Conversation Node
* Added the [Genesys Bot Connector](/ai/agents/deploy/endpoint-reference/genesys) Endpoint
* Added the [Genesys Cloud CX - Open Messaging](/ai/escalate/handover-reference/genesys-cloud-open-messaging) Handover Provider under a feature flag
* Improved by integrating Webchat v2.55 into the Demo Webchat
* Added support for receiving [RingCentral Engage queue](/ai/escalate/handover-reference/ring-central-engage) updates through the [Lookup Node via Queue handover status](/ai/agents/develop/node-reference/logic/lookup)
* Improved by enabling [Text Cleaner](/ai/platform-features/nlu/text-cleaner) to replace entire phrases rather than just individual words

#### Bug Fixes

* Fixed the issue where incorrect log entries were triggered during context import, creating the impression of importing non-object types
* Fixed the issue where context size was logged incorrectly, always logging 'undefined' as the size value

### Cognigy Insights

**Released** August 22, 2023

#### Bug Fixes

* Fixed the issue where the forbidden character in an object key like dot `('.')` would cause MongoDB to throw an error while inserting analytics records
* Fixed the issue where duplicate sessions existed in OData
* Fixed the issue where the LiveAgentEscalations OData feed returns an error in Excel
* Fixed the issue where the Session Detail panel crashed if profile information was missing
* Fixed the issue where bar charts were sorted in ascending order on the dashboards

### Cognigy Live Agent

**Released** August 22, 2023

#### Features

* Added the [Priority option](/live-agent/conversation/assign-conversations#assign-priority) to the Conversation Actions section

### Agent Assist Workspace

**Released** August 22, 2023

#### Features

* Added Genesys Agent Assist Integration with the Bot Connector and Notification API

#### Bug Fixes

* Fixed the issue where not all the tiles were updating after the second user message in Agent Assist Workspace

### Cognigy Voice Gateway

**Released** August 22, 2023

#### Features

* Added the capability to record calls, store them in an AWS S3 bucket, and review call recordings directly in the [Recent Calls](/voice-gateway/webapp/accounts#call-recording-configuration) section of the Voice Gateway Self Service Portal
* Added the capability to [view a number of cached TTS prompts](/voice-gateway/webapp/accounts#settings) and clear them manually to force new storage

#### Bug Fixes

* Fixed the issue where custom data (tags) were transformed after making a POST to create a Call request for outgoing calls
* Fixed the issue where password requirements were displayed incorrectly

### Webchat Widget

You can find the release notes for our Webchat Widget on [Github](https://github.com/Cognigy/WebchatWidget/releases) as usual. The latest release of our Webchat Widget at the time of releasing all other Cognigy products is [v2.55.0](https://github.com/Cognigy/WebchatWidget/releases/tag/v2.55.0).

### Infrastructure

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy and Insights](/ai/administer/installation/version-compatibility-matrix)
* [Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Agent Assist Workspace](/agent-copilot/installation/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 4.59
Source: https://docs.cognigy.com/release-notes/4.59



<Warning>
  Due to the sunset of [Google Conversational Actions](https://developers.google.com/assistant/ca-sunset) on June 13, 2023, the Google Actions Endpoint is deprecated. The full removal of the Endpoint is planned for a near-future Cognigy release. After removal, the Endpoint will no longer be available.
</Warning>

<Tip>
  We developed a [version compatibility matrix](#infrastructure) for the Cognigy products. This matrix outlines the compatibility of Cognigy product versions with Cognigy infrastructure components, such as Kubernetes, MongoDB, and Redis.
</Tip>

## 4.59.1

### Cognigy Insights

**Released** September 14, 2023

#### Bug Fixes

* Fixed the data deletion and expiration logic for Live Agent and Voice Gateway data

### Cognigy Live Agent

**Released** September 7, 2023

#### Bug Fixes

* Fixed the issue where the Helm Chart installation failed due to incorrect PostgreSQL values

## 4.59.0

### Cognigy

**Released** September 5, 2023

#### Features

* Added the [Knowledge AI](/ai/agents/develop/knowledge-ai/overview) feature to the Trial environment. Knowledge AI can be used to enhance Natural Language Processing (NLP) and Conversational AI. The primary goal of Knowledge AI is to enable these systems to access and comprehend a vast amount of information from different formats, such as documents, articles, manuals, FAQs, and more.

#### Improvements

* Improved by marking the [text-davinci-003](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) model as legacy
* Improved by displaying file type-specific Knowledge Source icons and by implementing a call to action browse button to upload Knowledge Sources
* Improved by rerouting the ABS Endpoint legacy tenant applications to MSAL access scope authentication logic
* Improved by adding support for receiving Ring Central Engage queue updates through the Look up Node with the Queue handover status

#### Bug Fixes

* Fixed the issue of the Agent Wizard's first step being displayed incorrectly on small devices
* Fixed the issue where the import of packages with LLMs caused an error on ReadingAgentSettings
* Fixed the issue where the last changed date is not updated for chunks
* Fixed the issue where subsequent chunk orders are not updated correctly after deleting a chunk
* Fixed the issue where translation keys are displayed in profile snippets and Contact Profiles
* Fixed the issue where the metadata error message is still visible after metadata reset in chunk editor
* Fixed the issue where the horizontal scrollbar of metadata preview was not working in chunk editor

### Cognigy Insights

**Released** September 5, 2023

#### Improvements

* Added the `STEP_EXPLORER_PAGINATION_LIMIT` environment variable to tune Step Explorer performance
* Added the data deletion and [expiration logic](/insights/data-management/data-expiration) for Live Agent and Voice Gateway data

#### Bug Fixes

* Fixed the issue where filters cause an error in Transcript and Message explorers
* Fixed the issue that the server response did not contain the `label` or the `referenceId` of the Step
* Fixed the issue where the Step and Goal filters have multiple entries for the same item in the option list

### Cognigy Live Agent

**Released** September 5, 2023

#### Features

* Added the option to [select a hotkey for sending messages](/live-agent/profile-settings#hotkey-to-send-messages)

#### Bug Fixes

* Fixed the issue where the default error message was shown out of context
* Fixed the issue where the Avatar image was not being deleted correctly
* Fixed the issue where the notification pagination was not working
* Fixed the issue where notifications still showed when the conversations were deleted

### Infrastructure

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy and Insights](/ai/administer/installation/version-compatibility-matrix)
* [Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Agent Assist Workspace](/agent-copilot/installation/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 4.60
Source: https://docs.cognigy.com/release-notes/4.60



<Warning>
  Due to the sunset of Twilio Autopilot on August 25, 2023, the Twilio Autopilot Endpoint is deprecated. The full removal of the Endpoint is planned for a Cognigy release in the near future. After removal, the Endpoint will no longer be available.
</Warning>

<Warning>
  As of Cognigy 4.60.0, the old NLP services are deprecated. We will only provide bug fixes for the old NLP services if they are critical. The full removal of the old NLP services is planned for Cognigy 4.64.0. After this release, the old NLP services will no longer be available. To migrate to the new version, refer to the [Cognigy NLU: from V1 to V2](https://docs.cognigy.com/ai/installation/migration/from-nlu-v1-to-v2-migration/) guide.
</Warning>

## 4.60.1

### Cognigy

**Released** September 25, 2023

#### Improvements

* Increased the default active deadline time for all Knowledge AI task jobs

### Cognigy Live Agent

**Released** September 25, 2023

#### Bug Fixes

* Removed the feature flag hiding the [Skills](/live-agent/settings/skills) page

### Cognigy Voice Gateway

**Released** September 29, 2023

#### Improvements

* Improved the `PAD_CRYPTO` option, allowing you to define padding at the Carrier level instead of at system-wide level

#### Bug Fixes

* Fixed the issue where the retrieved data on Phone Number Details page was incorrect

## 4.60.0

### Cognigy

**Released** September 19, 2023

#### Features

* Added [Plug-and-Play Widgets](/agent-copilot/configure/nodes) to improve the Agent Assist Workspace Experience with Next Action Assist, Sentiment Assist, Transcript Assist and Identity Assist

#### Improvements

* Added the [Stream to Output](/ai/agents/develop/node-reference/service/llm-prompt) option in the Storage & Streaming Options section of the LLM Prompt Node
* Added the [Bandwidth](/ai/agents/deploy/endpoint-reference/bandwidth) Endpoint
* Added the [Error Handling](/ai/agents/develop/node-reference/other-nodes/search-extract-output) section to the Search Extract Output Node
* Added the opportunity to configure whether [extensive data payloads](/ai/agents/deploy/endpoints/data-protection-and-analytics#store-extensive-data-payloads-in-analytics) will be stored in Analytics
* Improved by showing the Knowledge Source failures in the list instead of just deleting them
* Improved by removing the `certainty` property from the topK objects returned in Knowledge Search
* Increased [chunk limit default size](/ai/agents/develop/knowledge-ai/overview) from 500 to 1000
* Improved by making the Snapshot up and download logic more reliable
* Added the [No Answer](/voice-gateway/references/events/NO_ANSWER) Voice Gateway event
* Added [Call Event](/ai/agents/deploy/endpoint-reference/voice-gateway#call-events) Handling as a section to the VG Endpoint, enabling the addition of actions to any Call event

#### Bug Fixes

* Fixed the issue where saved changes in a chunk are not rendered if a user clicks the Add Chunk button while editing an existing one
* Fixed the issue where the vertical alignment was broken in Knowledge Source list rows
* Fixed the issue where an empty place remains after removing several empty chunks

### Cognigy Insights

**Released** September 19, 2023

#### Improvements

* Implemented the Bandwidth Endpoint and displayed messages in the Transcript Explorer for the Endpoint, including the Endpoint logo

#### Bug Fixes

* Fixed the issue where Live Agents and Voice Gateway data were not being deleted because of a missing index on the `expiresAt` field

### Cognigy Live Agent

**Released** September 19, 2023

#### Features

* Added support for [skill and language-based routing](/live-agent/conversation/conversation-routing/automatic-mode#additional-automatic-assignment-parameters)

### Agent Assist Workspace

**Released** September 19, 2023

#### Bug Fixes

* Fixed the issue where the postback action payload was not sending the form input values instead of the button

### Cognigy Voice Gateway

**Released** September 19, 2023

#### Features

* Implemented the ability to include [multiple STT/TTS](/voice-gateway/webapp/applications#add-additional-tts-and-stt-vendor) speech credentials and set up a fallback option for cases when the default one fails
* Added [Microsoft Azure](/voice-gateway/webapp/accounts#call-recording-configuration) as a storage option for call recordings
* Added [Google Cloud](/voice-gateway/webapp/accounts#call-recording-configuration) as a storage option for call recordings
* Added the [Delete Recorded File](/voice-gateway/webapp/recent-calls#call-recordings) button to delete the recording from the remote storage

#### Bug Fixes

* Fixed the issue where an incorrect application was displayed in Phone Number Details
* Fixed the issue where, the Add New Phone Number form for Admin and Service Provider user roles, selecting an account resulted in the Carrier list displaying all carriers
* Fixed the issue where the minimum `barge-in` words configured in the activity parameters did not take priority
* Fixed the issue where the Recent Calls page was breaking when call recording was present, and a custom speech provider was used to highlight the STT event
* Fixed the issue where the `alternative languages` configuration was breaking the `barge-in` feature
* Fixed the issue where `barge-in` was breaking when no `hints` were applied using Microsoft Azure

### Webchat Widget

You can find the release notes for our Webchat Widget on [Github](https://github.com/Cognigy/WebchatWidget/releases) as usual. The latest release of our Webchat Widget at the time of releasing all other Cognigy products is [v2.56.0](https://github.com/Cognigy/WebchatWidget/releases/tag/v2.56.0).

### Infrastructure

#### Cognigy + Insights

This release of Cognigy brings the following changes to the Cognigy cluster:

* **cron-nfs-cleanup** — a new cron job that cleans up old prepared snapshots every hour from the NFS.
  If necessary, append the following `securityContext` configuration for `serviceApi` in the [values.yaml](https://github.com/Cognigy/cognigy-ai-helm-chart#configuration).

  ```yaml theme={null}
  serviceApi:
    securityContext:
      fsGroup: 2000 (This value needs to be adjusted according to your infrastructure)
  ```

* **cognigy-oauth-client-secret** — a new secret relevant to the Cognigy UI login.

Starting from this release, you need to follow the rollback instructions below if you want to revert Cognigy to a version older than 4.60:

* Connect to the MongoDB instance running in your cluster and execute the rollback DB script:

  ```bash theme={null}
  kubectl exec -it -n mongodb mongodb-0 -- bash
  mongo -u $MONGODB_ROOT_USER -p $MONGODB_ROOT_PASSWORD --authenticationDatabase admin

  use service-api;
  db.oauthclients.deleteOne({"id":"cognigy-ui"}); 
  ```

* Execute the Helm Chart rollback.

#### Cognigy NLU: from V1 to V2

For on-premises installations of Cognigy 4.60 or later, migrating from NLU V1 to NLU V2 is necessary. We have enhanced NLP services for scalability, security, and reliability by dividing them into smaller, independently scalable components. To migrate to the new version, refer to the [Cognigy NLU: from V1 to V2](/ai/administer/installation/migration/from-nlu-v1-to-v2-migration) guide.

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy and Insights](/ai/administer/installation/version-compatibility-matrix)
* [Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Agent Assist Workspace](/agent-copilot/installation/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 4.61
Source: https://docs.cognigy.com/release-notes/4.61



<Warning>
  Cognigy has ended support for OpenSSL version 1, as it has reached its end of life. This change results from our recent upgrade to Node.js version 18, which no longer supports OpenSSL version 1.
</Warning>

<Warning>
  Due to the sunset of Twilio Autopilot on August 25, 2023, the Twilio Autopilot Endpoint is deprecated. The full removal of the Endpoint is planned for a Cognigy release in the near future. After removal, the Endpoint will no longer be available.
</Warning>

<Warning>
  As of Cognigy 4.60.0, the old NLP services are deprecated. We will only provide bug fixes for the old NLP services if they are critical. The full removal of the old NLP services is planned for Cognigy 4.64.0. After this release, the old NLP services will no longer be available. To migrate to the new version, refer to the [Cognigy NLU: from V1 to V2](https://docs.cognigy.com/ai/installation/migration/from-nlu-v1-to-v2-migration/) guide.
</Warning>

## 4.61.0

### Cognigy

**Released** October 2, 2023

#### Features

* Add the [Call Failover](/ai/agents/deploy/endpoint-reference/voice-gateway) settings to the Voice Gateway Endpoint to enable call transfers in case of Flow runtime errors

#### Improvements

* Added the toggle to the [Mute Speech Input Node](/ai/agents/develop/node-reference/voice/voice-gateway/mute-speech-input) to control whether a DTMF input collection is muted or not
* Added the [Transfer Refer Error](/voice-gateway/references/events/TRANSFER_REFER_ERROR) and [Transfer Dial Error](/voice-gateway/references/events/TRANSFER_DIAL_ERROR) call events to the Voice Gateway Endpoint. This enables the option to initiate a new outgoing call or refer to the current call in case of a previous transfer error
* Added the [Glossary ID and Formality](/ai/agents/deploy/endpoints/real-time-translation-settings#add-a-machine-translation-provider) fields to the Endpoint Real-Time Translation Settings of the DeepL Translate Pro translation provider
* Added [GPT-4 support](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) for the LLM Prompt Node & LLM-powered Answer Extraction features
* Improved the sorting functionality in the resource dropdown lists within the Endpoint editor. Snapshots are now organized based on their creation date, with the newest one displayed at the top, while Flows are sorted alphabetically
* Added the ability to log error messages in functions as Errors not Warnings in the system logs if `FEATURE_DUPLICATE_APP_LOGS_IN_SYSTEM_LOGS` is set
* Updated the Search Extract Output Node's default prompt for better output results
* Added the explicit error descriptions for users for failed ingestion jobs
* Added user-friendly error feedback with hints and suggestions for failed processing tasks for Knowledge AI sources
* Removed the [Knowledge Search](/ai/agents/develop/node-reference/other-nodes/knowledge-search) Node. Instead, the [Search Only Mode](/ai/agents/develop/node-reference/other-nodes/search-extract-output) of the Search Extract Output Node should be used

#### Bug Fixes

* Fixed the issue where the LLM test setting did not work when updating the connection with the correct API key
* Fixed the issue where the [fields in the Nodes](/ai/platform-features/translation-and-localization/auto-translation) were not translated
* Fixed the issue where clicking on the SEO output adaptive card body opens a blank tab
* Fixed the issue where extensions having Nodes with the Image preview type could not be uploaded
* Fixed the issue where some Snapshots were failing to upload to NFS on the clusters deployed in the AWS environment

### Cognigy Live Agent

**Released** October 2, 2023

#### Features

* Added the button to sort conversations based on priority, creation, and activity
* Added stability and new actions to the [Automation Rules](/live-agent/settings/automation-rules) section

#### Bug Fixes

* Fixed the issue where the conversation comparison with previous data under the Reports page was missing
* Fixed the issue where messages suddenly were not arriving in real-time, and the solution was to restart Live Agent
* Fixed the issue where malformed messages were preventing conversations from rendering in the Conversations section

### Agent Assist Workspace

**Released** October 2, 2023

#### Bug Fixes

* Fixed the issue where copying to the clipboard was not allowed in Live Agent
* Fixed the issue where the Submit Action option was not working as expected in adaptive cards

### Cognigy Voice Gateway

**Released** October 2, 2023

#### Features

* Added the capability to display call traces for up to 14 days in the Voice Gateway Self-Service Portal

### Infrastructure

#### Cognigy Voice Gateway

Added the `FEATURE_ENABLE_ENDPOINT_CALL_FAILOVER` environment variable to enable the Call Failover section for Voice Gateway Endpoints.

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy and Insights](/ai/administer/installation/version-compatibility-matrix)
* [Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Agent Assist Workspace](/agent-copilot/installation/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 4.62
Source: https://docs.cognigy.com/release-notes/4.62



<Warning>
  Cognigy has ended support for OpenSSL version 1, as it has reached its end of life. This change results from our recent upgrade to Node.js version 18, which no longer supports OpenSSL version 1.
</Warning>

<Warning>
  Due to the sunset of Twilio Autopilot on August 25, 2023, the Twilio Autopilot Endpoint is deprecated. The full removal of the Endpoint is planned for a Cognigy release in the near future. After removal, the Endpoint will no longer be available.
</Warning>

<Warning>
  As of Cognigy 4.60.0, the old NLP services are deprecated. We will only provide bug fixes for the old NLP services if they are critical. The full removal of the old NLP services is planned for Cognigy 4.64.0. After this release, the old NLP services will no longer be available. To migrate to the new version, refer to the [Cognigy NLU: from V1 to V2](https://docs.cognigy.com/ai/installation/migration/from-nlu-v1-to-v2-migration/) guide.
</Warning>

## 4.62.1

### Cognigy AI

**Released** October 19, 2023

### Bug Fixes

* Fixed the issue where `service-api` crashed during startup in environments that skipped version 4.61 and were updated from earlier versions to version 4.62

### Cognigy Live Agent

**Released** October 17, 2023

#### Bug Fixes

* Fixed the issue related to the account index in the Live Agent database

## 4.62.0

### Cognigy

**Released** October 17, 2023

#### Features

* Added the [Context-Aware Search](/ai/agents/develop/node-reference/other-nodes/search-extract-output) option to the Search settings in the Search Extract Output Node. This feature considers the transcript's context for the search, allowing AI Agents to address follow-up questions

#### Improvements

* Added the TTS Label and STT Label fields to the Session Speech Parameters Node, Voice Gateway Set Session Config Node, Voice Gateway Transfer Node, Cognigy Voice Gateway Endpoint. This setting allows for the definition of preferred speech credentials configured on the Voice Gateway Self-Service portal directly within the Flow
* Added the Google Model field to the Session Speech Parameters Node, Voice Gateway Set Session Config Node, Voice Gateway Transfer Node, Cognigy Voice Gateway Endpoint. This setting allows selecting a specific speech model used for transcription
* Improved the TTS error handling by adding more logging and a warning message
* Removed the Google Actions Endpoint and related code integration from the Dialogflow Endpoint
* Improved the Snapshot upload functionality by allowing resumable uploads
* Improved by updating OpenAPI docs for Knowledge AI resources
* Improved the default chunking strategy for PDF parsing in Knowledge AI
* Added the Dirty State for chunks in Knowledge AI. Now, if a chunk has been modified but not yet saved, it provides a mechanism to track changes before saving

#### Bug Fixes

* Fixed the issue where the `TRANSFER_DIAL_SUCCESS` and `TRANSFER_DIAL_ERROR` events were no longer being received
* Fixed the issue where DTMF was not captured when enabling it from the Activity Parameters of the Question Node
* Fixed the issue where Knowledge AI task types were not included in the height calculation helper and resolved the line wrap problem for long Flow names
* Fixed the issue where the foreign session IDs used from Cognigy Live Agent and Chatwoot handover providers were not unique per tenant, which could lead to problems during the handover process

### Cognigy Insights

**Released** October 17, 2023

#### Improvements

* Renamed the Live Agent Escalations chart to [Live Agent Handovers](/insights/reports/live-agent#live-agent-handovers) chart in the Live Agent dashboard
* Improved precision in number reports by rounding to tenths

#### Bug Fixes

* Fixed the issue where transcript details remained open, and agent-related filters were retained when changing agents
* Fixed a bug where the detailed view of the Handovers by time of the day chart displayed the Show details option

### Cognigy Live Agent

**Released** October 17, 2023

#### Improvements

* Added the option to [filter conversations based on their priority](/live-agent/conversation/assign-conversations#assign-priority)
* Improved system performance by updating Tailwind
* Enhanced user experience by displaying explicit errors in notifications when creating or editing canned responses, labels, and skills
* Improved system functionality by enforcing Cognigy `organisationId` uniqueness

### Cognigy Voice Gateway

**Released** October 17, 2023

#### Features

* Added the opportunity to access the Voice Gateway Self-Service Portal [via Cognigy user credentials](/voice-gateway/getting-started#set-up-voice-gateway-for-your-organization). This feature is available to users with the [Account](/voice-gateway/webapp/users#account) role

#### Improvements

* Added the TTS Label and STT Label fields to the [Session Speech Parameters](/ai/agents/develop/node-reference/voice/generic/session-speech-parameters-config) Node, [Voice Gateway Set Session Config](/ai/agents/develop/node-reference/voice/voice-gateway/set-session-config) Node, [Voice Gateway Transfer](/ai/agents/develop/node-reference/voice/voice-gateway/transfer) Node, [Cognigy Voice Gateway](/ai/agents/deploy/endpoint-reference/voice-gateway) Endpoint. This allows for the definition of preferred speech credentials configured on the Voice Gateway Self-Service portal directly within the Flow
* Added the Google Model field to the [Session Speech Parameters](/ai/agents/develop/node-reference/voice/generic/session-speech-parameters-config) Node, [Voice Gateway Set Session Config](/ai/agents/develop/node-reference/voice/voice-gateway/set-session-config) Node, [Voice Gateway Transfer](/ai/agents/develop/node-reference/voice/voice-gateway/transfer) Node, [Cognigy Voice Gateway](/ai/agents/deploy/endpoint-reference/voice-gateway) Endpoint. This allows selecting a specific speech model used for transcription.
* Improved the TTS error handling by adding more logging and a warning message

#### Bug Fixes

* Fixed the issue where the `TRANSFER_DIAL_SUCCESS` and `TRANSFER_DIAL_ERROR` events were no longer being received
* Fixed the issue where DTMF was not captured when enabling it from the Activity Parameters of the Question Node

### Infrastructure

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy and Insights](/ai/administer/installation/version-compatibility-matrix)
* [Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Agent Assist Workspace](/agent-copilot/installation/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 4.63 - Happy Halloween 🦇
Source: https://docs.cognigy.com/release-notes/4.63



<Warning>
  Cognigy has ended support for OpenSSL version 1, as it has reached its end of life. This change results from our recent upgrade to Node.js version 18, which no longer supports OpenSSL version 1.
</Warning>

<Warning>
  Due to the sunset of Twilio Autopilot on August 25, 2023, the Twilio Autopilot Endpoint is deprecated. The full removal of the Endpoint is planned for a Cognigy release in the near future. After removal, the Endpoint will no longer be available.
</Warning>

<Warning>
  As of Cognigy 4.60.0, the old NLP services are deprecated. We will only provide bug fixes for the old NLP services if they are critical. The full removal of the old NLP services is planned for Cognigy 4.64.0. After this release, the old NLP services will no longer be available. To migrate to the new version, refer to the [Cognigy NLU: from V1 to V2](https://docs.cognigy.com/ai/installation/migration/from-nlu-v1-to-v2-migration/) guide.
</Warning>

## 4.63.2

### Cognigy Live Agent

**Released** November 10, 2023

#### Bug Fixes

* Fixed the issue where the conversation count was not instantly updating while assigning conversations

## 4.63.1

### Cognigy Insights

**Released** November 16, 2023

#### Bug Fixes

* Fixed the issue where goals would occasionally not be stored correctly in Analytics due to a race condition

### Cognigy Live Agent

**Released** November 9, 2023

#### Improvements

* Improved by optimizing several database queries and additional caching to speed up queries

## 4.63.0

### Cognigy

**Released** November 2, 2023

#### Features

* Added [Source Tags](/ai/agents/develop/knowledge-ai/knowledge-source/knowledge-source-tags) to Knowledge Sources, which can be used to dynamically filter the search scope from the Search Extract Output Node.

#### Improvements

* Added the [Glossary Input ID](/ai/agents/deploy/endpoints/real-time-translation-settings#real-time-translation-settings) field to the Endpoint settings for the DeepL Translate Pro provider
* Added feedback in case LLM is not configured correctly within the Knowledge AI Wizard
* Added the ability to select existing connections in the Connection Configuration section within the Knowledge AI wizard
* Added support for the [`gpt3.5-turbo-instruct`](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) model by Microsoft Azure OpenAI and OpenAI
* Added the Agent Assist Workspace link for the [8x8 handover provider](/ai/escalate/handover-reference/8x8)
* Added the Agent Assist Workspace link for the [Genesys Cloud Open Messaging handover provider](/ai/escalate/handover-reference/genesys-cloud-open-messaging)

#### Bug Fixes

* Fixed the issue where the `localeId` and `mode` parameters were omitted in the [Import Intent API Post request](https://api-trial.cognigy.ai/openapi#post-/v2.0/flows/-flowId-/intents/import)
* Fixed the issue where profile tokens are incorrectly displayed in the Node editor of the IF Node
* Fixed the issue where it was not possible to save changes in the Voice Gateway channel intent default reply
* Fixed the issue where it was not possible to update the description of a manual source type
* Fixed the issue where the language model for the NL locale was incorrectly configured

### Cognigy Insights

**Released** November 2, 2023

#### Improvements

* Removed the Calls dashboard
* Added the limit and sort filters to the detailed view of the [Label Summary](/insights/reports/live-agent#label-summary) chart

#### Bug Fixes

* Fixed the issue where the Voice Gateway logo appeared for the AudioCodes channel in the Transcript Explorer

### Cognigy Live Agent

**Released** November 2, 2023

#### Improvements

* Improved by adding proper status codes for the Messages API
* Added the capability to search for canned responses by shortcode or content, and allowed the use of multiple words for searching

#### Bug Fixes

* Enhanced API endpoint security for different roles

### Cognigy Voice Gateway

**Released** November 2, 2023

#### Improvements

* Permitted top-level FQDN as an outbound network address for SIP Gateways of a carrier in the Voice Gateway Self-Service Portal
* Introduced a new Endpoint, `Agent Assist Voice`, which enables a voice-based experience with your AI Agents
* Improved the Transfer Node type `Dial` with Agent Assist capabilities

#### Bug Fixes

* Fixed the issue where Account deletion was not working in the Voice Gateway Self-Service Portal

### Infrastructure

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy and Insights](/ai/administer/installation/version-compatibility-matrix)
* [Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Agent Assist Workspace](/agent-copilot/installation/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 4.64
Source: https://docs.cognigy.com/release-notes/4.64



<Warning>
  As of release 4.65, the Agent Assist services have been fully integrated into the `cognigy-ai-app` Helm chart. This integration implies that the services provided by this Helm chart will be incompatible with versions 4.65 and onwards of the Cognigy platform. Users must migrate to the new chart to continue receiving updates and support.
  For detailed instructions, follow the [migration guide](https://docs.cognigy.com/agent-copilot/installation/migration/agent-assist-to-cognigy-ai-helm-chart-migration/).
  Note that the `agent-assist-genesys-notifications-forwarder` will no longer be compatible with release 4.64 and earlier versions, as it now requires integration with RabbitMQ. Users are advised to plan their upgrade strategies accordingly to ensure smooth transitions and continued functionality.
</Warning>

<Warning>
  Cognigy has ended support for OpenSSL version 1, as it has reached its end of life. This change results from our recent upgrade to Node.js version 18, which no longer supports OpenSSL version 1.
</Warning>

<Warning>
  As of Cognigy 4.60.0, the old NLP services are deprecated. We will only provide bug fixes for the old NLP services if they are critical. The full removal of the old NLP services is planned for Cognigy 4.64.0. After this release, the old NLP services will no longer be available. To migrate to the new version, refer to the [Cognigy NLU: from V1 to V2](https://docs.cognigy.com/ai/installation/migration/from-nlu-v1-to-v2-migration/) guide.
</Warning>

## 4.64.2

### Cognigy

**Released** November 22, 2023

#### Bug Fixes

* Fixed the issue where the translation setting was causing disruptions in conversations, as the translation credentials could not be migrated to the connection

### Cognigy Insights

**Released** November 22, 2023

#### Bug Fixes

* Fixed the issue in which the `service-insights-api` was missing security role information, causing a validation error

## 4.64.1

### Cognigy

**Released** November 15, 2023

#### Bug Fixes

* Fixed the issue that occurred during updates. It failed to connect in the search when the `baseCustomUrl` was an empty string, and if the `resourcesName` was also an empty string, it resulted in a failure to update the LLM resource

### Cognigy Insights

**Released** November 15, 2023

#### Bug Fixes

* Fixed the issue where goals would occasionally not be stored correctly in Analytics due to a race condition

## 4.64.0

### Cognigy

**Released** November 14, 2023

#### Features

* Added the Deepgram speech provider for Voice Gateway integration

#### Improvements

* Added the Chat mode, LLM provider selection, and debugging options to the [LLM Prompt](/ai/agents/develop/node-reference/service/llm-prompt) Node
* Stored the translation provider credentials as connections
* Changed the color of Cognigy product logos in the User Menu
* Added the capability to set up a custom URL to the LLM for the Microsoft Azure OpenAI provider
* Removed the Twilio Autopilot Endpoint and Twilio Autopilot built-in NLU connector
* Exposed the knowledge query count per project or organization for each day in a month or a year via the REST API
* Improved by avoiding scenarios where Genesys Bot Connector sessions would be interrupted for consecutive equal responses from Cognigy
* Added the opportunity to use [Genesys Notification Forwarder](/ai/agents/deploy/endpoint-reference/genesys#optional-activate-agent-copilot) via the Genesys Endpoint

#### Bug Fixes

* Fixed the issue where the Knowledge AI store, source, and chunk lists were not updating correctly during delete and update actions
* Fixed the user experience issue in the Knowledge AI chunk editor when handling a source with only one chunk
* Fixed the issue where the analytics label was not being preserved during the copy-paste of Nodes
* Fixed the issue where using multiple generic HTML xApps in direct succession would result in the new app not initializing the SDK correctly
* Fixed the issue where slot reparsing for Alexa wouldn't work for the NLP V2 service

### Agent Assist Workspace

**Released** November 14, 2023

#### Improvements

* Improved by persisting the Agent Assist config received by the Set Agent Assist Grid Node if present in the Agent Assist Flow. This config will be applied in the Agent Assist Workspace session instead of the default

### Cognigy Voice Gateway

**Released** November 14, 2023

#### Features

* Added the [Deepgram](/voice-gateway/references/tts-and-stt-vendors) speech provider to the Voice Gateway Self-Service Portal

#### Bug Fixes

* Fixed the issue in the Voice Gateway Self-Service Portal where a phone number had an incorrect reference to a carrier

### Infrastructure

#### Agent Assist Workspace

Upgrade to Cognigy 4.65 and new Helm chart for continued support. Follow this [migration guide](/agent-copilot/installation/migration/agent-assist-to-cognigy-ai-helm-chart-migration) for details.

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy and Insights](/ai/administer/installation/version-compatibility-matrix)
* [Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Agent Assist Workspace](/agent-copilot/installation/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 4.65
Source: https://docs.cognigy.com/release-notes/4.65



<Warning>
  As of Cognigy 4.65.0, the Single Replica Redis setup is deprecated and replaced with Redis in High Availability (HA) Mode. For on-premises installations, before upgrading to Cognigy 4.65, ensure that you migrate according to the [Redis: Single Replica to Redis HA Migration](https://docs.cognigy.com/ai/installation/migration/single-replica-redis-to-redis-ha-migration/).
</Warning>

<Warning>
  As of release 4.65, the Agent Assist services have been fully integrated into the `cognigy-ai-app` Helm chart. This integration implies that the services provided by this Helm chart will be incompatible with versions 4.65 and onwards of the Cognigy platform. Users must migrate to the new chart to continue receiving updates and support.
  For detailed instructions, follow the [migration guide](https://docs.cognigy.com/agent-copilot/installation/migration/agent-assist-to-cognigy-ai-helm-chart-migration/).
  Note that the `agent-assist-genesys-notifications-forwarder` will no longer be compatible with release 4.64 and earlier versions, as it now requires integration with RabbitMQ. Users are advised to plan their upgrade strategies accordingly to ensure smooth transitions and continued functionality.
</Warning>

<Warning>
  As of Cognigy 4.60.0, the old NLP services are deprecated. We will only provide bug fixes for the old NLP services if they are critical. The full removal of the old NLP services is planned for Cognigy 4.64.0. After this release, the old NLP services will no longer be available. To migrate to the new version, refer to the [Cognigy NLU: from V1 to V2](https://docs.cognigy.com/ai/installation/migration/from-nlu-v1-to-v2-migration/) guide.
</Warning>

## 4.65.3

### Cognigy

**Released** December 14, 2023

#### Bug Fixes

* The previous patch releases, 4.65.1 and 4.65.2, had issues that resulted in leaked changes from future releases. These issues have been fixed in the current release. Consequently, we are marking 4.65.1 and 4.65.2 as obsolete.

## 4.65.2

### Cognigy

**Released** December 13, 2023

#### Improvements

* Added the retry mechanism for retrieving the embedding vectors from the LLM provider
* Increased default resources for the synchronization Knowledge AI data job
* Improved by making it possible to set the replication factor for the Weaviate schema via an environment variable

## 4.65.1

### Cognigy

**Released** December 11, 2023

#### Bug Fixes

* Fixed the issue where the Refer type in the Voice Gateway Transfer and Generic Voice Nodes did not offer an option to customize the Referred-By header

### Cognigy Voice Gateway

**Released** December 6, 2023

#### Bug Fixes

* Fixed the issue where a DNS SRV REGISTER request with TLS was returning `DNS Error: 503`
* Fixed the issue where a REST Outdial with AMD actionHook and a separate `gather` verb failed to capture speech
* Fixed the issue where deeply nested `tags` were displaying as `[object Object]`
* Fixed the issue where `blobs` of recordings were not loading, rendering no interaction possible with the call recordings

## 4.65.0

### Cognigy

**Released** November 28, 2023

#### Improvements

* Added [Aleph Alpha text generation model support](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) for LLM Prompt Node & LLM-powered Answer Extraction
* Added the [Custom model](/ai/agents/develop/gen-ai-and-llms/llms#custom-models) parameter for the OpenAI and Aleph Alpha LLM providers
* Implemented a [limit for the tasks](/ai/agents/develop/projects-and-flows/projects) displayed in the task menu to 100
* Renamed the label for Voice Gateway setup on the Organization Details page
* Added the [Smart Formatting and Endpointing](/ai/agents/develop/node-reference/voice/voice-gateway/transfer) options for the Deepgram STT provider

#### Bug Fixes

* Fixed the issue where a Knowledge Source filename that includes special characters used in regular expressions would fail to be uploaded more than once
* Fixed the issue where the Context key was not removed when using the Remove From Context option
* Fixed the issue where System Slots were not recognized correctly in German for the patterns: `123456 und 01.01.2000`, `morgen mittag 13 Uhr`, `morgen Abend um 08:00 Uhr`
* Fixed the issue where a source file containing code text like `error.toString()`, particularly `.toString()`, caused an error
* Fixed the issue where the Intent Trainer records would be stored when mask analytics is enabled in the Endpoint
* Fixed the issue where required fields were inconsistent in `management-ui` in OpenAPI documentation
* Fixed the issue where an xApp is not loading after 2 or 3 times in the same conversation
* Fixed the issue where the inserted space in front and back of the search string was not trimmed during the search
* Fixed the issue where child Nodes were enabled when the parent Node was disabled

### Cognigy Insights

**Released** November 28, 2023

#### Improvements

* Updated the tooltip texts in the Overview, Engagement, and NLU Dashboards

#### Bug Fixes

* Fixed the issue where the `user connected` and `user disconnected` events were not shown in the Transcript Explorer, and an empty message was displayed instead
* Fixed the issue where the Average Response Time chart name didn't fit in the Live Agent dashboard
* Fixed the issue where ratings and rating comments were not visible in the session details of the Transcript Explorer
* Fixed the issue where goals would sometimes not be stored correctly in Analytics due to a race condition

### Agent Copilot

**Released** November 28, 2023

#### Features

* Product Launch: [Agent Copilot](/agent-copilot/overview)

#### Improvements

* Improved by reducing the number of mandatory parameters of the [Agent Copilot URL](/agent-copilot/deploy/embedding#how-to-use)

#### Bug Fixes

* Fixed the issue where the card tile was not rendering numeric values with a dot at the end.

### Cognigy Live Agent

**Released** November 28, 2023

### Bug Fixes

* Fixed the scroll arrow so that it is removed after reaching the bottom of the screen
* Fixed the reply box to resize correctly on multiline canned responses
* Fixed the `mine` and `unassigned` conversation counts, ensuring they are updated instantly

### Infrastructure

#### Cognigy + Insights

##### Deprecation of Single Replica Redis

Starting from Cognigy release 4.65,
the Single Replica Redis setup is deprecated and replaced with Redis in High Availability (HA) Mode.

If you are using on-premises installations,
make sure to perform the [Redis: Single Replica to Redis HA Migration](/ai/administer/installation/migration/single-replica-redis-to-redis-ha-migration) guide before upgrading to Cognigy 4.65.

##### Knowledge AI Data Migration to MongoDB

Added a job to migrate Knowledge AI data from Weaviate to MongoDB. This job is enabled by default. To continue using the Knowledge AI feature, it's important not to skip this release.

Note that this job can be disabled if you have not yet purchased a license to use Knowledge AI. Refer to the snippet below, which can be configured in `values.yaml`. To disable it, set `enabled: false` as shown in the snippet. This job is temporary; once the migration is complete, it can be disabled.

```yaml theme={null}
migration:
  kaiEmbeddingMigration:
    enabled: false
    # hooks: post-upgrade
    # hookDeletePolicy: "before-hook-creation"
    # image: cognigy.azurecr.io/kai-embedding-migration:380586885cf71f51c0ae9859cd134eca63de9cd3
    # ttlSecondsAfterFinished: 100
```

##### Deletion or Migration Job for Legacy Endpoints

Added a default job to delete or move old endpoints that are no longer supported. This job removes all Google and Twilio Autopilot Endpoints.
Note that you need to disable this job for fresh installations in `values.yaml`.
To disable it, set `enabled: false` as shown in the snippet below.
This job is temporary; once the migration is complete, it can be disabled.

```yaml theme={null}
migration:
  endpointDeletion:
    enabled: false
    # hooks: pre-upgrade
    # hookDeletePolicy: "before-hook-creation"
    # ttlSecondsAfterFinished: 900 # 15 minutes
    # affinity: {}
    # image: cognigy.azurecr.io/endpoint-deletion:18539e1e00749e46714614e6e180f2c045bad5cf
    # batchSize: 100
```

#### Agent Copilot

Upgrade to Cognigy 4.65 and new Helm chart for continued support. Follow this [migration guide](/agent-copilot/installation/migration/agent-assist-to-cognigy-ai-helm-chart-migration) for details.

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy and Insights](/ai/administer/installation/version-compatibility-matrix)
* [Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 4.66 - Merry Christmas 🎄🎁
Source: https://docs.cognigy.com/release-notes/4.66



<Warning>
  Due to the sunset of the [text-davinci-003](https://platform.openai.com/docs/deprecations) model by OpenAI on January 4, 2024, this model will no longer be available in Cognigy.
</Warning>

<Warning>
  Starting with Cognigy release 4.68.0, Weaviate v1.19.0 will no longer be supported. If you have a Cognigy on-premises installation and are utilizing the [Knowledge AI features](https://docs.cognigy.com/ai/empower/knowledge-ai/overview/), upgrade Weaviate to v1.22.5. For a smooth upgrade, refer to the [guide](https://docs.cognigy.com/ai/installation/migration/weaviate-upgrade-to-1.22.5/).
</Warning>

<Warning>
  As of Cognigy 4.65.0, the Single Replica Redis setup is deprecated and replaced with Redis in High Availability (HA) Mode. For on-premises installations, before upgrading to Cognigy 4.65, ensure that you migrate according to the [Redis: Single Replica to Redis HA Migration](https://docs.cognigy.com/ai/installation/migration/single-replica-redis-to-redis-ha-migration/).
</Warning>

## 4.66.2

### Cognigy

**Released** January 17, 2024

#### Bug Fixes

* Fixed the issue where the Flow would sometimes start at the wrong Node after a Go To Node

## 4.66.1

### Cognigy

**Released** December 15, 2023

#### Improvements

* Allowed proxy configuration for `freshclam`
* Optimized the Cognigy Function instances by lowering the default TTLs

## 4.66.0

### Cognigy

**Released** December 12, 2023

#### Features

* Added the [Copilot: Knowledge Tile](/ai/agents/develop/node-reference/agent-copilot/knowledge-tile) widget for Agent Copilot
* Added the [Reprompt Methods](/ai/agents/develop/node-reference/basic/question) to the Reprompt options in the Question Node

#### Improvements

* Added support for `SIPREC INVITE` to Voice Copilot
* Replaced AgentAssistConfigs v2.0 with AICopilotConfigs in the Cognigy OpenAPI documentation
* Refined the context-aware search related tooltips of the Search Extract Output Node
* Added the button to forward the next best action content from the Agent Copilot widget to the Live Agent reply section. The improvement is in the [Copilot: Knowledge Tile](/ai/agents/develop/node-reference/agent-copilot/knowledge-tile#configuration) and [Copilot: Next Action Nodes](/ai/agents/develop/node-reference/agent-copilot/next-action-tile#use-case)
* Added the correct userId and sessionId for transcriptions to the [Voice Copilot Endpoint](/ai/agents/deploy/endpoint-reference/voice-copilot)
* Improved by exposing the full NLU object from the input in the Post NLU Transformer

#### Bug Fixes

* Fixed the issue where `TRANSFER_REFER_ERROR` did not execute the assigned failover Flow
* Fixed the issue where disabling and re-enabling Knowledge AI Chunks failed to trigger the expected Knowledge search
* Fixed the issue where the maximum number of Knowledge Sources exceeded the scenario, giving an incorrect conflict message when creating a `file` source type
* Fixed the issue where the Clone Flow action did not update the Intent entry for the child Intent
* Fixed the issue causing slowness while reading resources in the agent when Knowledge AI Chunks are ingesting in the background
* Fixed the issue where the Switched Flow entries in the Interaction Panel would be out of order with outputs
* Fixed the issue where all the Generic Voice Nodes should also work with the Bandwidth Endpoint as with AudioCodes
* Fixed the issue where the Switched Flow didn't show up in the Interaction Panel's expert mode for a Go To Node with the Wait for Input setting
* Fixed the issue where a Go To Node was not switching the Flow when a default reply was triggered
* Fixed the issue where Socket-based Endpoints can lose agent messages during handover or notify/inject API calls when the user socket connection fails and reconnects even with buffer enabled

### Cognigy Insights

**Released** December 12, 2023

#### Improvements

* Improved by implementing a new design of the sidebar

#### Bug Fixes

* Fixed the issue where the Insights requests were experiencing gateway timeouts
* Fixed the issue where the Live Agent and Calls data were not deleted even if the request was valid

### Cognigy Voice Gateway

**Released** December 12, 2023

#### Features

* Added STT latency visualization to the Voice Gateway Self-Service Portal
* Added support for `SIPREC INVITE` and the capability to choose a default `SIPREC` application to the Voice Gateway Self-Service Portal
* Added support for Outbound call tracing

#### Bug Fixes

* Fixed the issue where recent call details crashed when MP3 was not available
* Fixed the issue where selecting a Custom Speech vendor was causing the Application form page to crash

### Infrastructure

#### Cognigy + Insights

The `jobSyncKnowledgeData` is a new cron job that has been introduced to sync Knowledge AI data from MongoDB to Weaviate. This job is disabled by default and can be enabled on demand to allow for the restoration or repair of data in Weaviate. The job is applicable only when running on an environment where Knowledge AI is enabled.

`Repair Mode` - enabling the cron job will sync the data from MongoDB to Weaviate during the configured schedule.

```yaml theme={null}
jobSyncKnowledgeData:
  enabled: true
  schedule: "0 1 * * *" # every day at 1 AM
```

`Restore Mode` - the job can be configured to restore the data in Weaviate after a complete cleanup of Weaviate by adding the following `JOB_ARGUMENTS` configuration to the `extraEnvVars` variable in the` values.yaml` file.

```yaml theme={null}
jobSyncKnowledgeData:
  enabled: true
  extraEnvVars:
    - name: JOB_ARGUMENTS
      value: "eyJhY3Rpb24iOiJyZXN0b3JlVmVjdG9yREIiLCJvcmdhbmlzYXRpb25JZHMiOltdLCJwcm9qZWN0SWRzIjpbXSwic3RvcmVJZHMiOltdLCJ0YXNrSWQiOiIiLCJ0cmFjZUlkIjoidHJhY2UtcmVzdG9yZVZlY3RvckRCIiwidXBkYXRlQmF0Y2hTaXplIjo0MH0="
```

#### Cognigy Voice Gateway

* Updated the StatefulSet of `feature-server` and added `NODE_OPTIONS` to the environment variables. By default, the value is set to `--max-old-space-size=450`.

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy and Insights](/ai/administer/installation/version-compatibility-matrix)
* [Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 4.67 - Happy New Year 🎉❄️
Source: https://docs.cognigy.com/release-notes/4.67



<Warning>
  Due to the sunset of the [text-davinci-003](https://platform.openai.com/docs/deprecations) model by OpenAI on January 4, 2024, this model will no longer be available in Cognigy.
</Warning>

<Warning>
  Starting with Cognigy release 4.68.0, Weaviate v1.19.0 will no longer be supported. If you have a Cognigy on-premises installation and are utilizing the [Knowledge AI features](https://docs.cognigy.com/ai/empower/knowledge-ai/overview/), upgrade Weaviate to v1.22.5. For a smooth upgrade, refer to the [guide](https://docs.cognigy.com/ai/installation/migration/weaviate-upgrade-to-1.22.5/).
</Warning>

<Warning>
  As of Cognigy 4.65.0, the Single Replica Redis setup is deprecated and replaced with Redis in High Availability (HA) Mode. For on-premises installations, before upgrading to Cognigy 4.65, ensure that you migrate according to the [Redis: Single Replica to Redis HA Migration](https://docs.cognigy.com/ai/installation/migration/single-replica-redis-to-redis-ha-migration/).
</Warning>

## 4.67.1

### Cognigy

**Released** January 16, 2024

#### Bug Fixes

* Fixed the issue where the Flow would sometimes start at the wrong Node after a Go To Node

### Cognigy Voice Gateway

**Released** January 10, 2024

#### Bug Fixes

* Fixed the issue where the recognizer label was not correctly identified, potentially leading to the use of incorrect speech credentials
* Reverted changes made in version 4.67.0 for the Label list on the Application page

### Infrastructure

#### Cognigy Voice Gateway

* Add variables to the StatefulSet `sbc-rtp` to control the start and end ports for RTP:
  * `.Values.rtpEngine.startPort` (default=40000)
  * `.Values.rtpEngine.endPort` (default=60000)
* Add an environment variable to the Drachtio server container in the feature-server StatefulSet and the `sbc-sip` StatefulSet to enable support for best-effort TLS:
  * `.Values.drachtio.supportBestEffortTls` (default=false)

## 4.67.0

### Cognigy

**Released** January 9, 2024

#### Features

* Added the [Generic Settings](/ai/agents/deploy/endpoint-reference/voice-gateway) section to the Voice Gateway Endpoint

#### Improvements

* Added the Referred By option to the Refer transfer type in both the Call Failover and Call Events (specifically for Transfer Dial Error and Transfer Refer Error events) sections within the Voice Gateway Endpoint
* Added the `nova-2` Deepgram STT model
* Added [ElevenLabs](/voice-gateway/references/tts-and-stt-vendors) as a Text-to-Speech (TTS) provider
* Improved the Voice Gateway Endpoint by adding a setting to trim `input.data` for better transcripts
* Added support to import the [secondary locale of the package](/ai/platform-features/packages) as the primary locale of the agent

#### Bug Fixes

* Fixed the memory leak issue in `service-nlp-ner` by replacing the axios package
* Fixed the issue where LLM prompts in chat mode could fail if inputs contained no text
* Fixed the issue where there were multiple entries in the Intent Trainer for the same Flow and input text
* Fixed the issue where the UX of filter selection for the Case Node was misleading
* Fixed the issue where the current year was not displayed in the Billable Conversations filter on the Organization Details page

### Cognigy Insights

**Released** January 9, 2024

#### Improvements

* Improved the UI of the Insights dashboards
* Changed the dashboards' width to use all the available space on the screen

#### Bug Fixes

* Fixed the issue where the maximum message count was incorrect when the first digit is smaller than the minimum message count

### Cognigy Live Agent

#### Features

* Added the capability for human agents to create draft messages. If a [reply](/live-agent/conversation/send-reply#send-replies-to-users) or a [private](/live-agent/conversation/send-reply#send-replies-to-other-human-agents) message is composed but not sent immediately, the message will be saved for 24 hours

#### Improvements

* Renamed Live Agent Assist Bot to [Agent Copilot Bot](/live-agent/assistants/agent-copilot-bot)

### Cognigy Voice Gateway

**Released** January 9, 2024

#### Bug Fixes

* Fixed the issue where the default label in the UI was not applied
* Fixed the issue where the selected language did not promptly map to the corresponding voices

### Infrastructure

#### Cognigy Voice Gateway

* Updated the `helm.sh/hook` annotations for the following `Secret` resources to `"helm.sh/hook": "pre-install,pre-upgrade"`:
  * voice-gateway-call-recording-credentials
  * voice-gateway-drachtio-secret
  * voice-gateway-jwt
  * voice-gateway-mysql-password
* Remove unused `replicaCount` from `spec.Values` for:
  * rtpEngine
  * rtpEngineSidecar
  * sbcSipSidecar
* With the release of 4.65.0, we introduced `StatefulSets` for `sbc-sip` and `sbc-rtp`. We recommend enabling `StatefulMode` as it improves the load balancing of traffic between the replicas.
  * The required changes in values.yaml:
    * `spec.Values.sbc.sip.useStatefulSet: true`
    * `spec.Values.sbc.rtp.useStatefulSet: true`
  ```yaml theme={null}
  sbc:
    sip:      
      useStatefulSet: true
    rtp:      
      useStatefulSet: true
  ```
* With the release of 4.65.0, we introduced Redis High Availability. We recommend enabling `redisHa`:
  * The required changes in values.yaml:
  ```yaml theme={null}
  cognigyEnv:
    FEATURE_REDIS_SENTINEL_MODE_ENABLED: "true"
    REDIS_SENTINEL_MASTER_NAME: "mymaster"
    REDIS_SENTINEL_HOST: "redis-ha-headless"
    REDIS_SENTINEL_PORT: "26379"
    
  redisHa:
    enabled: true
  metrics:
    serviceMonitor:
      enabled: true
  ```
* With the release of 4.64.0, we disabled Grafana via a feature flag since we are using `cognigy-monitoring-stack`, which already includes Grafana. In the current release, we are removing all Grafana-related resources from the Helm charts:
  * Service: `grafana/service.yaml`
    * Deployment: `grafana/deployment.yaml`
    * ConfigMap: `grafana/config-map.yaml`
    * StorageClasses:
      * `templates/gcp/grafana/storage-class.yaml`
      * `templates/azure/grafana/storage-class.yaml`
      * `templates/aws/grafana/storage-class.yaml`

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 4.68
Source: https://docs.cognigy.com/release-notes/4.68



<Warning>
  With the 4.71 release, DNS names for Agent Copilot will change. If you have an on-premises installation, refer to the [migration guide](https://docs.cognigy.com/agent-copilot/installation/migration/agent-copilot-new-dns-names-migration/).
</Warning>

<Warning>
  As of Cognigy 4.69.0, the `topK.order` property of Knowledge Chunks in the search result generated by [Search Extract Output](/ai/agents/develop/node-reference/other-nodes/search-extract-output) Nodes is deprecated.
</Warning>

<Warning>
  Due to the sunset of the [text-davinci-003](https://platform.openai.com/docs/deprecations) model by OpenAI on January 4, 2024, this model will no longer be available in Cognigy.
</Warning>

<Warning>
  Starting with Cognigy release 4.68.0, Weaviate v1.19.0 will no longer be supported. If you have a Cognigy on-premises installation and are utilizing the [Knowledge AI features](https://docs.cognigy.com/ai/empower/knowledge-ai/overview/), upgrade Weaviate to v1.22.5. For a smooth upgrade, refer to the [guide](https://docs.cognigy.com/ai/installation/migration/weaviate-upgrade-to-1.22.5/).
</Warning>

## 4.68.2

### Cognigy

**Released** January 29, 2024

#### Bug Fixes

* Fixed the issue where Web Page Import consistently resulted in the error `Could not open web page at...`

### Infrastructure

**Released** February 5, 2024

#### Cognigy Voice Gateway

* Parameterized the `drachtio-certs` secret name in the `sbc-sip` DaemonSet and StatefulSet via `.Values.sbc.sip.certSecretName`

## 4.68.1

### Infrastructure

**Released** January 31, 2024

#### Cognigy Voice Gateway

* Added a new environment variable to the `sbc-outbound` deployment to support best-effort TLS. By default, it is disabled via `.Values.sbcOutbound.supportBestEffortTls: false`

### Cognigy

**Released** January 23, 2024

#### Bug Fixes

* Fixed the issue where the last message in LLM streams would sometimes not be displayed
* Fixed the issue where fetching trainer records failed when removing duplicates

## 4.68.0

### Cognigy

**Released** January 23, 2024

#### Features

* Added the [Web Page](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/web-page) source type for Knowledge AI. Now you can use web pages for creating Knowledge Sources
* Agent Copilot can now be directly embedded in the [Salesforce Agent Console](https://github.com/Cognigy/salesforce-integrations)

#### Improvements

* Improved by exposing transcript in [Code Nodes](/ai/for-developers/code/overview) as `lastConversationEntries`
* Improved by adding Knowledge AI resources into [Packages](/ai/platform-features/packages)
* Removed the Azure Deployment ID and advanced configuration of the Microsoft Speech provider from the Generic and Voice Gateway Nodes, as these settings can now be configured in the Voice Gateway Self-Service Portal

#### Bug Fixes

* Fixed the issue where the log level would not be applied correctly
* Fixed the issue where the Flow would sometimes start at the wrong Node after a Go To Node
* Added error details when uploading a Lexicon with a synonym that already exists
* Fixed the issue where the task manager did not display the correct number of tasks when there were child tasks
* Fixed the issue where Knowledge tile queries couldn't be submitted by pressing Enter
* Fixed the issue where `undefined` was displayed as the source in the Agent Copilot Knowledge tile
* Fixed the issue where Handover settings for Salesforce were not validated and forwarded properly
* Fixed the issue where TTS and STT parameters in the Set Session Config Node could not be cleared by adding the Default option

### Cognigy Insights

**Released** January 23, 2024

#### Improvements

* Improved by opening Insights in a new tab rather than the current one

### Cognigy Live Agent

**Released** January 23, 2024

#### Bug Fixes

* Fixed the issue in the Live Agent UI where the old Agent Assist name was used instead of Agent Copilot
* Fixed the issue of missing translations for the duplicated Canned Responses error
* Fixed the issue where the content of Agent Copilot's Next Action and Knowledge Tiles would replace the existing message in the reply box instead of appending to it

### Cognigy Voice Gateway

**Released** January 23, 2024

#### Bug Fixes

* Fixed the issue where the hang-up custom headers were not working
* Fixed the issue where the Label field of the speech provider on the Application page was incorrectly applied

### Infrastructure

#### Agent Copilot

Added a [migration guide](/agent-copilot/installation/migration/agent-copilot-new-dns-names-migration) for DNS name changes for Agent Copilot services.

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 4.69
Source: https://docs.cognigy.com/release-notes/4.69



<Warning>
  With the 4.71 release, DNS names for Agent Copilot will change. If you have an on-premises installation, refer to the [migration guide](https://docs.cognigy.com/agent-copilot/installation/migration/agent-copilot-new-dns-names-migration/).
</Warning>

<Warning>
  As of Cognigy 4.69.0, the `topK.order` property of Knowledge Chunks in the search result generated by [Search Extract Output](/ai/agents/develop/node-reference/other-nodes/search-extract-output) Nodes is deprecated.
</Warning>

<Warning>
  Due to the sunset of the [text-davinci-003](https://platform.openai.com/docs/deprecations) model by OpenAI on January 4, 2024, this model will no longer be available in Cognigy.
</Warning>

<Warning>
  Starting with Cognigy release 4.68.0, Weaviate v1.19.0 will no longer be supported. If you have a Cognigy on-premises installation and are utilizing the [Knowledge AI features](https://docs.cognigy.com/ai/empower/knowledge-ai/overview/), upgrade Weaviate to v1.22.5. For a smooth upgrade, refer to the [guide](https://docs.cognigy.com/ai/installation/migration/weaviate-upgrade-to-1.22.5/).
</Warning>

## 4.69.1

### Cognigy Live Agent

**Released** February 13, 2024

#### Bug Fixes

* Fixed the issue causing handover failures when a Question Node is present in the Flow

### Cognigy Voice Gateway

**Released** February 15, 2024

#### Bug Fixes

* Fixed the issue where outbound calls transferred via `REFER` did not include the custom headers

## 4.69.0

### Cognigy

**Released** February 6, 2024

#### Improvements

* Introduced [new global roles](/ai/administer/access/admin-center/access-control#before-you-begin), specifically designed for support purposes. These roles are: `fullSupportUser`, `basicSupportUser`, `projectAssigner`
* Improved by displaying a spinner in the UI while Knowledge Store deletion is in progress
* Improved by displaying a spinner in the UI while Knowledge Source deletion or disabling is in progress
* Added the option to set a custom background for the Agent Copilot workspace via the [Copilot: Set Grid](/ai/agents/develop/node-reference/agent-copilot/set-grid) Node
* Removed the `order` property from the topK results returned in Knowledge Search
* Improved the performance of Knowledge Chunk insertion and deletion operations
* Introduced new voices from Microsoft Azure and ElevenLabs speech providers

#### Bug Fixes

* Fixed the issue where some error messages are ignored in Kibana
* Fixed the issue where a couple of translations are missing in the Date Picker Node
* Fixed the issue where the bot was not answering for few minutes after deleting an active Snapshot and starting a new conversation
* Fixed the issue where texts are not wrapped properly in the Interaction Panel
* Fixed the issue where trainer records are not displayed for texts when the Flow executes through the Go To or Execute Flow Nodes with no Intent mapping
* Fixed the issue by adding missing locale-specific Yes/No confirmation static words
* Fixed the issue where Fetching Intent Trainer Record failed due to duplicate records awaiting deletion
* Fixed the issue where the AI Agent stops responding after deleting an active snapshot and starting a new chat
* Fixed the issue where text in Interaction Panels didn't wrap correctly. We made sure that text breaks properly by using the `wordBreak: break-word` property
* Fixed the issue where the `TRANSFER_DIAL_SUCCESS` call event occurred at the end of the call instead of at the time of the transfer

### Cognigy Insights

**Released** February 6, 2024

#### Improvements

* Improved by adding a scroll-up button to the dashboards
* Improved by updating translation strings

### Cognigy Live Agent

**Released** February 6, 2024

#### Bug Fixes

* Fixed the issue where links in Agent Copilot bot messages were rendered in the same color as the background
* Fixed the issue where empty pre-chat details could cause the Salesforce integration to behave incorrectly

### Cognigy Voice Gateway

**Released** February 6, 2024

#### Features

* Added the option for [failover in cases when the Cognigy endpoint is unavailable](/voice-gateway/webapp/applications#call-forwarding)
* Added the option to [download a PCAP and check Register status](/voice-gateway/webapp/carriers#outbound-authentication) on the Carrier page

#### Improvements

* Improved Text-To-Speech performance by implementing pre-caching capabilities. This improvement allows the voice agent to receive and process audio data in parallel, leading to a smoother and more responsive speech experience
* Added logic that considers DTMF signals when they are used in combination with User Input Timeout
* Introduced new voices from Microsoft Azure and ElevenLabs speech providers

#### Bug Fixes

* Fixed the issue where `organisationId` was not added to the redirect URL when using Cognigy AI login
* Fixed the issue where the timer start of the gather verb was applied incorrectly for the Deepgram Speech provider, causing long prompts to be interrupted
* Fixed the issue where a phone number saved under an Account without an Application selected did not have a Service Provider SID, causing the phone number to return a `404` response

### Infrastructure

#### Cognigy Voice Gateway

* Added a new environment variable to the `feature-server` StatefulSet. By default, it is enabled via `.Values.featureServer.preCacheAudio: true`

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 4.70
Source: https://docs.cognigy.com/release-notes/4.70



<Warning>
  With the 4.71 release, DNS names for Agent Copilot will change. If you have an on-premises installation, refer to the [migration guide](https://docs.cognigy.com/agent-copilot/installation/migration/agent-copilot-new-dns-names-migration/).
</Warning>

<Warning>
  As of Cognigy 4.69.0, the `topK.order` property of Knowledge Chunks in the search result generated by [Search Extract Output](/ai/agents/develop/node-reference/other-nodes/search-extract-output) Nodes is deprecated.
</Warning>

<Warning>
  Due to the sunset of the [text-davinci-003](https://platform.openai.com/docs/deprecations) model by OpenAI on January 4, 2024, this model will no longer be available in Cognigy.
</Warning>

## 4.70.0

### Cognigy

**Released** February 20, 2024

#### Improvements

* Added the Custom URL field to [Aleph Alpha LLM configuration](/ai/agents/develop/gen-ai-and-llms/providers/aleph-alpha#add-models)
* Made package uploads more user-friendly by keeping extracted content when navigating during the upload process
* Improved the [LLM Prompt](/ai/agents/develop/node-reference/service/llm-prompt) Node by adding JSON Mode, detailed results storage, stream copy storage, seed parameters, and custom options
* Improved by [forwarding custom attributes](/ai/escalate/handover-reference/genesys-cloud-open-messaging), including queue, languages, skills, and manually defined attributes, into the Genesys inbound message flow. This data can be retrieved by using the [Get Participant Data action](https://help.mypurecloud.com/articles/get-participant-data-action/) in the Genesys flow design
* Added [Deepgram](/voice-gateway/references/tts-and-stt-vendors) as a Text-to-Speech (TTS) provider

#### Bug Fixes

* Fixed the issue where the intent quality indicator would display `This intent has some overlap with Intent 'null'`
* Fixed the issue where Flow deletion was failing due to a transaction error
* Fixed the agent assist flow and the agent assist bot name to align with the new name Agent Copilot
* Fixed the issue with double submission and inconsistent template-changing behavior in xApps
* Fixed the issue where Voice Gateway incorrectly assigned the user ID when placing outbound calls
* Fixed the issue where the selected voice in the TTS settings of an old Set Session Config Node was missing
* Fixed the issue where selecting Execute Flow as an Action in the Call Events setup of the Voice Gateway Endpoint did not require Flow selection, leading to execution issues

### Cognigy Insights

**Released** February 20, 2024

#### Features

* Introduced a new [horizontal filters toolbar](/insights/global-filters)

#### Improvements

* Enhanced by opening the Cognigy product in a new tab rather than the current one
* Changed the shadow style of the Agent selector
* Redesigned the [Transcript Explorer](/insights/explorers/conversations)
* Renamed the Sessions Count tile to Total Sessions
* Changed the naming of the explorers section in the sidebar from Explorer to Explorers

#### Bug Fixes

* Fixed the issue where Power BI would show an error when a Cognigy Flow contained an empty Say Node

### Cognigy Live Agent

**Released** February 20, 2024

#### Improvements

* Removed the Live Agent onboarding screen that was previously used for creating accounts

#### Bug Fixes

* Fixed the issue where a user accessing Live Agent with a URL from a different Cognigy organization URL was not able to log in using the right account
* Fixed the issue where a file uploaded via Cognigy Webchat could not be deleted
* Fixed the issue where certain text elements within the resolve action dropdown were not translated correctly

### Cognigy Voice Gateway

**Released** February 20, 2024

#### Improvements

* Added [Deepgram](/voice-gateway/references/tts-and-stt-vendors) as a Text-to-Speech (TTS) provider

### Infrastructure

#### Cognigy Voice Gateway

* Added new argument to `drachtio` container in the `sbc-sip` StatefulSet and DaemonSet to reject `REGISTER` without sip-realm. It can be controlled via `.Values.drachtio.rejectRegisterWithNoRealm`, by default, it is enabled.

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 4.71
Source: https://docs.cognigy.com/release-notes/4.71



<Danger>
  With Release 4.71.0, we have introduced an incorrect default behavior for how meta-data for Cognigy Handovers are cached. This leads to flaky behavior in Handover sessions for Cognigy in all versions of our 4.71 and 4.72 lineup. Immediately update to the newest patch versions or set the configuration below:

  ```yaml theme={null}
  cognigyEnv:
  LRU_CACHE_ENABLED: "false"
  ```
</Danger>

<Warning>
  With the 4.71 release, DNS names for Agent Copilot will change. If you have an on-premises installation, refer to the [migration guide](https://docs.cognigy.com/agent-copilot/installation/migration/agent-copilot-new-dns-names-migration/).
</Warning>

## 4.71.4

### Cognigy

<Accordion title="Cognigy Handover problem solved in 4.71.4">
  This version solves an issue with an incorrect caching behavior of Cognigy Handovers by deactivating the cache by default. If you are upgrading to this version, the following configuration, which was necessary in the older version of 4.71 is **NOT necessary**:

  ```yaml theme={null}
  cognigyEnv:
      LRU_CACHE_ENABLED: "false"
  ```

  **Released** April 5, 2024
</Accordion>

### Bug Fixes

* Deactivated the default meta-data caching option for Cognigy Handovers (LRU\_CACHE\_ENABLED: false)

### Cognigy Live Agent

**Released** April 4, 2024

#### Bug Fixes

* Fixed the issue where human agents were unable to log in to Live Agent under a high load when automatic assignment was enabled

## 4.71.3

### Cognigy

<Danger>
  With Release 4.71.0, we have introduced an incorrect default behavior for how meta-data for Cognigy Handovers are cached. This leads to flaky behavior in Handover sessions for Cognigy in all versions of our 4.71 and 4.72 lineup. Immediately update to the newest patch versions or set the configuration below:

  ```yaml theme={null}
  cognigyEnv:
  LRU_CACHE_ENABLED: "false"
  ```
</Danger>

**Released** March 27, 2024

#### Bug Fixes

* Fixed the issue where messages were randomly missing during a handover to the RingCentral Engage provider

### Infrastructure

**Released** March 13, 2024

#### Cognigy Live Agent

* Fixed another issue where the `existingSecretKey` parameter within the `cognigyOAuth` block would not be applied

## 4.71.2

### Cognigy

<Danger>
  With Release 4.71.0, we have introduced an incorrect default behavior for how meta-data for Cognigy Handovers are cached. This leads to flaky behavior in Handover sessions for Cognigy in all versions of our 4.71 and 4.72 lineup. Immediately update to the newest patch versions or set the configuration below:

  ```yaml theme={null}
  cognigyEnv:
  LRU_CACHE_ENABLED: "false"
  ```
</Danger>

**Released** March 14, 2024

#### Bug Fixes

* Fixed the issue where handover requests were failing without displaying any errors in the logs

### Infrastructure

**Released** March 13, 2024

#### Cognigy Live Agent

* Fixed the issue where the `existingSecretKey` parameter within the `cognigyOAuth` block would not be applied

## 4.71.1

### Cognigy

<Danger>
  With Release 4.71.0, we have introduced an incorrect default behavior for how meta-data for Cognigy Handovers are cached. This leads to flaky behavior in Handover sessions for Cognigy in all versions of our 4.71 and 4.72 lineup. Immediately update to the newest patch versions or set the configuration below:

  ```yaml theme={null}
  cognigyEnv:
  LRU_CACHE_ENABLED: "false"
  ```
</Danger>

**Released** March 8, 2024

#### Bug Fixes

* Fixed the issue where Portuguese dates following the `Junho de 93` pattern were not recognized correctly
* Fixed the issue where it is not possible to use a Custom URL for Azure LLM provider with domains other than `azure.com`
* Fixed the issue where the Live Agent handover was failing

### Cognigy Live Agent

**Released** March 8, 2024

#### Bug Fixes

* Fixed the issue where push notifications were not working

## 4.71.0

### Cognigy

<Danger>
  With Release 4.71.0, we have introduced an incorrect default behavior for how meta-data for Cognigy Handovers are cached. This leads to flaky behavior in Handover sessions for Cognigy in all versions of our 4.71 and 4.72 lineup. Immediately update to the newest patch versions or set the configuration below:

  ```yaml theme={null}
  cognigyEnv:
  LRU_CACHE_ENABLED: "false"
  ```
</Danger>

**Released** March 5, 2024

#### Improvements

* Improved the Package Upload capability to resume on connection failure
* Improved the default password policy for the organization to include a minimum of 1 special character and require a minimum length of 12 characters
* Added new voices for text-to-speech for the Deepgram provider

#### Bug Fixes

* Fixed the issue where models are not marked outdated after changes to case-sensitive Intent mapping
* Fixed the issue where outdated NLU of non-primary locale is not reflected in the Flow list
* Fixed the issue where several copies of the same Knowledge Source were uploaded by clicking the Next button multiple times in the Knowledge AI wizard
* Fixed the endpoint name for Agent Copilot
* Fixed the issue where entryId validation was missing when adding Keyphrases or Slots to Lexicon entries
* Fixed the issue of occasional failure in trained model invalidation

### Cognigy Insights

**Released** March 5, 2024

#### Improvements

* Moved the locale filters back to the additional filters panel
* Improved by hiding the grain label in the widget's detailed view
* Removed the redundant indent from the Transcript Explorer details
* Changed the UI of the filters panel

#### Bug Fixes

* Fixed the issue where the header of the Transcript Explorer was overlapping table content when transcript details were open
* Fixed the issue where `service-analytics-conversation` was logging `Validation failed. Missing field 'organisationId'`
* Fixed the issue where the first and last message timestamps in the Transcript Explorer details weren't adjusted to the current time zone

### Cognigy Live Agent

**Released** March 5, 2024

#### Bug Fixes

* Fixed the issue where the line break was not correctly encoded in the Reply to bot section
* Fixed the issue preventing Live Agent upgrade with a non-root database user

### Cognigy Voice Gateway

**Released** March 5, 2024

#### Features

* Implemented [streaming](/voice-gateway/streaming) capability with the ElevenLabs Speech provider

#### Bug Fixes

* Fixed the issue where an error occurred during a `say` task with pre-cached audio using Deepgram
* Fixed the issue where the Transcribe feature was not working

### Infrastructure

#### Cognigy + Insights

If you are not using `cognigyLiveAgent.existingSecret` in your `values` local file, then you need to add the `helm.sh/resource-policy: keep` annotation to the `cognigy-live-agent-credentials` secret.
You can do this by using the following command. Replace `<namespace>` with the namespace where Cognigy is deployed.

```bash theme={null}
kubectl annotate -n=<namespace> secret cognigy-live-agent-credentials helm.sh/resource-policy=keep
```

#### Cognigy Voice Gateway

* Change the default value of `.Values.api.enablePrepareCognigyData` to `true`.
  * This change will modify the startup behavior of the API server.
  * The `admin` user will now have a randomly generated password instead of `admin`.
  * A user named 'cognigy-ai' is being created for Interaction Panel calls.
  * If not already present by that name, a new Service Provider named `IP Test Call Manager` will be created. Rename your existing service provider, used for the Interaction Panel, to `IP Test Call Manager`.
  * An API token is created and connected to the new service provider.
* Add Ingress for `Homer`
* Add Middleware `ipwhitelisthomer`
* Add Middleware `homerauth`
* Update `redis` to `7.2.4-debian-11-r3`
* Update `redis high availability` images:
  * `redis` image to `7.2.4-debian-11-r3`
  * `redis-sentinel` to `7.2.4-debian-11-r4`
  * `redis-exporter` to `1.56.0-debian-11-r1`
* Update `feature-server` Deployment and StatefulSet change the `OTEL` environment variable from `"1"` to `"true"`
* Update the `feature-server` Deployment and StatefulSet templates for TTS streaming from `value: {{ if .Values.featureServer.enableTtsElevenlabsStreaming }}{{ print 0 | quote }}{{ else }}{{ print 1 | quote }}{{ end }} to {{ not .Values.webapp.enableTtsElevenlabsStreaming | quote }}`
* Update the `api-server` Deployment TTS streaming template from `value: {{ if .Values.featureServer.enableTtsElevenlabsStreaming }}{{ print 0 | quote }}{{ else }}{{ print 1 | quote }}{{ end }}` to `{{ not .Values.webapp.enableTtsElevenlabsStreaming | quote }}`

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 4.72
Source: https://docs.cognigy.com/release-notes/4.72



<Danger>
  With Release 4.71.0, we have introduced an incorrect default behavior for how meta-data for Cognigy Handovers are cached. This leads to flaky behavior in Handover sessions for Cognigy in all versions of our 4.71 and 4.72 lineup. Immediately update to the newest patch versions or set the configuration below:

  ```yaml theme={null}
  cognigyEnv:
  LRU_CACHE_ENABLED: "false"
  ```
</Danger>

<Warning>
  As of April 2, 2024, several Microsoft Azure OpenAI API versions are deprecated. For more information, read [Retiring soon](https://learn.microsoft.com/en-us/azure/ai-services/openai/api-version-deprecation#retiring-soon). Ensure that you do not use these versions. If you are using them, switch to the latest API versions before the retirement date.
</Warning>

<Warning>
  With the 4.71 release, DNS names for Agent Copilot will change. If you have an on-premises installation, refer to the [migration guide](https://docs.cognigy.com/agent-copilot/installation/migration/agent-copilot-new-dns-names-migration/).
</Warning>

## 4.72.4

### Cognigy

<Check>
  This version solves an issue with an incorrect caching behavior of Cognigy Handovers by deactivating the cache by default. If you are upgrading to this version, the following configuration, which was necessary in the older version of 4.72 is **NOT necessary**:

  ```yaml theme={null}
  cognigyEnv:
      LRU_CACHE_ENABLED: "false"
  ```
</Check>

**Released** April 5, 2024

### Bug Fixes

* Deactivated the default meta-data caching option for Cognigy Handovers (LRU\_CACHE\_ENABLED: false)

## 4.72.3

### Cognigy

<Danger>
  With Release 4.71.0, we have introduced an incorrect default behavior for how meta-data for Cognigy Handovers are cached. This leads to flaky behavior in Handover sessions for Cognigy in all versions of our 4.71 and 4.72 lineup. Immediately update to the newest patch versions or set the configuration below:

  ```yaml theme={null}
  cognigyEnv:
  LRU_CACHE_ENABLED: "false"
  ```
</Danger>

**Released** April 2, 2024

#### Bug Fixes

* Fixed the issue where the `TRANSFER_DIAL_SUCCESS` event was mistakenly sent at the beginning of a new outbound call initiated via REST API call
* Fixed the issue where the Set Activity Parameters of Voice Gateway were not applied correctly

### Voice Gateway

**Released** April 23, 2024

#### Bug Fixes

* Changed the image of `freeswitch-mrf` to `4.69.1`:
  * Fixed the issue where simultaneous `start transcribe` requests in AWS Polly could cause problems
  * Fixed the issue where the order of events for the `bsearch` algorithm was incorrect, potentially causing missing events

## 4.72.2

### Cognigy

<Danger>
  With Release 4.71.0, we have introduced an incorrect default behavior for how meta-data for Cognigy Handovers are cached. This leads to flaky behavior in Handover sessions for Cognigy in all versions of our 4.71 and 4.72 lineup. Immediately update to the newest patch versions or set the configuration below:

  ```yaml theme={null}
  cognigyEnv:
  LRU_CACHE_ENABLED: "false"
  ```
</Danger>

**Released** March 27, 2024

#### Bug Fixes

* Fixed the issue where call events for outbound calls were missing when they were performed without a parent call

### Cognigy Live Agent

**Released** April 8, 2024

#### Bug Fixes

* Fixed the issue where the Supervisor Overview Report always showed 0 conversations in the Created or Queued statuses

### Voice Gateway

**Released** March 26, 2024

#### Features

* The [ElevenLabs Streaming TTS](/voice-gateway/streaming) feature is not available. We have switched your settings to traditional TTS

### Infrastructure

#### Voice Gateway

* Updated the template for `feature-server` `StatefulSet` and `Deployment`, specifically for the container `freeswitch`:
  * Remove `- --log-level` from `args` to support the R70 image of freeSWITCH-mrf
  * Set `.Values.webapp.enableTtsElevenlabsStreaming: false`

## 4.72.1

### Cognigy

<Danger>
  With Release 4.71.0, we have introduced an incorrect default behavior for how meta-data for Cognigy Handovers are cached. This leads to flaky behavior in Handover sessions for Cognigy in all versions of our 4.71 and 4.72 lineup. Immediately update to the newest patch versions or set the configuration below:

  ```yaml theme={null}
  cognigyEnv:
  LRU_CACHE_ENABLED: "false"
  ```
</Danger>

**Released** March 26, 2024

#### Bug Fixes

* Fixed the issue where messages were randomly missing during a handover to the RingCentral Engage provider
* Fixed the issue where Agent Copilot tiles were showing the Copy button in Live Agent

### Cognigy Live Agent

**Released** April 4, 2024

#### Bug Fixes

* Fixed the issue where the activity message shows the wrong assignee based on availability

### Infrastructure

**Released** March 25, 2024

#### Voice Gateway

* Added the internal infrastructure changes to improve Voice Gateway reliability

## 4.72.0

### Cognigy

<Danger>
  With Release 4.71.0, we have introduced an incorrect default behavior for how meta-data for Cognigy Handovers are cached. This leads to flaky behavior in Handover sessions for Cognigy in all versions of our 4.71 and 4.72 lineup. Immediately update to the newest patch versions or set the configuration below:

  ```yaml theme={null}
  cognigyEnv:
  LRU_CACHE_ENABLED: "false"
  ```
</Danger>

**Released** March 20, 2024

#### Features

* Added [Webchat v3](/webchat/v3/overview) as a preview feature. This is a new improved version of Webchat with updated design and various settings for widget customization. We encourage you to try it out and provide us with feedback.

#### Improvements

* Implemented the [LLM Entity Extract](/ai/agents/develop/node-reference/other-nodes/llm-entity-extract) Node and added the [LLM Entity Extraction Options](/ai/agents/develop/node-reference/basic/question) to the Question Node
* Added the capability to activate the Copy button within the [Copilot: Next Action Tile](/ai/agents/develop/node-reference/agent-copilot/next-action-tile) and [Copilot: Knowledge Tile](/ai/agents/develop/node-reference/agent-copilot/knowledge-tile) for all handover providers except Live Agent. For Salesforce, ensure you update your Salesforce integrations using [this guide](https://github.com/Cognigy/salesforce-integrations?tab=readme-ov-file#update)
* Extended the functionality of the Copilot: Next Action Tile Node to include the [Knowledge Assist](/ai/agents/develop/node-reference/agent-copilot/next-action-tile) options
* Introduced the parent-child concept for Knowledge Source creation tasks
* Upgraded the vg-core package to version 0.1.64 with new Deepgram voices
* Updated Demo Webchat to version 2.54.0
* Improved by preventing the deletion of a Snapshot when it is associated with an Endpoint
* Improved by modifying field labels and descriptions for the Request Rating Node
* Enabled the [Agent Copilot Embedding URL](/ai/escalate/handover-reference/genesys-cloud-open-messaging) for the Genesys Open Message handover provider
* Modified the `service-handover` logs related to Salesforce handover to explicitly state the failure reason when the handover fails
* Added the capability to [select a handover provider](/ai/agents/deploy/endpoint-reference/voice-gateway#) for the Voice Gateway Endpoint. After handover initiation, a `gather` verb is sent to collect transcripts and forward them to the human agent
* Updated Deepgram text-to-speech voices

#### Bug Fixes

* Fixed the issue where Russian Date Slots were not recognized
* Fixed the issue where the Code Editor Context Menu was getting hidden in specific cases
* Fixed the issue where a Live Follow session would not stop following when the browser was reloaded or reopened after being closed
* Fixed the issue where the Empty Text With Data in the Intent Default Reply doubles `data` in response
* Fixed the issue where the Create new User button was displayed to unauthorized users
* Fixed the issue where the billable conversations chart showed incorrect dates for January 30–31 when the current month is February
* Fixed the issue where Flow rendering would crash due to certain invalid configurations
* Fixed the issue where multiple date slots separated by commas were not being recognized
* Fixed the issue where importing CSV files with uppercase file extensions as Lexicon did not work
* Fixed the issue where Knowledge AI source creation and ingestion tasks couldn't be canceled through the UI
* Fixed the issue where the `last edited` column displayed only the ID instead of the username for basic users
* Fixed the issue where base-role users could encounter a forbidden error when attempting to read the Endpoint or the Search Extract Output Node due to missing `read` permissions
* Fixed the issue with the API documentation for creating a Lexicon entry being incorrect
* Fixed the issue where some organization details, such as the organization name, were temporarily stored in Redis after the organization was deleted
* Fixed the issue where query parameters were missing from the indexSnippets API request
* Fixed the issue where the extension sidebar overlapped with the Interaction Panel
* Fixed the issue where the Speech-to-Text (STT) and Text-to-Speech (TTS) select fields in the Node editor were not disabled when no localization was added
* Fixed the issue where the list of languages in DeepL was outdated
* Fixed the issue where the Deepgram speech-to-text with the nova-2-phonecall model was incorrectly saved in the Set Session Config Node
* Fixed the issue where the Salesforce handover provider was failing after a few seconds

### Cognigy Live Agent

**Released** March 20, 2024

#### Improvements

* Added the UI element to [change the size of the conversation chat interface](/live-agent/conversation/other-operations#expand-or-collapse-the-chat-area)
* Added new activity messages to clarify the reasons for conversation assignment, such as changes in agent availability or queued conversations

#### Bug Fixes

* Fixed the issue where sidebar information would overlap content on smaller screens, primarily occurring on various Settings screens

### Cognigy Voice Gateway

**Released** March 20, 2024

#### Improvements

* Updated Deepgram text-to-speech voices

#### Bug Fixes

* Fixed the issue preventing the deletion of an application

### Infrastructure

#### Cognigy + Insights

Added the `FEATURE_ENABLE_HANDOVER_SETTINGS_FOR_VG2` environment variable to control visibility of Handover settings in the VG endpoint. It is enabled by default.

#### Voice Gateway

* Added an environment variable `DISABLE_RATE_LIMIT` to the `api-server` deployment. It can be enabled via `.Values.api.disableRateLimit: true`.
* Added a new environment variable, `JAMBONZ_TMP_CACHE_FOLDER`, to `freeswitch-modules` to specify the directory for storing cached streaming files.
* Changed the default folder path for ElevenLabs streaming text-to-speech files to `tts-cache-files`, which is displayed in the Tracing.

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 4.73
Source: https://docs.cognigy.com/release-notes/4.73



<Warning>
  As of April 2, 2024, several Microsoft Azure OpenAI API versions are deprecated. For more information, read [Retiring soon](https://learn.microsoft.com/en-us/azure/ai-services/openai/api-version-deprecation#retiring-soon). Ensure that you do not use these versions. If you are using them, switch to the latest API versions before the retirement date.
</Warning>

<Warning>
  With the 4.71 release, DNS names for Agent Copilot will change. If you have an on-premises installation, refer to the [migration guide](https://docs.cognigy.com/agent-copilot/installation/migration/agent-copilot-new-dns-names-migration/).
</Warning>

## 4.73.2

### Cognigy Voice Gateway

**Released** May 7, 2024

#### Bug Fixes

* Fixed the issue where enabling Barge In could cause incorrect transcriptions to arrive later in a call
* Removed trailing silence trimming from Microsoft TTS to fix delays during calls. This trimming is no longer needed in recent versions and might cause issues
* Disabled the precache audio feature due to potential issues it may cause

### Infrastructure

#### Cognigy Voice Gateway

* Disabled the environment variable `VG_TTS_TRIM_SILENCE` by default: `.Values.featureServer.trimTts: false`
* Changed the default value of `.Values.featureServer.preCacheAudio` to `false`

## 4.73.1

### Cognigy

**Released** April 11, 2024

#### Improvements

* Improved stability by redesigning the logic for the Salesforce handover provider

### Cognigy Live Agent

**Released** April 8, 2024

#### Bug Fixes

* Fixed the issue where the Supervisor Overview Report always showed 0 conversations in the Created or Queued statuses

### Cognigy Voice Gateway

**Released** April 11, 2024

#### Bug Fixes

* Fixed race condition in AWS Polly transcription: resolved an issue where concurrent `start transcribe` requests could lead to unexpected behavior
* Fixed the issue where the order of events for the `bsearch` algorithm was incorrect, potentially causing data corruption on startup

### Infrastructure

**Released** April 10, 2024

#### Voice Gateway

* Added the environment variable `ANCHOR_MEDIA_ALWAYS` to the `feature-server` StatefulSet and Deployment. It can be enabled via `.Values.featureServer.anchorMediaAlways: true`. Default is `false`
* Changed the image of `freeswitch-mrf` to `4.69.1`

## 4.73.0

### Cognigy

**Released** April 2, 2024

#### Improvements

* Enabled [color selection](/webchat/v3/configuration#colors) for multiple new colors in the Webchat Layout section within Webchat v3 Endpoints
* Added the option to [specify other logos and titles](/webchat/v3/configuration#webchat-layout) for the message header in the Webchat Layout section within Webchat v3 Endpoints
* Implemented the option to create a [Web Page](/ai/agents/develop/knowledge-ai/knowledge-source/text-formats/web-page) source type using the Knowledge AI wizard
* Added the [Error Handling](/ai/agents/develop/node-reference/service/llm-prompt) section to the LLM Prompt Node
* Added the capability to use [new Anthropic models](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature)
* Added the capability to override the model version in [Custom Model Options](/ai/agents/develop/node-reference/service/llm-prompt) within the LLM Prompt Node. For example, you can overwrite the Legacy Anthropic model with a new one
* Improved the tooltip description for the [STT Hints parameter](/ai/agents/develop/node-reference/voice/voice-gateway/parameter-details) in the Set Session Config Node
* Improved the translation of Agent Copilot sentiment tiles based on the Flow execution language
* Improved the HTTP Request Node by adding [Timeout](/ai/agents/develop/node-reference/service/http-request) and [Retry Attempts](/ai/agents/develop/node-reference/service/http-request) fields
* Added the UI Preferences section to the Agent Copilot Tile Nodes. Now you can modify fonts for the Agent Copilot [Plug-and-Play](/agent-copilot/configure/nodes) widgets
* Updated Demo Webchat v2 to version 2.59.1

#### Bug Fixes

* Fixed the issue where the Chat mode in the LLM Prompt Node would cause a trim error when using Voice Gateway
* Fixed the issue where users were frequently logged out of Cognigy when utilizing OIDC (OpenID Connect) Single Sign-On (SSO) authentication
* Fixed the issue where JWT was exposed in the URL during WebSocket connections for the Interaction Panel
* Fixed the issue where the configured background image for Demo Webchat v3 was not displayed
* Fixed the issue where the grain property was sometimes missing in English date Slots
* Fixed the issue where English Date slots with the pattern `from 1:00 PM for three hours` were not recognized correctly
* Fixed the issue where the Copilot: Knowledge Tile was not rendering in some scenarios
* Fixed the issue where `Hi` was mistakenly displayed in the initial render of the Copilot: Knowledge Tile
* Fixed the issue where channel outputs could become corrupted when the 'Always remove No-Translation Markers' setting was enabled in Azure Bot Services
* Fixed the issue where the Expiration Value fields in the Management UI allowed values greater than the maximum value
* Fixed the issue where the `entrypoint` and `disabled` properties were ignored during the export and import of Intents
* Fixed the issue where the style configuration for Adaptive Cards was not displayed correctly
* Fixed the issue where the form data from the Agent Copilot adaptive card was occasionally not sent to Cognigy as a postback action
* Fixed the issue where handover was inconsistent and causing problems due to internal cache enabled

### Cognigy Insights

**Released** April 2, 2024

#### Bug Fixes

* Fixed an issue where the Contains Step filter value in the Transcript Explorer was empty in some cases

### Cognigy Live Agent

**Released** April 2, 2024

#### Improvements

* Implemented customization of Live Agent user session duration through the `AUTH_TOKEN_LIFESPAN` environment variable

#### Bug Fixes

* Fixed the issue where a human agent was unable to interact with Agent Copilot inlined messages until the actual message was marked as read
* Fixed the issue where forwarded Agent Copilot messages were not cleared from the reply box when the conversation changed
* Fixed the issue where attachments from WhatsApp were not listed on the Attachments page
* Fixed the issue where the Auto Resolve Action setting, with a value set to less than 5 minutes, was executed later than expected
* Fixed the issue where human agents were unable to log in to Live Agent under a high load when automatic assignment was enabled
* Fixed the issue where the activity message displayed the wrong assignee based on availability

### Infrastructure

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 4.74
Source: https://docs.cognigy.com/release-notes/4.74



<Danger>
  Starting from Cognigy 4.75.0, the `URLToken` parameter is mandatory for the Socket.IO and Webchat Endpoints. This parameter must be included along with `userId` and `sessionId` in the connection parameters. This functionality is automatically handled in `@cognigy/socket-client`, beginning with 4.5.3, and in Webchat v2.36.0. Ensure that you upgrade if you are using earlier versions.
</Danger>

<Warning>
  In Cognigy 4.79.0, the Cognigy Journeys feature will be removed.
</Warning>

<Warning>
  In Cognigy 4.75.0, Weaviate will be removed as a vector database. If you have a Cognigy on-premises installation and are utilizing the [Knowledge AI features](https://docs.cognigy.com/ai/empower/knowledge-ai/overview/), switch to Qdrant before upgrading to version 4.75.0. For a smooth migration, follow the [provided guide](https://docs.cognigy.com/ai/installation/migration/from-weaviate-to-qdrant/).

  The migration has a side effect that alters the `distance` value. Decision-making in the Flow will be ineffective post-migration if distance is used. If you are using this value in your Flows, refer to [recommendations](https://docs.cognigy.com/ai/empower/knowledge-ai/distance-value-change-after-migration/) for configuring this value.
</Warning>

## 4.74.4

### Cognigy

**Released** May 17, 2024

#### Improvements

* Added the ability to [forward Genesys bot messages to the end user](/ai/escalate/handover-reference/genesys-cloud-open-messaging) while using the Genesys Cloud Open Messaging handover provider. These messages will be injected into the chat. To enable this feature for on-premises installations, use the `GENESYS_CLOUD_OM_HANDLE_BOT_MESSAGE: "true"` feature flag

## 4.74.3

### Cognigy Insights

**Released** April 24, 2024

#### Bug Fixes

* Implemented internal logging to monitor important security-related actions within the OData Endpoint

## 4.74.2

### Infrastructure

#### Cognigy Voice Gateway

* Added the `FEATURE_ENABLE_HANDOVER_SETTINGS_FOR_VG2` environment variable to control visibility of Handover settings in the Voice Gateway Endpoint. It is enabled by default.

## 4.74.1

### Cognigy

**Released** April 19, 2024

#### Bug Fixes

* Fixed the issue that caused an exception when using the `Skip if the result is already in Context of a Slot Filler` toggle

### Cognigy Live Agent

**Released** April 23, 2024

#### Bug Fixes

* Fixed the issue where the activity message shows the wrong assignee based on availability

### Cognigy Voice Gateway

**Released** May 7, 2024

#### Bug Fixes

* Fixed the issue where enabling Barge In could cause incorrect transcriptions to arrive later in a call
* Removed trailing silence trimming from Microsoft TTS to fix delays during calls. This trimming is no longer needed in recent versions and might cause issues
* Fixed the issue where Voice Gateway integrations would not behave correctly when the `FEATURE_DISABLE_RESET_FLOW` feature flag was enabled

### Infrastructure

#### Cognigy Voice Gateway

* Disabled the environment variable `VG_TTS_TRIM_SILENCE` by default: `.Values.featureServer.trimTts: false`

## 4.74.0

### Cognigy

**Released** April 16, 2024

#### Improvements

* Improved user experience for [creating LLMs](/ai/agents/develop/gen-ai-and-llms/llms#model-types) and also enabled custom models for the LLM providers
* Added support for [`text-embedding-3-large`](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature), which can be hosted on either OpenAI or Microsoft Azure. Now, you can use [these embeddings for NLU training and scoring](/ai/platform-features/nlu/external/nlu-connectors/overview)
* Enhanced by redesigning the Salesforce conversation polling for better stability
* Expanded the German localization in the product
* Made the API version field required for Azure OpenAI when creating models
* Improved by maintaining the connection open when the conversation disconnects for Genesys Cloud Guest Chat

#### Bug Fixes

* Fixed the issue where the updated chunk metadata value was not being displayed until the page was refreshed
* Fixed the issue by updating the regular expression to handle cases where a user enters a US Social Security Number (SSN) with more than 9 digits
* Fixed the issue of inconsistent handover and associated problems caused by the enabled internal cache
* Fixed the issue where users could create text intent records with empty strings
* Fixed the issue where the Genesys Cloud Chat handover would fail, causing a service restart due to an uncontrolled exception while creating messages with an empty body

### Cognigy Insights

**Released** April 16, 2024

#### Bug Fixes

* Fixed the issue where OData was returning numbers instead of strings in the LiveAgentEscalations collection

### Cognigy Live Agent

**Released** April 16, 2024

#### Improvements

* Added performance improvements to conversations and calculations for the Reports
* Improved Live Agent performance in high-load scenarios by reducing WebSocket traffic

#### Bug Fixes

* Fixed the issue where the Supervisor Overview Report always showed 0 conversations in the Created or Queued statuses
* Fixed the issue that caused the account's auto-resolve setting to display an activity message for resolved conversations multiple times. This behavior occurred specifically under a high load

### Cognigy Voice Gateway

**Released** April 16, 2024

#### Bug Fixes

* Disabled the `pre cache audio` feature due to potential issues it may cause
* Fixed the issue where users with the Service Provider role could create and delete resources of another Service Provider. Now, users with the Service Provider role can only create and delete resources within their own project in the Voice Gateway Self-Service Portal

### Infrastructure

#### Cognigy Voice Gateway

* Changed the default value of `.Values.featureServer.preCacheAudio` to `false`

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 4.75
Source: https://docs.cognigy.com/release-notes/4.75



<Danger>
  Starting from Cognigy 4.75.0, the `URLToken` parameter is mandatory for the Socket.IO and Webchat Endpoints. This parameter must be included along with `userId` and `sessionId` in the connection parameters. This functionality is automatically handled in `@cognigy/socket-client`, beginning with 4.5.3, and in Webchat v2.36.0. Ensure that you upgrade if you are using earlier versions.
</Danger>

<Warning>
  * The Cognigy Journeys feature will be removed in version 4.77.
  * The Any Slots feature will be removed in version 4.83. Instead, use the [LLM Entity Extraction](/ai/agents/develop/node-reference/other-nodes/llm-entity-extract) Node to identify Slots in the user input via LLMs.
</Warning>

<Warning>
  In Cognigy 4.75.0, Weaviate will be removed as a vector database. If you have a Cognigy on-premises installation and are utilizing the [Knowledge AI features](https://docs.cognigy.com/ai/empower/knowledge-ai/overview/), switch to Qdrant before upgrading to version 4.75.0. For a smooth migration, follow the [provided guide](https://docs.cognigy.com/ai/installation/migration/from-weaviate-to-qdrant/).

  The migration has a side effect that alters the `distance` value. Decision-making in the Flow will be ineffective post-migration if distance is used. If you are using this value in your Flows, refer to [recommendations](https://docs.cognigy.com/ai/empower/knowledge-ai/distance-value-change-after-migration/) for configuring this value.
</Warning>

## 4.75.4

### Live Agent

**Released** May 27, 2024

#### Bug Fixes

* Fixed the issue where the Agent Availability Node incorrectly returned 0 for the number of available agents even when at least one human agent was online

## 4.75.3

### Live Agent

**Released** May 27, 2024

#### Improvements

* Added internal improvements to enhance Live Agent performance

## 4.75.2

### Live Agent

**Released** May 24, 2024

#### Improvements

* Improved the speed of retrieving inbox collaborators, which also speeds up working with the `inbox_members` API

#### Bug Fixes

* Fixed the issue where human agents could not log in to Live Agent, and auto assignment was occurring when clicking Unassign in the Conversation Details
* Fixed the issue where profile and inbox avatars were not displayed

## 4.75.1

### Cognigy

**Released** May 17, 2024

#### Improvements

* Added the ability to [forward Genesys bot messages to the end user](/ai/escalate/handover-reference/genesys-cloud-open-messaging) while using the Genesys Cloud Open Messaging handover provider. These messages will be injected into the chat. To enable this feature for on-premises installations, use the `GENESYS_CLOUD_OM_HANDLE_BOT_MESSAGE: "true"` feature flag

### Insights

**Released** May 17, 2024

#### Bug Fixes

* Fixed the issue where Insights would crash when in the Endpoints filter 'Interaction Panel' was selected together with one additional Endpoint

### Live Agent

**Released** May 22, 2024

#### Improvements

* Improved by adding new activity messages about assignments within the conversation. These messages will appear when the selected team is changed, and when auto-assignment occurs as the conversation is created

#### Bug Fixes

* Fixed the issue where when the [Allow auto assignment inbox setting](https://docs.cognigy.com/live-agent/conversation/conversation-routing/automatic-mode/#automatic-assignment) is enabled with the [Consider conversation count](https://docs.cognigy.com/live-agent/conversation/conversation-routing/automatic-mode/#additional-automatic-assignment-parameters) setting checked conversations were assigned to agents who did not belong to the inbox

## 4.75.0

### Cognigy

**Released** April 30, 2024

#### Improvements

* Added the [xApp Overlay](/xApps/build/overlay) feature for Webchat v3
* Added the [`knowledgeAdmin`](/ai/administer/access/members#project-roles) project-level role that permits users to create, update, read, and delete the knowledge resources within a project
* Improved UX consistency for required fields in the Connection form
* Improved by adding a [default retry mechanism of 3 attempts](/ai/agents/develop/gen-ai-and-llms/llms#retry-mechanism) when contacting an LLM provider for the following Nodes and features:<br />
  Nodes: LLM Prompt, LLM Entity Extract, GPT Conversation, Search Extract Output.<br />
  Features: Question Node Slot mapping, Question Node LLM reprompting, Sentiment Analysis.<br />
  This retry mechanism can be controlled using the following environment variables: `GENERATIVE_AI_RETRY_OPTIONS_NUMBER_OF_RETRIES`, `GENERATIVE_AI_RETRY_OPTIONS_MIN_TIMEOUT`, `GENERATIVE_AI_RETRY_OPTIONS_MAX_TIMEOUT`, `GENERATIVE_AI_RETRY_OPTIONS_FACTOR`

#### Bug Fixes

* Fixed the issue where after making changes to a connection, the flow execution would continue using the old connection tokens. This would lead to errors and unexpected behavior, especially with the LLM prompt Node
* Fixed the issue of a 403 error message briefly appearing in the UI upon loading a page
* Fixed the issue where the NLU intents form reset button was clickable even when the form state was empty
* Fixed the issue where the Reprompt Method configuration for older question Nodes would not be processed correctly
* Fixed the issue where metadata was not being forwarded to the HTML and iframe tiles in Agent Copilot
* Fixed the issue where a Knowledge Source with web page import, which could have more than 100 chunks, could not be created

### Cognigy Insights

**Released** April 30, 2024

#### Improvements

* Removed [Interaction Panel and Playbook data](/insights/global-filters) from Insights by default and added a toggle to include it in the Endpoints filter

#### Bug Fixes

* Fixed the issue where the Transcript Details page was crashing due to an unexpected payload

### Cognigy Live Agent

**Released** April 30, 2024

#### Improvements

* Added the possibility of using custom sounds for message notification
* Added a configuration option to set a timespan after which the agent is automatically logged out if they are away or set to away or offline
* Added the capability to [change the conversation inbox individually and through bulk actions](/live-agent/conversation/assign-conversations#change-inboxes)
* Added a reporting events model to Live Agent OData to retrieve stats about deleted or unassigned conversations

#### Bug Fixes

* Fixed the issue where conversation reports were broken upon removing an inbox member
* Fixed the issue where users with the Supervisor role could gain API access to resources restricted to the Administrator role

### Cognigy Voice Gateway

**Released** April 30, 2024

#### Bug Fixes

* Fixed the issue where DTMF input was being triplicated in instances where Genesys and UDP were utilized
* Fixed the issue where enabling Barge In could cause incorrect transcriptions to arrive later in the call

### Infrastructure

#### Cognigy

##### Migration from Weaviate to Qdrant

The Weaviate vector database has been removed. If you have a Cognigy on-premises installation and are utilizing the [Knowledge AI features](/ai/agents/develop/knowledge-ai/overview), switch to Qdrant before upgrading to version 4.75.0. For a smooth migration, follow the [provided guide](/ai/administer/installation/migration/from-weaviate-to-qdrant).

The migration has a side effect that alters the `distance` value. Decision-making in the Flow will be ineffective post-migration if distance is used. If you are using this value in your Flows, refer to [recommendations](/ai/agents/develop/knowledge-ai/distance-value-change-after-migration) for configuring this value.

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 4.76
Source: https://docs.cognigy.com/release-notes/4.76



<Danger>
  Starting from Cognigy 4.75.0, the `URLToken` parameter is mandatory for the Socket.IO and Webchat Endpoints. This parameter must be included along with `userId` and `sessionId` in the connection parameters. This functionality is automatically handled in `@cognigy/socket-client`, beginning with 4.5.3, and in Webchat v2.36.0. Ensure that you upgrade if you are using earlier versions.
</Danger>

<Warning>
  * The Cognigy Journeys feature will be removed in version 4.77.
  * The Any Slots feature will be removed in version 4.83. Instead, use the [LLM Entity Extraction](/ai/agents/develop/node-reference/other-nodes/llm-entity-extract) Node to identify Slots in the user input via LLMs.
</Warning>

## 4.76.3

### Live Agent

**Released** May 27, 2024

#### Bug Fixes

* Fixed the issue where the Agent Availability Node incorrectly returned 0 for the number of available agents even when at least one human agent was online

### Infrastructure

#### Cognigy Voice Gateway

* Add `PRIVATE_IP_ONLY` to `feature-server` StatefulSet and Deployment and `sbc-sip` StatefulSet and DaemonSet. It can be controlled via `.Values.drachtio.privateIpOnly`. The default value is `false`.

## 4.76.2

### Cognigy

**Released** May 17, 2024

#### Improvements

* Added the ability to [forward Genesys bot messages to the end user](/ai/escalate/handover-reference/genesys-cloud-open-messaging) while using the Genesys Cloud Open Messaging handover provider. These messages will be injected into the chat. To enable this feature for on-premises installations, use the `GENESYS_CLOUD_OM_HANDLE_BOT_MESSAGE: "true"` feature flag

### Live Agent

**Released** May 24, 2024

#### Improvements

* Improved the speed of retrieving inbox collaborators, which also speeds up working with the `inbox_members` API

#### Bug Fixes

* Fixed the issue where human agents could not log in to Live Agent, and auto assignment was occurring when clicking Unassign in the Conversation Details
* Fixed the issue where profile and inbox avatars were not displayed

### Cognigy Voice Gateway

**Released** May 23, 2024

#### Bug Fixes

* Fixed the issue for a possible race condition using the `gather` verb
* Fixed the issue in `vg-drachtio-server` where a part of the entrypoint was missing, causing unresolved IP addresses in Azure clusters with a load balancer
* Improved the handling of TTS audio files in the cache by creating a new namespace based on their file extension

### Infrastructure

#### Cognigy Voice Gateway

* Added the `PRIVATE_VOIP_NETWORK_CIDR` environment variable to `sbc-inbound` and `sbc-outbound` deployments. This variable can be used for a private network (for example, `172.31.0.0/16`)
* Increased the `api-server` rate limit to 1000 within 4 minutes

## 4.76.1

### Cognigy

**Released** May 16, 2024

#### Bug Fixes

* Fixed the issue where text messages from the RingCentral Engage channel configuration were being transferred incorrectly in Handover transcripts

### Live Agent

**Released** May 22, 2024

#### Improvements

* Improved by adding new activity messages about assignments within the conversation. These messages will appear when the selected team is changed, and when auto-assignment occurs as the conversation is created

#### Bug Fixes

* Fixed the issue where when the [Allow auto assignment inbox setting](https://docs.cognigy.com/live-agent/conversation/conversation-routing/automatic-mode/#automatic-assignment) is enabled with the [Consider conversation count](https://docs.cognigy.com/live-agent/conversation/conversation-routing/automatic-mode/#additional-automatic-assignment-parameters) setting checked conversations were assigned to agents who did not belong to the inbox

## 4.76.0

### Cognigy

**Released** May 14, 2024

#### Improvements

* Added [Intent descriptions](/ai/platform-features/nlu/external/intent-reranking) to the `input.nlu.scores` object to improve intent ranking
* Updated the `@cognigy/extension-tools` package. Now, this package includes the latest features added to the Extension Tools framework
* Improved the Genesys OM handover provider to separately display end-user and AI Agent messages within Genesys
* Updated Demo Webchat to version 2.59.2
* Renamed Agent to Project in the Cognigy UI

#### Bug Fixes

* Fixed the issue where the host header for the Cognigy UI was not validated
* Fixed the issue where Slots were not matched for a period of time after an internal TTL expired
* Fixed the issue of the confusing user experience with Flow and Node selection in the Go To Node

### Cognigy Live Agent

**Released** May 14, 2024

#### Improvements

* Improved processing of queued conversations when reassignment logic is enabled for the inbox and the account has a maximum conversation limit enabled
* Improved by deactivating the agent auto logout by default

### Cognigy Voice Gateway

**Released** May 14, 2024

#### Improvements

* Updated `vg-core` to `0.1.70` adding new Microsoft STT languages
* Updated the base image to `node:18.20-alpine3.19`
* Improved by supporting Azure Speech Services with private endpoints

### Infrastructure

#### Cognigy Voice Gateway

* Add `ENABLE_ON_PREM_AZURE_SPEECH` to webapp deployment. It can be controlled via `.Values.webapp.enableOnPremisesAzureSpeech`. The default value is `false`

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 4.77
Source: https://docs.cognigy.com/release-notes/4.77



<Warning>
  * The Cognigy Journeys feature will be removed in Cognigy 4.79.0.
  * The Chatbase analytics integration will be removed in Cognigy 4.82.0. The Chatbase platform hasn't been supported since 2021.
  * The Any Slots feature will be removed in Cognigy 4.83.0. Instead, use the [LLM Entity Extraction](/ai/agents/develop/node-reference/other-nodes/llm-entity-extract) Node to identify Slots in the user input via LLMs.
  * The `facebook-bot-messenger` module within a Code Node will be removed in Cognigy 4.83.0.
</Warning>

## 4.77.1

### Infrastructure

#### Cognigy Voice Gateway

* Add `PRIVATE_IP_ONLY` to `feature-server` StatefulSet and Deployment and `sbc-sip` StatefulSet and DaemonSet. It can be controlled via `.Values.drachtio.privateIpOnly`. The default value is `false`

## 4.77.0

### Cognigy

**Released** May 28, 2024

#### Features

* The [Webchat v3 Beta](/webchat/v3/overview) is now available. This means that most of the errors from the Webchat v3 Preview have been fixed, and features from v2 are now included in v3 Beta.
  With this release, the 18-month countdown to the deprecation of Webchat v2 begins.
  During this time, only important security updates will be provided for v2, with no further feature additions

#### Improvements

* Added the [Google Gemini](/ai/agents/develop/gen-ai-and-llms/model-support-by-feature) provider for LLMs
* Added the Microsoft Agent Copilot extension to the Cognigy Marketplace
* Released Webchat v2.60
* Added translations for languages other than English
* Renamed the UI term `agent` to differentiate between a Project, an AI Agent, and a human agent
* Updated the OpenAPI documentation for the Cognigy API
* Added the display of the `Handover was canceled` message in the chat when the Genesys Open Messaging handover is terminated by the human agent or due to end user inactivity

#### Bug Fixes

* Fixed the issue where the handover status was not updating from `Queued` when the end user disconnected, despite the human agent replying first
* Fixed the issue where selecting different voices with labels using Activity Parameters on the Voice Gateway channel did not return the correct language
* Fixed the issue where the NLU V2 migration script encountered `KeyError('deepTraining')`
* Fixed the issue where the Say Node autocompletion for default channel structured content would always require an `_image` template to be configured
* Fixed the issue where the `Content Security Policy` HTTP header was missing on the UI
* Fixed the issue where upgrading the `mssql service-ai` package caused problems with trusting the SQL Server certificate
* Fixed the issue where fetching the call counter API timed out
* Fixed the incorrect type in `@cognigy/extension-tools` v0.16.1
* Fixed the issue showing the incorrect Agent Copilot favicon

### Cognigy Live Agent

**Released** May 28, 2024

#### Improvements

* Added content normalization to messages for better matching with Automation Rules, handling different user input variations

#### Bug Fixes

* Fixed the issue with the Live Agent OData endpoint. Previously, querying the conversation entity without using `$select` caused errors in PowerBI and Excel. Now, the `custom_attributes` and `additional_attribute` fields return as strings

### Cognigy Voice Gateway

**Released** May 28, 2024

#### Improvements

* Added [on-premises installation support](/voice-gateway/webapp/speech-services#on-premises-speech-services) for the following speech service providers: Microsoft Azure Speech Services, Deepgram, and Nuance

#### Bug Fixes

* Fixed the issue in `sbc-outbound` where non-audible calls occurred after declining a call. This issue happened when using TLS and multiple gateways for outbound calls

### Infrastructure

#### Cognigy Voice Gateway

* Add `DEEPGRAM_SPEECH_USE_SINGLE_CONNECTION` to the `feature-server` StatefulSet and Deployment for the container FreeSWITCH. It can be controlled via `.Values.freeswitch.deepgramSttUseSingleConnection`. The default value is `true`
* Add `DISABLE_HTTP2_FOR_TTS_STREAMING` to the `feature-server` StatefulSet and Deployment for the container freeSWITCH. It can be controlled via `.Values.freeswitch.disableHttp2ForTtsStreaming`. The default value is `true`
* Change the default value of `.Values.webapp.enableOnPremisesAzureSpeech` to `true`
* Change the default value of `.Values.webapp.enableOnPremisesDeepgram` to `true`
* Change the default value of `.Values.webapp.enableOnPremisesNuance` to `true`

#### Version Compatibility Matrix

The version compatibility matrix was updated for the following Cognigy products:

* [Cognigy, Insights, and Agent Copilot](/ai/administer/installation/version-compatibility-matrix)
* [Live Agent](/live-agent/installation/deployment/version-compatibility-matrix)
* [Cognigy Voice Gateway](/voice-gateway/installation/version-compatibility-matrix)


# 4.78
Source: https://docs.cognigy.com/release-notes/4.78



<Warning>
  * The Cognigy Journeys feature will be removed in Cognigy 4.79.0.
  * The Chatbase analytics integration will be removed in Cognigy 4.82.0. The Chatbase platform hasn't been supported since 2021.
  * The Any Slots feature will be removed in Cognigy 4.83.0. Instead, use the [LLM Entity Extraction](/ai/agents/develop/node-reference/other-nodes/llm-entity-extract) Node to identify Slots in the user input via LLMs.
  * The `facebook-bot-messenger` module within a Code Node will be removed in Cognigy 4.83.0.
</Warning>

## 4.78.3

### Infrastructure

**Released** July 11, 2024

#### Cognigy Voice Gateway

* Added a list of German voicemail phrases to the `feature-server` ConfigMap for Answering Machine Detection (AMD)

### Cognigy

**Released** June 26, 2024

#### Bug Fixes

* Fixed the issue where the analytics `inputText` and `inputData` fields would not be overwritten by the Overwrite Analytics Node or Code Node

## 4.78.2

### Cognigy Insights

**Released** June 24, 2024

#### Bug Fixes

* Fixed the issue in `service-insights-api` that caused it to fail due to the `ERR_HTTP_HEADERS_SENT` error thrown in the proxy middleware

### Cognigy Voice Gateway

**Released** June 24, 2024

#### Bug Fixes

* Fixed the issue where Google-based Speech-To-Text conversions used an incorrect default model, causing inaccuracies in transcriptions. The default model has been updated to `latest_short` to improve recognition accuracy

## 4.78.1

### Cognigy

**Released** June 21, 2024

#### Improvements

* Added the opt-out mechanism for Gravatar interaction to protect privacy for profile pictures. To enable this mechanism, use the `FEATURE_DISABLE_GRAVATAR_FOR_PROFILE_PICTURES=true` feature flag

#### Bug Fixes

* Fixed the issue where both Webchat v2 and Webchat v3 were not responding in cases where the Agent Copilot flowId was configured but no handover provider was selected

### Cognigy Voice Gateway

**Released** June 20, 2024

#### Bug Fixes

* Fixed the issue where the `AudioCodes` disconnect event caused a `503 Service Unavailable` error instead of the expected `200 OK` response

### Infrastructure

#### Cognigy Voice Gateway

* Fixed the issue where the image name of `feature-server` was incorrect

## 4.78.0

### Cognigy

**Released** June 12, 2024

#### Features

* Added charts for [monitoring the number of Knowledge Chunks and Queries metrics](/ai/administer/access/admin-center/dashboard) according to a [new license](/ai/agents/develop/knowledge-ai/activate#apply-for-a-license) for the Knowledge AI feature. To ingest new Knowledge Sources after the v4.78.0 update, a new license key for the Knowledge AI feature should be updated in the Cognigy installation, and the maximum Knowledge Chunks quota should be allocated for the organization.

#### Improvements

* Added the functionality to record user acceptance of the [Privacy Notice](/webchat/v3/features#privacy-notice) within the [Contact Profile](/ai/agents/analyze/contact-profiles#profile-schema)
* Added the [Minimize](/webchat/v3/features#controlling-the-chat-window) button at the top of the Chat screen within the Webchat v3 interface
* Updated the Webchat v3 `preview` tag to `beta`
* Renamed the Organization Details page to [Admin Center](/ai/administer/access/admin-center/overview)
* Removed the Knowledge AI trial notification from the UI
* Improved the Agent Copilot section in the Endpoint Settings to differentiate between configurations for the Agent Copilot Whisper and Agent Copilot workspace
* Updated the Agent Copilot configuration. Adding the Agent Copilot config via API is no longer mandatory. In the Copilot section of the Endpoint Settings, select [Workspace > None](/agent-copilot/getting-started/overview) and use a [Copilot: Set Grid Node](/ai/agents/develop/node-reference/agent-copilot/set-grid) to set up the workspace within your Flow
* Added support for multiple languages in addition to English
* Increased the size limit of a custom field in the Cognigy OData Endpoint. Now, the maximum length of a custom field is 1024 characters

#### Bug Fixes

* Fixed the issue of insufficient input validation enabling HTML injection in notification emails
* Fixed the issue where deleting a connection selected in the HTTP Request Node breaks Node execution, even after setting the authType to `none`
* Fixed the issue w